PDA

View Full Version : Planned Rating Changes



Pages : [1] 2

Bill Gletsos
20-06-2004, 07:18 PM
Over the past 3-4 mths the ACF Ratings Officers have been experimenting with a number of possible changes to the rating system.

The main aims being :
a) To better handle the impact on players who play rapidly improving players (or for that matter rapidly declining players)
b) To better handle players who are new to the normal list even though they may be active on the rapid list and vice versa.

With regards to a) this is achieved by generating an intermediate rating for all players. For the calculation of the player's rating his rating at the start of the period is still used but the opponents intermediate rating is used as the opponents rating rather than their rating at the start of the rating period.

For example back in 2002 and up to the March 2003 rating period Ronald Yu's rating had been between 1691 in April 2002 and 1776 in March 2003. Now for the June 2003 ratings Ronalds opponents had their results calculated based on his March 2003 rating of 1776 yet he was clearly playing at a much higher level. It therefore seems reasonable for Ronalds opponents to treat him not as 1776 but as a player rated more like 2050. BTW Ronalds performance rating during this period was 2225. Testing shows that it is indeed better to do this as it leads to better predictive accuracy. It should be noted that the USCF uses this intermediate rating idea in their rating calculations.

With regards b) this is achieved by increasing the RD factor of the player when his rapid is used as the seed rating for his first normal games or vice versa in the case of his first rapid games. This increase in RD was already done previously but it was found to improve predictive accuracy if we increased it more. Note increasing it to the point of maximum RD(350) actually reduces predictive accuracy, which we already knew and I had mentioned previously on this board.

Lastly we are investigating whether we should raise the rating floor from is current value of zero to possibly 100.

Cat
20-06-2004, 07:34 PM
Well done, it's good to see you finally recognise there is a problem! What criterion will you use to determine that a player meets the status of 'rapidly improving'? And how will this work in a pool of players that are generally improving as a cohort, eg groups of juniors that mainly play each other, as might apply at the Gardiner Chess Centre, for example?

Bill Gletsos
20-06-2004, 07:56 PM
Well done, it's good to see you finally recognise there is a problem!
Ha ha ha. Now why doesnt it surprise me that you would be the first to reply. :whistle:


What criterion will you use to determine that a player meets the status of 'rapidly improving'?
The system already detects rapidly improving players provided they have a non trivial number of games actually submitted for rating. As to how it works I already explained it in the post. :hand:


And how will this work in a pool of players that are generally improving as a cohort, eg groups of juniors that mainly play each other, as might apply at the Gardiner Chess Centre, for example?
Havent we been through this before.
I suggest you look back at the previous threads. :hand:

Alan Shore
20-06-2004, 08:30 PM
Sounds like a good idea Bill, I know a lot of players have suffered from playing underrated players and hopefully the new technique might better protect against possible future deflation.

Having the rating floor at 100 is a great addition too, I and a couple of others suggested implementing it on a previous thread so thanks for considering it.

I was wondering though, with the way you want to shape the rating system would I be correct in assuming you would want optimally, a figure for each player that represents their true ability closer than simply their past performances?

Bill Gletsos
20-06-2004, 08:38 PM
Sounds like a good idea Bill, I know a lot of players have suffered from playing underrated players and hopefully the new technique might better protect against possible future deflation.
There isnt any evidence of deflation using the current method and you would want to maintain that with any changes to the system.


Having the rating floor at 100 is a great addition too, I and a couple of others suggested implementing it on a previous thread so thanks for considering it.
Yeah I'm still looking at this.


I was wondering though, with the way you want to shape the rating system would I be correct in assuming you would want optimally, a figure for each player that represents their true ability closer than simply their past performances?
This is always a tricky question.
E.g make the players new rating their performance rating for the period. Afetr all that must surely represent their true ability.
Bzzzt. Wrong. :doh:

What rating systems come down to is merging the players current performance into his previous performance. Hopefully doing it in such a way that their new rating is the best predictor of their future results.
Of course you want everyones rating to meet that criteria.
Easier said than necessarily done. ;)

PHAT
20-06-2004, 09:21 PM
What rating systems come down to is merging the players current performance into his previous performance. Hopefully doing it in such a way that their new rating is the best predictor of their future results.
Of course you want everyones rating to meet that criteria.
Easier said than necessarily done. ;)


"Hopfully doing it it in such a way ..." :eek:

"Easier said than necessarily done." :eek:

This reeks of mediocracy. Should it be rewarded with a kiss?

Edit: Lemme say I am not having a go at your hard work, just your system.

Bill Gletsos
20-06-2004, 09:24 PM
"Hopfully doing it it in such a way ..." :eek:

"Easier said than necessarily done." :eek:

This reeks of mediocracy. Should it be rewarded with a kiss?
I was simply trying to explain to BD the issues.
Also I never said that was the any particular system.
I was talking in general terms.

Edit: Let me say I am having a go at you. ;) :hand:

Cat
20-06-2004, 09:30 PM
Ha ha ha. Now why doesnt it surprise me that you would be the first to reply. :whistle:

Because after criticising you continually for not listening, I wanted to be the first to congratulate you for doing so.


The system already detects rapidly improving players provided they have a non trivial number of games actually submitted for rating. As to how it works I already explained it in the post. :hand:

No, you've given a single example without any detail of how you arrived at 2050. Is it linked to volatility?


Havent we been through this before.
I suggest you look back at the previous threads. :hand:

No, you've never suggested this idea before. It will clearly have a beneficial effect.

ursogr8
20-06-2004, 09:32 PM
Over the past 3-4 mths the ACF Ratings Officers have been experimenting with a number of possible changes to the rating system.

The main aims being :
a) To better handle the impact on players who play rapidly improving players (or for that matter rapidly declining players)
b) To better handle players who are new to the normal list even though they may be active on the rapid list and vice versa.

With regards to a) this is achieved by generating an intermediate rating for all players. For the calculation of the player's rating his rating at the start of the period is still used but the opponents intermediate rating is used as the opponents rating rather than their rating at the start of the rating period.

.

Bill

Another rating thread eh. Will be good for some post counts of certain people.

BTW, Christopher Wallis is a case that bears looking at. He has just had two super tournaments after the last rating calc. and he is going to cause a major distortion to our Interclub team selections (we run 25 teams).

starter

Cat
20-06-2004, 09:38 PM
"Hopfully doing it it in such a way ..." :eek:

"Easier said than necessarily done." :eek:

This reeks of mediocracy. Should it be rewarded with a kiss?

Edit: Lemme say I am not having a go at your hard work, just your system.

I'd say this would be the mechanism he used to correct the ratings on the Gold Coast, which lead to an average increase of around 65 pts.

Bill Gletsos
20-06-2004, 09:39 PM
Because after criticising you continually for not listening, I wanted to be the first to congratulate you for doing so.
I wasnt sure how to take your orignal comment.


No, you've given a single example without any detail of how you arrived at 2050. Is it linked to volatility?
You simply do an initial calculation using the standard Glicko2 formula. It gives the 2050.



No, you've never suggested this idea before. It will clearly have a beneficial effect.
You misunderstand me.
You were obviously referring to supposed isolated rating pools.
It was that I was referring to when I said read the previous threads.

There are lots of ideas we may be looking at.
I just dont bother wasting my time discussing them on the BB.

Once tests have been done and if the results positive then I'll comment on them.

Bill Gletsos
20-06-2004, 09:42 PM
I'd say this would be the mechanism he used to correct the ratings on the Gold Coast
Then you would be wrong.
Of course I had explained at the time what we did.
I'm not going to waste my time repeating it.

Cat
20-06-2004, 09:45 PM
There are lots of ideas we may be looking at.
I just dont bother wasting my time discussing them on the BB.

Once tests have been done and if the results positive then I'll comment on them.

I don't think I've been wasting you're time, Bill - I hope I've been providing impetus for change, simply throwing up some ideas!

Bill Gletsos
20-06-2004, 09:48 PM
Bill

Another rating thread eh. Will be good for some post counts of certain people.
I was sorely tempted just to wait until it was implemented before commenting so as not to put up with stupid replys.(DR this is not referring to you so far)


BTW, Christopher Wallis is a case that bears looking at. He has just had two super tournaments after the last rating calc. and he is going to cause a major distortion to our Interclub team selections (we run 25 teams).
Unfortunately there isnt much that can be done for him so far based on reported results. Even under the new system he would only be about 1656.
Eugene Schon on the other hand would be around 1238.

Cat
20-06-2004, 09:50 PM
Then you would be wrong.
Of course I had explained at the time what we did.
I'm not going to waste my time repeating it.


Then maybe you should, run it retrospectively, it might tidy up many anomalies.

Bill Gletsos
20-06-2004, 09:52 PM
Then maybe you should, run it retrospectively, it might tidy up many anomalies.
I'm way ahead of you on that.
Naturally for the proposed changes I have already recalculated all the ACT and GC related corrections.

ursogr8
20-06-2004, 09:56 PM
I was sorely tempted just to wait until it was implemented before commenting so as not to put up with stupid replys.(DR this is not referring to you so far)


Unfortunately there isnt much that can be done for him so far based on reported results. Even under the new system he would only be about 1656.
Eugene Schon on the other hand would be around 1238.

Bill
What about just pushing the button and having a June 20th ratings version to replace the June version. It is just a formula and a data-base isn't it? Or does it have manual intervention?

starter

Cat
20-06-2004, 10:06 PM
I'm way ahead of you on that.
Naturally for the proposed changes I have already recalculated all the ACT and GC related corrections.


Thanks Bill, I'll await the results with interest. Gee, I guess its too late to apply for the Olympiad?

Bill Gletsos
20-06-2004, 11:26 PM
Bill
What about just pushing the button and having a June 20th ratings version to replace the June version. It is just a formula and a data-base isn't it? Or does it have manual intervention?

starter
There is no manual intervention.
However it wont be implemented yet.
I have a few more tests I want to run.

Plus I'm still looking at long inactive players and their April 2000 150 point uplift.

Kevin Bonham
21-06-2004, 12:02 AM
"Hopfully doing it it in such a way ..." :eek:

"Easier said than necessarily done." :eek:

This reeks of mediocracy. Should it be rewarded with a kiss?

If it had been less equivocal you would be saying it reeked of arrogance.

arosar
22-06-2004, 12:55 PM
Look...I feel a bit left our if I don't say anything in a ratings thread. Besides, it's not fair at all when only certain persons can increase their post counts. So, I'm just yappin' on here for no particular reason at all. Anyways, if everyone reckons it's a good idea, then it must be a good idea eh?

AR

Oepty
22-06-2004, 01:29 PM
My first reaction when I saw the thread was, NOT ANOTHER RATING SYSTEM CHANGE. But on reading the proposed change it does seem a reasonably good idea. Bill just stop having these good ideas, we do need a rating system that isn't continually being changed sometime.
Scott

Rincewind
22-06-2004, 01:44 PM
Any change which can be shown to increase the predictivity of the rating pool as a whole must be (by definition) a good idea.

arosar
22-06-2004, 01:56 PM
What's this 'predictivity' business? Youse all mean predictability, don't you?

AR

Rincewind
22-06-2004, 02:11 PM
What's this 'predictivity' business? Youse all mean predictability, don't you?

Not predictability. Any rating system should be entirely predictable - ie for a known set of inputs the output can be calculated with certainty. Predictivity is a fuction of the system to produce ratings which match the future performance of players.

Any change which increases the predictivity (means that players perform more closely to their current published rating) is a good one.

PHAT
22-06-2004, 03:51 PM
Any change which increases the predictivity (means that players perform more closely to their current published rating) is a good one.

S, what is the exact part of the Glicko system that extrapolates to the midpoint of the current rating period? :rolleyes:

Rincewind
22-06-2004, 04:38 PM
S, what is the exact part of the Glicko system that extrapolates to the midpoint of the current rating period? :rolleyes:

None of it. It doesn't make any attempt to trend rating movement across rating periods.

Bill's proposed change will go part of the way there by using a 1st appoximation of a players current period rating, instead of their last period rating when performing the calculation of the current rating. That is all.

It is sort of like using the rating midpoint of the last rating period to determine the current rating - but only by analogy. This is not strictly speaking true.

PHAT
22-06-2004, 05:38 PM
None of it. It doesn't make any attempt to trend rating movement across rating periods.

I know. That is why I asked the question.




Bill's proposed change will go part of the way there by using a 1st appoximation of a players current period rating, ...

An all this time Bill et al. have been saying there was no need to extrapolate or account for trends, because "Glicko handles it." This capitulation is a precious and sweet victory. :owned:

Rincewind
22-06-2004, 06:07 PM
An all this time Bill et al. have been saying there was no need to extrapolate or account for trends, because "Glicko handles it." This capitulation is a precious and sweet victory. :owned:

That's why I said it wasn't strictly true. The change doesn't involve extrapolation at all. Which is why I originally described it as a 1st approximation. The 1st approximation is based on results already received and not an extrapolation based on trend analysis.

You can claim any change in the rating system as a capitulation by Bill. However, the fact is it is Bill and Graham's work which lead to this change and as it is not based on trend analysis I think your claim of a victory is disingenuous trolling.

Bill Gletsos
22-06-2004, 06:16 PM
I know. That is why I asked the question.
Ok so you know it makes no use of trend analysis.


An all this time Bill et al. have been saying there was no need to extrapolate or account for trends, because "Glicko handles it." This capitulation is a precious and sweet victory. :owned:
What a load of crap.
Your claim of victory is a hollow claim. :hand:
Sort of like the French victory at the Battle of Waterloo.
Oh thats right they lost. :owned:

You just admitted above it makes no use of trend analysis.
Your so called tri-rank system supposedly did take into account trends but in your only attempt at it previously where you published expected ratings you failed abysmally. :whistle:

Cat
23-06-2004, 12:02 PM
I know. That is why I asked the question.




An all this time Bill et al. have been saying there was no need to extrapolate or account for trends, because "Glicko handles it." This capitulation is a precious and sweet victory. :owned:

Sadly, we're all losers, especially Asutralian Chess. Because of denial and intransigence its taken over 12 months to get to a situation where a problem is (at least to some degree) acknowledged and to a point which appears vaguely reasonable.

As many of the corrections have been made in an essentially ad hoc manner, points given to correct what ELO created 4yr ago (nothing to do with Glicko you understand), it is the inactive players that have benefited most. I've gained 200 pts for doing nothing - where's the link between effort & reward I ask myself.

Now this is not to say I am not grateful for the changes, its been a real help for the Gold Coast. But if we try to delude ourselves all is well in the garden, fail to recognise the difficulties we have found ourselves in, and don't take steps to monitor the system to ensure it's working reasonably effectively, then we could well find ourselves in the same situation in a few years.

It also says a lot about how Australian Chess is run, that such a simple issue cannot be tackled directly without a load of recrimination and abuse. Is this any way for any sporting association to behave? What hope does Australian Chess have if it cannot face up to problems that are patently obvious to the vast majority of players, where a system has become more precious than the game?

The rating system and the ratings look very different from 12 months ago, even though we were told there was no problem back then. Which is right? Either there was a problem and its been improved, or there was no problem -in which case one wonder why all these changes have been necessary?

Bill, it doesn't matter if things aren't working perfectly, no one is blaming you for the problems. All one is asking is some respect can be shown for the issues raised and proper attempts are made to answer those questions instead of this endless obsfication. Bill, you are not the Pentagon, you have nothing to hide. We will respect you more if you learn to respond sympathetically to peoples concerns. If at times attacks have seemed strident, it is you who have made it so.

I'm happy to see how things work out, and I'll be monitoring the results later in the year to gaun some measure of the effects of the changes. Maybe this is the time to end all our differences and recognise we are all on the same side, all seeking the same ends, a better environment in which to enjoy our chess and one that encourages everyone to compete.

Alan Shore
23-06-2004, 12:14 PM
The rating system and the ratings look very different from 12 months ago, even though we were told there was no problem back then. Which is right? Either there was a problem and its been improved, or there was no problem -in which case one wonder why all these changes have been necessary?

Aren't you being a bit harsh on Bill, David? The ratings look different since they were increased to be brought more in line with FIDE standards.. it's just a figure to represent playing strength, relative to your opponents and now it's becoming more relative to global rating standards rather than just Australian standards. The fact your rating has increased from not playing doesn't mean all that much since everyone else's rating was also increased.

As for the other concerns, they're being addressed. Bill still maintain there is absolutely no deflation inherent in Glicko, but I think it sounds a little like John Howard when he once said 'There will be no GST, never ever!' :confused: Time will tell I suppose.

Rincewind
23-06-2004, 12:46 PM
Aren't you being a bit harsh on Bill, David? The ratings look different since they were increased to be brought more in line with FIDE standards.. it's just a figure to represent playing strength, relative to your opponents and now it's becoming more relative to global rating standards rather than just Australian standards. The fact your rating has increased from not playing doesn't mean all that much since everyone else's rating was also increased.

As for the other concerns, they're being addressed. Bill still maintain there is absolutely no deflation inherent in Glicko, but I think it sounds a little like John Howard when he once said 'There will be no GST, never ever!' :confused: Time will tell I suppose.

I don't know that Dr Dave is talking about deflation and FIDE alignments. I suspect he is pushing the rating system should use the player's age as a parameter argument.

The point with rating systems is that something is better than nothing, and some systems are better than others. It is all a matter of relative improvement.

I don't think Bill ever said Glicko can't be improved upon, just that there is nothing that has proven to be better. Bill and others have been busy looking at a number of possible tweeks to the system to see if any of them might be better. It appears one has turned up as showing an increased predictivity over the status quo and so it seems likely this will be adopted. I suspect that most of the tweeks tested were not shown to have any benefit (or some detriment) and so were discarded.

Coming up with new concepts is good - imagination is more important than knowledge (Einstein) - however, you have to back them up with proof if you want anyone to take any notice of you.

Cat
23-06-2004, 01:44 PM
I don't know that Dr Dave is talking about deflation and FIDE alignments. I suspect he is pushing the rating system should use the player's age as a parameter argument.

The point with rating systems is that something is better than nothing, and some systems are better than others. It is all a matter of relative improvement.

I don't think Bill ever said Glicko can't be improved upon, just that there is nothing that has proven to be better. Bill and others have been busy looking at a number of possible tweeks to the system to see if any of them might be better. It appears one has turned up as showing an increased predictivity over the status quo and so it seems likely this will be adopted. I suspect that most of the tweeks tested were not shown to have any benefit (or some detriment) and so were discarded.

Coming up with new concepts is good - imagination is more important than knowledge (Einstein) - however, you have to back them up with proof if you want anyone to take any notice of you.

That's all fair enough except that parts of the 'solution' has been to simply dump points into the system, whereby inactive players have experienced rating increases in real terms and maybe relative to the players who are active. That will be part of the test of the changes to demonstrate those playing the game can improve their rating more effectively than waiting for the corrections should they arrive.


As I say, the ratings system should reward effort & performance. Let's see what the future holds.

Bill Gletsos
23-06-2004, 01:50 PM
That's all fair enough except that parts of the 'solution' has been to simply dump points into the system, whereby inactive players have experienced rating increases in real terms and maybe relative to the players who are active.
It is statements like this that demonstrate what a total goose you are when it comes to discussing ratings. :whistle:

Ratings are based on differences. Hence if everyone gets the same points (the 150 in April 2000 and the 70 in March 2004) added it makes no damn difference. :doh:


That will be part of the test of the changes to demonstrate those playing the game can improve their rating more effectively than waiting for the corrections should they arrive.
Those changes were to maintain FIDE relativity not as corrections.

ursogr8
23-06-2004, 02:21 PM
There is no manual intervention.
However it wont be implemented yet.
A I have a few more tests I want to run.


B Plus I'm still looking at long inactive players and their April 2000 150 point uplift.

Bill

A and B and probably important, but not time-urgent.
In the interim, why not just press the button?

starter

Bill Gletsos
23-06-2004, 02:29 PM
Sadly, we're all losers, especially Asutralian Chess. Because of denial and intransigence its taken over 12 months to get to a situation where a problem is (at least to some degree) acknowledged and to a point which appears vaguely reasonable.

As many of the corrections have been made in an essentially ad hoc manner, points given to correct what ELO created 4yr ago (nothing to do with Glicko you understand), it is the inactive players that have benefited most. I've gained 200 pts for doing nothing - where's the link between effort & reward I ask myself.
The problem is as usual that you misrepresent the facts.
Because the important thing with ratings is the difference between them and not the absolute value then since everyone got the 150 points back in April 2000 (except Ian Rogers) it made no difference.
As regards you specifically you did not appear on the rating lists with a published rating untill December 2000. It is therefore superious at best to suggest you received those 150 points due to inactivity since you only first appeared on the list in April 2000.



Now this is not to say I am not grateful for the changes, its been a real help for the Gold Coast. But if we try to delude ourselves all is well in the garden, fail to recognise the difficulties we have found ourselves in, and don't take steps to monitor the system to ensure it's working reasonably effectively, then we could well find ourselves in the same situation in a few years.
Unfounded speculation based as usual on no evidence.



It also says a lot about how Australian Chess is run, that such a simple issue cannot be tackled directly without a load of recrimination and abuse. Is this any way for any sporting association to behave? What hope does Australian Chess have if it cannot face up to problems that are patently obvious to the vast majority of players, where a system has become more precious than the game?
We have never claimed the system is perfect.
All we have done with any hackneyed theory people had to improve the system was to demand that it be backed up by actual proof. In virtually all cases this was not done.



The rating system and the ratings look very different from 12 months ago, even though we were told there was no problem back then. Which is right? Either there was a problem and its been improved, or there was no problem -in which case one wonder why all these changes have been necessary?
Unfortunately David, you seem to lack the understanding that before making changes you need to know whether they are benefical or not, not just lets make the change to try it and see how it goes.

This has been said before but you and others just seem to conveniently forget it and then raise it again much later to hopefully convince a new crop of uninformed readers. Therefore I'll point out the changes that have been made since August 2000.

1) In December 2000 the Glicko system was introduced to replace the Elo system. The Glicko was much better at handling juniors than the Elo.
2) Prior to the calculation of the August 2000 list reduced the aging of the RD factor and recalculated all previous rating peiords to take account of it.
3) The description of the Glicko2 system was finally publicaly released and in January 2002 we started testing the Glicko2 system and implemented it for the December 2002 list. All previous rating periods were recalulated. The Glicko2 was superior to handling juniors than the Glicko.
4) In 2003 we switched to producing rating lists 4 times a year. We had hoped the Glicko and Glicko2 systems would resolve the problems of the ACT junior problem caused by the Elo system over many years. This turned out not be the case. We therefore took about carrying out a study to determine the best means of accomplishing this. This was implemented after the September 2003 rating period as a special November 2003 list.
5) We had been aware of a possible anomaly in the system with regards players rated over 2200 for some time. This was investigated at the same time as we were looking at the ACT junior problem. A solution to this anomaly was introduced for the September 2003 list with all previous rating periods from Dec 2000 were recalculated.
6) In March 2004 all players received a 70 point uplift to try and align their ratings with FIDE.
7) Prior to the September 2004 ratings we are looking at implementing the changes described in the first post.


Now you may call these ad-hoc. I disagree
Whats important is assigning priority to implmenting changes that affect the majority of players, not minorities.

The implementation of the Glicko was planned.
Once we became aware of the Glicko2 system we planned for its implementation but we had to wait for it to be publically released. As soon as it was we started investigating it.
Although we were aware of the ACT junior problem we had to wait for a number of rating periods under Glicko to see if it would alleviate it. However by this time we were planning the implmentation of the Glicko2. Initial testing suggested this may solve the ACT junior problem. However it was determined whilst doing the Glicko2 testing from Jan 2002 till implementation in Sept 2002 that this would not be the case. Therefore at the same time as the ACF Council authorised us to implement the Glicko2 they also authorised us to take steps to resolve the ACT problem.
Resolving this was far from a trivial issue.
We became aware of an anomaly regarding 2200 and above players in late 2002 and started investigating it, but the ACT junior problem was considered more important an issue to be resolved first.
However once our method of recitifing the ACT junior problem was determined and tests run we were able to finalise the solution to the anomaly for 2200+ players. The solution was implemented for the Sept 2003 list and shortly thereafter the corrections for the ACT junior problem were implented in the Nov 2003 list.

Now I mentioned back in the March 2004 ratings thread in amongst all the crap you were posting in a query (#263) by starter regarding a particular junior and his followup in #273 about the use of rapid ratings to seed normal and vice versa. In my reply #275 on the 6th April 2004 I made it clear to him why this was done. I also explained to him that perhaps increasing the RD would be benefical but would require further testing but not expect an answer in the short term.
Well blow me down guess what, I did the testing and announced it in this thread amongst other things.

Finally as for the use of intermediate ratings I first mentioned this back in June 2003 in the Rating Theory 1001 thread on the old BBS. However with regards the ACT junior problem, the 2200+ anomaly and the FIDE realignment investigation it got put on the back burner until this year. To therefore suggest this is ad-hoc is a joke.

The trouble is you need to be more observant.


Bill, it doesn't matter if things aren't working perfectly, no one is blaming you for the problems. All one is asking is some respect can be shown for the issues raised and proper attempts are made to answer those questions instead of this endless obsfication. Bill, you are not the Pentagon, you have nothing to hide. We will respect you more if you learn to respond sympathetically to peoples concerns. If at times attacks have seemed strident, it is you who have made it so.
The problem remains, you continually make unfounded and unsupported statements or accusations.
You then seem to wonder why your theories are given no credence.



I'm happy to see how things work out, and I'll be monitoring the results later in the year to gaun some measure of the effects of the changes.
Monitor all you want.
Just try and make sure your conclusions are based on fact and not fantasy and personal beliefs.
Be aware however that although the proposed changes are beneficial to all players the effect on the majority of players is minor.



Maybe this is the time to end all our differences and recognise we are all on the same side, all seeking the same ends, a better environment in which to enjoy our chess and one that encourages everyone to compete.
The only differences I'm having are with you and Matt. :wall:
I'm getting along fine with Kevin, Barry, starter and most people. :hand:

Cat
23-06-2004, 02:39 PM
It is statements like this that demonstrate what a total goose you are when it comes to discussing ratings.

Ratings are based on differences. Hence if everyone gets the same points (the 150 in April 2000 and the 70 in March 2004) added it makes no damn difference. :doh:

These are quotes from players I have had communication with in the last few weeks. When asked to compete in our club championship one player said;
"the last time I competed in a competition like that I lost 300 rating points. It's not that I'm worred about my rating per se, it just that I'm trying to get my rating back up to 2000 so I can compete in the Queensland Championships. I nearly missed out last time because of my rating".

Another player said;

"I've never been so highly ranked. It's because I haven't been playing any competition, the corrections have increased my ratings. If I start competing my rating will probably fall."

These players weren't complaining about the rating system, they were just giving a straightforward answer to my questions. What I like to see is players saying they need to compete to gain rating points and have some degree of confidence this might be achieved.

When I looked at the Gold Coast Classic (1998-2002) the average adult lost 4 rating points for every 2 pts gained by a junior. I'll be using BJC's calculator to perform the same calculation this year. Let's hope we can demonstrate that they have about even odds of a rating increase.

Bill Gletsos
23-06-2004, 03:02 PM
These are quotes from players I have had communication with in the last few weeks. When asked to compete in our club championship one player said;
"the last time I competed in a competition like that I lost 300 rating points. It's not that I'm worred about my rating per se, it just that I'm trying to get my rating back up to 2000 so I can compete in the Queensland Championships. I nearly missed out last time because of my rating".
How can he get his rating up if he isnt competing.
I assume you mean he just isnt prepared to compete in the the Gold Coast Championship in which case maybe you should run it in multiple divisions like North Sydney, St.George and a number of other clubs do.



Another player said;

"I've never been so highly ranked. It's because I haven't been playing any competition, the corrections have increased my ratings. If I start competing my rating will probably fall."
This one might be in for a rude shock when we implement the motion passed by the ACF at the October 2003 Council meeting which said:
"That the ACF Ratings Officers are authorised to deduct rating points from players who had been inactive over the period 1980-1999 to offset any anomalies caused by the 150 point bonus added to players back in April 2000. The method of determination and the number of points to be deducted is left to the ACF Rating Officers to decide."



These players weren't complaining about the rating system, they were just giving a straightforward answer to my questions. What I like to see is players saying they need to compete to gain rating points and have some degree of confidence this might be achieved.
With regards player one above you may need to restructure your tournaments to make it attractive to him to play. i.e. instead of just running big swisses all the time try different formats to cater to your members varying needs.
As for player number 2, this one may require a sudden loss of rating points to get him to resume. ;)


When I looked at the Gold Coast Classic (1998-2002) the average adult lost 4 rating points for every 2 pts gained by a junior.
Yes, a whole thread on the old BBS raised serious questions regarding the method you employed so the less said about it the better.


I'll be using BJC's calculator to perform the same calculation this year. Let's hope we can demonstrate that they have about even odds of a rating increase.
I'll wait with baited breath. ;) :whistle:

rob
23-06-2004, 03:04 PM
Over the past 3-4 mths the ACF Ratings Officers have been experimenting with a number of possible changes to the rating system.

The main aims being :
a) To better handle the impact on players who play rapidly improving players (or for that matter rapidly declining players)
b) To better handle players who are new to the normal list even though they may be active on the rapid list and vice versa.

With regards to a) this is achieved by generating an intermediate rating for all players. For the calculation of the player's rating his rating at the start of the period is still used but the opponents intermediate rating is used as the opponents rating rather than their rating at the start of the rating period.

For example back in 2002 and up to the March 2003 rating period Ronald Yu's rating had been between 1691 in April 2002 and 1776 in March 2003. Now for the June 2003 ratings Ronalds opponents had their results calculated based on his March 2003 rating of 1776 yet he was clearly playing at a much higher level. It therefore seems reasonable for Ronalds opponents to treat him not as 1776 but as a player rated more like 2050. BTW Ronalds performance rating during this period was 2225. Testing shows that it is indeed better to do this as it leads to better predictive accuracy. It should be noted that the USCF uses this intermediate rating idea in their rating calculations.


This seems like an excellent enhancement to the ratings system.

A WA example of a rapid riser whose victims would benefit from the suggested change:

In a 6 rnd swiss in May junior Thomas Donaldson (then 1625) played 3 high rated players: drew with AH 1911, RM 1973, TH 2047. In June he is 1683 and in the last 6 rnd weekender he beat AH 1892, GD 1768, unr (1600ish), drew with RM 1949, MP 2013, TH 2059. As Thomas's opportunities to play rated games are severely hampered by living near Albany (5 or so hours from Perth) his rating is unlikely to rise quickly enough to reflect his strength (he plays a lot of internet chess).

Each state probably has one or more similar examples.

Bring it on Bill :)

Garvinator
23-06-2004, 03:19 PM
[I][B]"the last time I competed in a competition like that I lost 300 rating points.
To lose three hundred rating points, the player would have had to have played very badly according to their rating expectation. If they played badly and didnt perform near to their rating, then they deserve to lose points.


I'll be using BJC's calculator to perform the same calculation this year. Let's hope we can demonstrate that they have about even odds of a rating increase.
David, just be careful when using barry calculator as it calculates under glicko, not glicko 2. this will mean that some results you achieve will not be accurate.

Rincewind
23-06-2004, 03:24 PM
David, just be careful when using barry calculator as it calculates under glicko, not glicko 2. this will mean that some results you achieve will not be accurate.

What you say is true and note that for various reasons no completely accurate rating calculator is possible, especially after Bill's proposed change.

That being said, I've found my Rating Estimator to be reasonably accurate. (Estimated my last rating shift by 2 points from memory).

Cat
23-06-2004, 03:53 PM
Bill, I know you've done your best, and we have to be satisfied with that. So let's not argue when we agree this represents something of an improvement. Let's look to the future and hope that your best is good enough.

Alan Shore
23-06-2004, 04:05 PM
To lose three hundred rating points, the player would have had to have played very badly according to their rating expectation. If they played badly and didnt perform near to their rating, then they deserve to lose points.

In a way I agree with this, the purpose of a rating is to give an accurate indication of the player's strength, not necessarily his past achievements. If he's playing like crap, his rating must go down, and a lot too - if he's good enough it'll rise again. Similarly with those performing very strongly - increase their rating to match their playing strength, glicko will automatically correct it if the player overshoots.

If we got some kind of $1 million government grant for chess ratings, then they should be updated after every game! Then you could think about how the player lost.. if they just left a piece en prise, give them a rating in accordance with their performance. Although I guess this means Nigel Short would have lost about 1000 points.. haha.

Garvinator
23-06-2004, 04:06 PM
Bill, I know you've done your best, and we have to be satisfied with that. So let's not argue when we agree this represents something of an improvement. Let's look to the future and hope that your best is good enough.
david, can you alter this post so you dont quote everything bill said, please :)

thank you david:)

Alan Shore
23-06-2004, 04:12 PM
david, can you alter this post so you dont quote everything bill said, please :)

Yeah, I hate when people quote something that's like 2000 words and their reply is only one sentence.. just quote the relevant parts or not at all..

PHAT
23-06-2004, 05:13 PM
The only differences I'm having are with you and Matt. :wall:
I'm getting along fine with Kevin, Barry, starter and most people. :hand:

"Most people", or "most people whom you can bamboozle?" CZ is a PhD mathematician. How come he has given up on ratings debates with you?

As for TRIRANK, it sits embrio-like in my bottom draw, an opportunistic beast awaiting further decrepitude in the Glicko and its labrenth of barely supporting fudge fatcors.

We all laugh/cry at our rediculously complicated tax system. It got that way precisely because of special treatment for vested interests. FFS, the Glicko is going the same way. It was a good idea for symetrical non-flux conditions that is applied to a system that is not symetrical and has a net perminant positive flux in component value. Is it any wonder that juniors are chronically underated and many adults don't want to play them. Is it any wonder that the elite bitch that one fluke game can see them tumble.

Garvinator
23-06-2004, 05:22 PM
Is it any wonder that juniors are chronically underated and many adults don't want to play them. Is it any wonder that the elite bitch that one fluke game can see them tumble.
this got tossed in the bin a while ago in my opinion. The fault lies in the fact that the juniors dont play enough normal rated games in a rating period. No system can chart progress if given insufficent data. As for junior v junior tournaments, well if they are not sent in for rating purposes, then the system again is not presented with all possible data for an accurate assessment.

Bill Gletsos
23-06-2004, 05:23 PM
"Most people", or "most people whom you can bamboozle?" CZ is a PhD mathematician. How come he has given up on ratings debates with you?
Just show what you know.
CZ and I have had a number of discussions regarding ratings, some quite recent.
I suspect if CZ would give up talking to anyone about any subject it would be you.


"As for TRIRANK, it sits embrio-like in my bottom draw, an opportunistic beast awaiting further decrepitude in the Glicko and its labrenth of barely supporting fudge fatcors.
TRIRANK was stillborn.
You can call them fudge factors, most would see them as enhancements.


"We all laugh/cry at our rediculously complicated tax system. It got that way precisely because of special treatment for vested interests. FFS, the Glicko is going the same way. It was a good idea for symetrical non-flux conditions that is applied to a system that is not symetrical and has a net perminant positive flux in component value. Is it any wonder that juniors are chronically underated and many adults don't want to play them. Is it any wonder that the elite bitch that one fluke game can see them tumble.
All you show as usual is your stupidity and lack of knowledge of anything chess related.

PHAT
23-06-2004, 05:47 PM
CZ and I have had a number of discussions regarding ratings, some quite recent.


And his opinions were ...?

Just joking, I won't make you lampoon yourself.

arosar
23-06-2004, 05:49 PM
this got tossed in the bin a while ago in my opinion. The fault lies in the fact that the juniors dont play enough normal rated games in a rating period. No system can chart progress if given insufficent data. As for junior v junior tournaments, well if they are not sent in for rating purposes, then the system again is not presented with all possible data for an accurate assessment.

So you're following all this, are you gray? You bas.tard got higher post count than me now.

AR

arosar
23-06-2004, 05:50 PM
And his opinions were ...?

Yeah, I am interested to know too...

AR

Bill Gletsos
23-06-2004, 06:20 PM
And his opinions were ...?

Just joking, I won't make you lampoon yourself.
All I'm prepared to say is that he wasnt criticising the system.

However I seriously doubt CZ had a discussion with you regarding ratings, therefore your original comments scurrilous innuendo.

Bill Gletsos
23-06-2004, 06:21 PM
Yeah, I am interested to know too...

AR
Ah yes the number one gossip.
Well you are out of luck.

Bill Gletsos
23-06-2004, 06:22 PM
this got tossed in the bin a while ago in my opinion. The fault lies in the fact that the juniors dont play enough normal rated games in a rating period. No system can chart progress if given insufficent data. As for junior v junior tournaments, well if they are not sent in for rating purposes, then the system again is not presented with all possible data for an accurate assessment.
Correct gray.
All Matt does is regurgitate the same old rubbish.

Cat
24-06-2004, 09:31 AM
this got tossed in the bin a while ago in my opinion. The fault lies in the fact that the juniors dont play enough normal rated games in a rating period. No system can chart progress if given insufficent data. As for junior v junior tournaments, well if they are not sent in for rating purposes, then the system again is not presented with all possible data for an accurate assessment.

Garvin, juniors play many more rated games than adults at our club. Most of our juniors have !! ratings indicating stability. They play many more games than adults, they play all they can, yet the adults don't experience the same problem. Quite simply the mechanisms within our rating system have been too blunt to properly depict rating changes in a rapidly improving junior population. Either we accept things the way they are to the detriment of all, or we attempt to explain and identify the problem to develop a solution. To some extent the changes Bill have brought in over the last 18 months have been attempting to accomodate this. Garvin, I suggest you return to your waste paper bin and try to develop a better grasp of the problem.

Bill Gletsos
24-06-2004, 12:33 PM
Garvin, juniors play many more rated games than adults at our club. Most of our juniors have !! ratings indicating stability. They play many more games than adults, they play all they can, yet the adults don't experience the same problem. Quite simply the mechanisms within our rating system have been too blunt to properly depict rating changes in a rapidly improving junior population. Either we accept things the way they are to the detriment of all, or we attempt to explain and identify the problem to develop a solution. To some extent the changes Bill have brought in over the last 18 months have been attempting to accomodate this. Garvin, I suggest you return to your waste paper bin and try to develop a better grasp of the problem.
I think you are misrepresenting what garvin is saying and what has been discussed previously.
It was pointed out that the juniors dont play many normal rated games.
It was suggested they need to play more normal rated games and you even agreed with that statement back on another thread.

Also juniors dont all improve at the same rate or at the same time.
To therefore suggest that the rating system cannot handle a group of players who are static, improving slowly, improving steadily and improving very quickly is not founded on fact.

Garvinator
24-06-2004, 01:59 PM
Garvin, juniors play many more rated games than adults at our club.
how many of these games are normal rated, instead of rapid rated?

PHAT
24-06-2004, 03:21 PM
All I'm prepared to say is that he wasnt criticising the system.


Not the [Glicko] system, but how fudge factors are applied. ;)

Bill Gletsos
24-06-2004, 03:37 PM
Not the [Glicko] system, but how fudge factors are applied. ;)
That also is incorrect.
CZ's only criticism was the timing of the 70 point uplift and the fact the NSWCA council voted not to increase the grade match divisions by 100 points. I pointed out to him that I had recommended the NSWCA Council increase the divisions by 100 points but could not get it supported by the other Council members.

Perhaps if you attended council meetings you would be able to function based on facts rather than supposition and innuendo.

Garvinator
24-06-2004, 07:38 PM
why dont you attend council meetings matt?

Cat
24-06-2004, 08:42 PM
how many of these games are normal rated, instead of rapid rated?
Both Garvin, they play many more games rapid & standard

Cat
24-06-2004, 08:51 PM
It was pointed out that the juniors dont play many normal rated games.
It was suggested they need to play more normal rated games and you even agreed with that statement back on another thread.

Yes I agreed that playing more games would help the problem, but they are still by far the more prolific players at our club at all types of chess, but their ratings are woeful compared to the adults, because the rating system has been so insenstive to rapid changes . The changes you've introduced over the last 18 months have certainly helped (if it wasn't so, why make the changes, you can't have it both ways), the question is by what degree?


Also juniors dont all improve at the same rate or at the same time.
To therefore suggest that the rating system cannot handle a group of players who are static, improving slowly, improving steadily and improving very quickly is not founded on fact.

Well if it was handling the changes, why dump 200 pts on some of our players ratings? It clearly wasn't handling the problem at all. Don't keep blaming the ELO, that was 4 year ago. If the Glicko had been effective at all, 4 year should have been enough. In 4 yrs a kid goes from 12 to 16. Wake up Bill.

PHAT
24-06-2004, 10:11 PM
why dont you attend council meetings matt?

1. They are mostly a waste of time. The talk happens via e-mail and the meeting is a rubber stamp on the concencus. The votes might as well be via e-mail as well.

2. I am involved in running a junior chess club on Thursdays and find it more needing of my time than the Bill & Peter show.

3. Meetings start at 7:30 in northern Sydney and run for 4 hours. Since I live somewhat south of the big smoke I have never got home before 1:00am.

4. I will go when there is a need to. So far, there has been two.

Bill Gletsos
25-06-2004, 12:12 AM
1. They are mostly a waste of time. The talk happens via e-mail and the meeting is a rubber stamp on the concencus. The votes might as well be via e-mail as well.

2. I am involved in running a junior chess club on Thursdays and find it more needing of my time than the Bill & Peter show.

3. Meetings start at 7:30 in northern Sydney and run for 4 hours. Since I live somewhat south of the big smoke I have never got home before 1:00am.

4. I will go when there is a need to. So far, there has been two.
Well you no longer have a need to attend any in future. Your position on council has been declared vacant in line with the constitution for failing to attend 3 consecutive council meetings. All of them without any apologies..

Alan Shore
25-06-2004, 12:18 AM
What position does Matthew have? (or rather, did he have, in line with Bill's post..)

Bill Gletsos
25-06-2004, 12:23 AM
Yes I agreed that playing more games would help the problem, but they are still by far the more prolific players at our club at all types of chess, but their ratings are woeful compared to the adults, because the rating system has been so insenstive to rapid changes .
It has been pointed out numerous times that they played very little games that were normally rated.


The changes you've introduced over the last 18 months have certainly helped (if it wasn't so, why make the changes, you can't have it both ways), the question is by what degree?
Changes wont make much difference if they dont play sufficient normal games.


Well if it was handling the changes, why dump 200 pts on some of our players ratings? It clearly wasn't handling the problem at all. Don't keep blaming the ELO, that was 4 year ago. If the Glicko had been effective at all, 4 year should have been enough. In 4 yrs a kid goes from 12 to 16. Wake up Bill.
You really have a comprehension problem dont you.
It has been pointed out that the problem caused by the Elo system could not be resolved in a satifactory time frame by the Glicko system because of the nature of the pool generated by the Elo system in the first place. Hence the corrections made for the ACT junior ratings and also the GC ratings. It has been pointed out that where as very few of the ACT juniors actually would have lost rating points if we had decided not to disadvantage any player by the correction this most definately was not the case with regards the FC. Many players should have lost rating points. However since the GC correction only effected less than 100 players it was deemed not to significant.

Bill Gletsos
25-06-2004, 12:27 AM
What position does Matthew have? (or rather, did he have, in line with Bill's post..)
Publicity officer.
However he only attended 2 out of 6 council meetings this year, sent no apologies for any of them and has been virtually non-existent in any email debates. In fact the last time he was involved in an email debate within council was around April 5th.

Cat
25-06-2004, 12:45 AM
It has been pointed out numerous times that they played very little games that were normally rated.

Bill, this is simply nonsense. When I looked at our rating pool, around 105 players, around 30 adults, the juniors had rating generally much more stable than the adults. They play many more games than the adults. You are living in a fantasy world where you cannot accept information that is uncomfortable to you.




It has been pointed out that the problem caused by the Elo system could not be resolved in a satifactory time frame by the Glicko system because of the nature of the pool generated by the Elo system in the first place.

4 yrs wasn't long enough. I reckon you could have given it another 4 yrs, and it still wouldn't have been long enough. It was never going to deal with the problem of junior ratings without major surgery.

Garvinator
25-06-2004, 12:49 AM
They play many more games than the adults. You are living in a fantasy world where you cannot accept information that is uncomfortable to you..
that is probably true, but were those games played in the normal system and not in the rapid system?

Bill Gletsos
25-06-2004, 12:52 AM
Bill, this is simply nonsense. When I looked at our rating pool, around 105 players, around 30 adults, the juniors had rating generally much more stable than the adults. They play many more games than the adults. You are living in a fantasy world where you cannot accept information that is uncomfortable to you.
Your saying it does not make it a fact. Many of your juniors had very few normally rated games.



4 yrs wasn't long enough. I reckon you could have given it another 4 yrs, and it still wouldn't have been long enough. It was never going to deal with the problem of junior ratings without major surgery.
Your failure to understand is not my problem. You just continually choose to ignore the facts.
I keep telling you that given the nature of the pool generated by the Elo system with regards the ACT juniors etc no rating system could correct it in a timely manner if at all without manual correction.

Cat
25-06-2004, 12:54 AM
that is probably true, but were those games played in the normal system and not in the rapid system?


On the current ACF list 29 of our juniors are represented with !! ratings, only 6 adults. These are normal games Garvin

Bill Gletsos
25-06-2004, 12:57 AM
On the current ACF list 29 of our juniors are represented with !! ratings, only 6 adults. These are normal games Garvin
That does not mean they have played many games but just that based on results they are relaible.
It really would help if you knew what you were yapping about.

Cat
25-06-2004, 12:59 AM
Your saying it does not make it a fact. Many of your juniors had very few normally rated games.
The figures are from the ACF masterfile Bill, your figures not mine.



Your failure to understand is not my problem. You just continually choose to ignore the facts.
I keep telling you that given the nature of the pool generated by the Elo system with regards the ACT juniors etc no rating system could correct it in a timely manner if at all without manual correction.

What do you mean by timely? anyway, lets not descend into further argument.

Cat
25-06-2004, 01:02 AM
That does not mean they have played many games but just that based on results they are relaible.
It really would help if you knew what you were yapping about.

What is means is that they played far more games than our adults, as you well know. You're simply being disingenuous.

Bill Gletsos
25-06-2004, 01:10 AM
What is means is that they played far more games than our adults, as you well know. You're simply being disingenuous.
That is all assumption on your part based on the !! or !'s etc.
All I am pointing out is that your using them as a measure of how many games a player has played is incorrect.

Garvinator
25-06-2004, 01:13 AM
can someone put up some cold hard stats plz instead of this back and forth stuff.

Bill Gletsos
25-06-2004, 01:14 AM
The figures are from the ACF masterfile Bill, your figures not mine.
You are not quoting figures at all just using the symbols to make assumptions.


What do you mean by timely? anyway, lets not descend into further argument.
I have explained it at length numerous times before.
As usual you are just a waste of my time as this has been covered in numerous previous threads.
Your failure to comprehend it is not my problem.

As for descending into further argument then dont continue to post rubbish.

Bill Gletsos
25-06-2004, 01:36 AM
can someone put up some cold hard stats plz instead of this back and forth stuff.
Here is a David Richards quote from another thread:

Juniors on the Gold Coast generally compete in Junior competition far more often than they do in open competition - FACT

Adults are more likely to be playing another adult at the ADULT club - FACT.


Now virtually all junior comps from the GC are rapids. Hence they play far more rapid games than normal games.

Then there was this:


We have around 120 ACF rated players on the Gold Coast, and as I say last time I checked 79% were juniors. Most of these players play at home. Its virtually impossible for them to progress in the rating system unless their outstanding, are prepared to travel, or recieve the biannual rating boost.
To which I replied:


This seems contradictory.
So most stay home. Therefore most will have accuarte ratings relative to other pool members.
Those that play away (ourside the pool) will gain points from those away games if they are truly underrated.

As highlighted previously last year, the Gold Coast seems to have many junior only rapids but very few junior only tournaments at normal time controls that can be rated in the Normal rating system. Hence the juniors only progress slowly on the normal list because they have a lack of results in the normal system.

This is not a rating system problem but one due to tournament logistics.

They should play more normal rated games.

David chose to ignore replying to virtually all of this except the last line
They should play more normal rated games

to which he replied

Agreed

Alan Shore
25-06-2004, 02:03 AM
Now virtually all junior comps from the GC are rapids. Hence they play far more rapid games than normal games.


But this is irrelevant, you were saying it was the lack of normal rated games juniors played relative to adults, not relative as a ratio of normal:rapid. The issue I believe was part of your argument was there was insufficient data regarding juniors playing normal rated games.

However, what David has failed to take into account are those juniors who do not have !! ratings. It's these that would have the largest percentage of misrepresented ratings. Those with !! and are still underrated due to rapid chess improvement would be the primary concern - something which Glicko is good at fixing quickly, after Bill's refinement.

So it appears you're both just looking at the issue from different perspectives but I really don't see a problem - there were minor flaws and they're being attended to by Bill's slight modifications, so surely we don't have a problem here.

David, if you are so hung up on ratings, I suggest you have a look at the QLD Junior ratings. Now they are abysmal compared to the ACF glicko system, (or they were, I think they've had a little refinement). Back when I was a junior (years ago now, maybe 98 or 99) I wrote a few articles debunking it but you should see how playing a lot of games (i.e. Gold Coast players) automatically made your rating increase as to maintain your rating you were only expected to perform at 47% rather than 50%. And when a junior plays about 100 games in a rating period.. well, you can imagine why so many Brisbane players used to perform so far above their rating when playing at the GC. (One tournament I performed nearly 500 points above my rating! And this was not a one-off occurrence.)

Bill Gletsos
25-06-2004, 02:11 AM
But this is irrelevant, you were saying it was the lack of normal rated games juniors played relative to adults, not relative as a ratio of normal:rapid. The issue I believe was part of your argument was there was insufficient data regarding juniors playing normal rated games.
Not totally correct.
I'm saying that they play far more rated games than normal games and in general very few normal games.

Alan Shore
25-06-2004, 02:29 AM
Not totally correct.
I'm saying that they play far more rated games than normal games and in general very few normal games.

I assume you mean 'rapid' instead of 'rated' there. As for 'in general' well, if you can show that regardless of any rapid activity then your point is well founded. If not, well.. David could be on to something in that regard. I still however agree with you that it shouldn't be an issue.

Cat
25-06-2004, 03:00 PM
However, what David has failed to take into account are those juniors who do not have !! ratings. It's these that would have the largest percentage of misrepresented ratings. Those with !! and are still underrated due to rapid chess improvement would be the primary concern - something which Glicko is good at fixing quickly, after Bill's refinement.

Well 29/70 juniors (approx) have stable ratings & only 6/35 adults. To suggest the lack of junior participation in standard games as the cause of the problem is utterly mistaken. All rating systems will become more efficient if greater numbers of observations are made, but we have to deal with the real world and develop a system that properly represents dynamic change.


David, if you are so hung up on ratings, I suggest you have a look at the QLD Junior ratings. Now they are abysmal compared to the ACF glicko system, (or they were, I think they've had a little refinement). Back when I was a junior (years ago now, maybe 98 or 99) I wrote a few articles debunking it but you should see how playing a lot of games (i.e. Gold Coast players) automatically made your rating increase as to maintain your rating you were only expected to perform at 47% rather than 50%. And when a junior plays about 100 games in a rating period.. well, you can imagine why so many Brisbane players used to perform so far above their rating when playing at the GC. (One tournament I performed nearly 500 points above my rating! And this was not a one-off occurrence.)]

Personally I think introducing juniors into a rating system early is counterproductive, even pernicious. Ratings have so little meaning in real terms to early developers that it is really pointless to use this kind of system. Yet for many juniors it becomes a point of focus and disappointment, and cannot be helpful to development.

It would be better for chess to become more closely aligned with the general educational curriculum and take a leaf out of the way music is developed, with a series of gradings which are seen as milestones of development. It would make much more sense given current understanding of learning behaviour.

Bill Gletsos
25-06-2004, 04:04 PM
Well 29/70 juniors (approx) have stable ratings & only 6/35 adults. To suggest the lack of junior participation in standard games as the cause of the problem is utterly mistaken. All rating systems will become more efficient if greater numbers of observations are made, but we have to deal with the real world and develop a system that properly represents dynamic change.
You cannot determine if a player is improving either slowly, steadily or rapidly if you have an insufficient number of results to work on. To suggest otherwise just demonstrates your lack of understanding of the issues involved.
Of course this has already been pointed out to you before on the old ACF BB.

Cat
25-06-2004, 06:11 PM
You cannot determine if a player is improving either slowly, steadily or rapidly if you have an insufficient number of results to work on. To suggest otherwise just demonstrates your lack of understanding of the issues involved.
Of course this has already been pointed out to you before on the old ACF BB.

One can do anything if one has the imagination, intellegence and determination to find a solution. 'Volatility' is one attempt to accomodate such a solution. The use of volatility is an aknowledgement of this reality.

arosar
25-06-2004, 06:21 PM
You know what would really be cool? It's if youse blokes actually started banging out your fancy multi-line equations! Ain't that how you mathematical types normally debate issues?

AR

Bill Gletsos
25-06-2004, 06:51 PM
One can do anything if one has the imagination, intellegence and determination to find a solution. 'Volatility' is one attempt to accomodate such a solution. The use of volatility is an aknowledgement of this reality.
All you are doing is demonstrating your complete lack of understanding of the issues.
If you have insufficient results to work with you cannot possibly determine if the result is an actual improvement or just a fluke result.

Cat
25-06-2004, 06:54 PM
All you are doing is demonstrating your complete lack of understanding of the issues.
If you have insufficient results to work with you cannot possibly determine if the result is an actual improvement or just a fluke result.

Stop complaining, you don't have insufficient results simply a blunt tool. Much more has been done with a lot less

Bill Gletsos
25-06-2004, 06:58 PM
Stop complaining, you don't have insufficient results simply a blunt tool. Much more has been done with a lot less
As usual you are just a complete waste of my time.

Bill Gletsos
25-06-2004, 08:48 PM
Here is the data for the GC
It shows the number of normal games and number of rapid games played in the last 4 rating periods and whether they are juniors ,adults or unknowns.

For supposedly rapidly improving players then anything less than 30 games in a 12 month period is pretty useless in determining if the players really are improving or just being lucky.


normal rapid
7 29 Junior
41 45 Junior
38 68 Junior
15 29 Junior
50 71 Junior
0 0 Junior
0 27 Junior
0 25 Junior
43 41 Junior
0 16 Junior
36 55 Junior
27 48 Junior
13 30 Junior
7 45 Junior
26 54 Junior
36 67 Junior
0 37 Junior
15 23 Junior
56 28 Junior
0 35 Junior
59 49 Junior
94 33 Junior
0 20 Junior
35 45 Junior
7 14 Junior
15 32 Junior
0 28 Junior
0 8 Junior
11 42 Junior
34 29 Junior
0 18 Junior
36 30 Junior
19 0 Junior
0 0 Junior
6 0 Junior
34 43 Junior
13 51 Junior
6 6 Junior
12 32 Junior
42 50 Junior
29 57 Junior
6 8 Junior
0 0 Junior
7 21 Junior
49 69 Junior
0 26 Junior
51 15 Junior
0 6 Junior
60 57 Junior
0 6 Junior
19 43 Junior
14 25 Junior
0 22 Junior
0 15 Junior
26 30 Junior
6 7 Junior
30 45 Junior
0 18 Junior
0 44 Junior
28 45 Junior
0 12 Junior
9 16 Junior
8 6 Junior
0 11 Junior
37 62 Junior
5 1 Adult
13 0 Adult
0 0 Adult
17 15 Adult
0 7 Adult
4 0 Adult
49 27 Adult
14 7 Adult
24 13 Adult
0 0 Adult
0 0 Adult
106 11 Adult
0 0 Adult
26 0 Adult
37 7 Adult
75 21 Adult
2 17 Adult
32 29 Adult
4 8 Adult
0 0 Adult
21 22 Adult
0 0 Adult
0 0 Adult
31 21 Adult
11 0 Adult
7 1 Adult
0 7 Adult
0 0 Adult
5 0 Adult
0 0 Adult
8 6 Adult
12 11 Unknown
16 9 Unknown
24 0 Unknown
0 0 Unknown
6 9 Unknown
18 21 Unknown
20 0 Unknown
0 0 Unknown
0 0 Unknown
24 47 Unknown
7 0 Unknown
6 9 Unknown
50 9 Unknown


By comparision most NSW and ACT juniors on the normal list would play far more than 30 games in a 12 month period. In fact many would play that in a single rating period.

So David just go away and stop continually wasting my time. :hand:

Cat
25-06-2004, 09:54 PM
Here is the data for the GC
It shows the number of normal games and number of rapid games played in the last 4 rating periods and whether they are juniors ,adults or unknowns.

For supposedly rapidly improving players then anything less than 30 games in a 12 month period is pretty useless in determining if the players really are improving or just being lucky.


normal rapid
7 29 Junior
41 45 Junior
38 68 Junior
15 29 Junior
50 71 Junior
0 0 Junior
0 27 Junior
0 25 Junior
43 41 Junior
0 16 Junior
36 55 Junior
27 48 Junior
13 30 Junior
7 45 Junior
26 54 Junior
36 67 Junior
0 37 Junior
15 23 Junior
56 28 Junior
0 35 Junior
59 49 Junior
94 33 Junior
0 20 Junior
35 45 Junior
7 14 Junior
15 32 Junior
0 28 Junior
0 8 Junior
11 42 Junior
34 29 Junior
0 18 Junior
36 30 Junior
19 0 Junior
0 0 Junior
6 0 Junior
34 43 Junior
13 51 Junior
6 6 Junior
12 32 Junior
42 50 Junior
29 57 Junior
6 8 Junior
0 0 Junior
7 21 Junior
49 69 Junior
0 26 Junior
51 15 Junior
0 6 Junior
60 57 Junior
0 6 Junior
19 43 Junior
14 25 Junior
0 22 Junior
0 15 Junior
26 30 Junior
6 7 Junior
30 45 Junior
0 18 Junior
0 44 Junior
28 45 Junior
0 12 Junior
9 16 Junior
8 6 Junior
0 11 Junior
37 62 Junior
5 1 Adult
13 0 Adult
0 0 Adult
17 15 Adult
0 7 Adult
4 0 Adult
49 27 Adult
14 7 Adult
24 13 Adult
0 0 Adult
0 0 Adult
106 11 Adult
0 0 Adult
26 0 Adult
37 7 Adult
75 21 Adult
2 17 Adult
32 29 Adult
4 8 Adult
0 0 Adult
21 22 Adult
0 0 Adult
0 0 Adult
31 21 Adult
11 0 Adult
7 1 Adult
0 7 Adult
0 0 Adult
5 0 Adult
0 0 Adult
8 6 Adult
12 11 Unknown
16 9 Unknown
24 0 Unknown
0 0 Unknown
6 9 Unknown
18 21 Unknown
20 0 Unknown
0 0 Unknown
0 0 Unknown
24 47 Unknown
7 0 Unknown
6 9 Unknown
50 9 Unknown


By comparision most NSW and ACT juniors on the normal list would play far more than 30 games in a 12 month period. In fact many would play that in a single rating period.

So David just go away and stop continually wasting my time. :hand:

Bill you have listed many juniors who have played between 20-50 rated games over the last 12/12, quite a reasonable number of games. The situation in NSW & Canberra is different for geographical reasons. It's much easier to travel to larger conurbations, the demographics are very different. In SE Queensland we have a sparce population spread out over a large geographical area. Getting people together is a harder task. We don't have the player populations to run multiple leagues like you suggest, even though the organisers do a great job.

Are you running a rating system for Australia or NSW? Just because things aren't like they are in Sydney, it doesn't mean it doesn't count. You seem to have an extremely Sydney-centric view of Australia. If you don't have respect or regard for the difficulties of other regions, where is the representation?

It is unaffordable for most of the parents of these children to pay the costs required in travelling to increase their children's participation. The children have other interest and commitments in school, etc, and can only further their chess though great sacrifice.

What is it all about, about the preservation of an abstract rating system, or getting kids and people involved in playing chess? It is so difficult for some of these kids to progress, the rating system is so punitive, it is little wonder they give up.

So why are the ratings so important to these kids? Because in an environment where kids are submerged in a culture of development, where they progress through musical grades, through higher levels of speech & drama, school achievement, etc, they quickly learn that their chess is on a road to nowhere. When they are rewarded to play instruments, for debating, etc, recieve achievement certificates for sport, all they have to measure their chess achievement is their bogged-down rating.

Now this is not important to the kids, because they've got other things to do - they're happy. But it is important to Australian chess, because chess looses out everytime.

If chess is to flourish in Australia, it has to find ways to compete. There are so many interests out there today, that chess will become squeezed more and more. It must become kid-friendly, encouraging, rewarding. The rating system is punitive and it is irrelevant to the needs of the majority of children. Yet it has been part and parcel of playing chess. You cannot walk within chess corridors without experiencing its weight. What other endeavour gives kids a label they have to trudge around with?

Finally, I'd like to know which of our adults played all these games. I can think of only 5 of our adults who have played more than 20 rated games, and another who really lives in Brisbane, but competes regularly on the coast. Maybe you're including everyone who has played at the Chess Centre?

Bill Gletsos
25-06-2004, 10:25 PM
Bill you have listed many juniors who have played between 20-50 rated games over the last 12/12, quite a reasonable number of games. The situation in NSW & Canberra is different for geographical reasons. It's much easier to travel to larger conurbations, the demographics are very different. In SE Queensland we have a sparce population spread out over a large geographical area. Getting people together is a harder task. We don't have the player populations to run multiple leagues like you suggest, even though the organisers do a great job.

Are you running a rating system for Australia or NSW? Just because things aren't like they are in Sydney, it doesn't mean it doesn't count. You seem to have an extremely Sydney-centric view of Australia. If you don't have respect or regard for the difficulties of other regions, where is the representation?

It is unaffordable for most of the parents of these children to pay the costs required in travelling to increase their children's participation. The children have other interest and commitments in school, etc, and can only further their chess though great sacrifice.

What is it all about, about the preservation of an abstract rating system, or getting kids and people involved in playing chess? It is so difficult for some of these kids to progress, the rating system is so punitive, it is little wonder they give up.

So why are the ratings so important to these kids? Because in an environment where kids are submerged in a culture of development, where they progress through musical grades, through higher levels of speech & drama, school achievement, etc, they quickly learn that their chess is on a road to nowhere. When they are rewarded to play instruments, for debating, etc, recieve achievement certificates for sport, all they have to measure their chess achievement is their bogged-down rating.

Now this is not important to the kids, because they've got other things to do - they're happy. But it is important to Australian chess, because chess looses out everytime.

If chess is to flourish in Australia, it has to find ways to compete. There are so many interests out there today, that chess will become squeezed more and more. It must become kid-friendly, encouraging, rewarding. The rating system is punitive and it is irrelevant to the needs of the majority of children. Yet it has been part and parcel of playing chess. You cannot walk within chess corridors without experiencing its weight. What other endeavour gives kids a label they have to trudge around with?
This has nothing to do with a NSW or any other form of centric thinking.
It has to do with your usual misrepresenttation of the facts.

If you want your juniors to have more accuarte normal ratings they need to play more normal rated games.
Even a moron like you should be able to see that.

I said they need to play more normal games.
A very significant number of them play less than 30 games in a 12 month period.

The GC runs many junior tournamennts it is just that most of them are rapids.

You keep making criticising the rating system for failures in your clubs tournament policy.
Why dont you try stopping complaining for once and ttake a leaf out of the ACT book and have juniors play more normal rated games.


Finally, I'd like to know which of our adults played all these games. I can think of only 5 of our adults who have played more than 20 rated games, and another who really lives in Brisbane, but competes regularly on the coast. Maybe you're including everyone who has played at the Chess Centre?
Nice try at misdirection but all players listed are members of the GC club as provided to me a committee member of the GC club.

Anyway I have more important things to do than waste my time on the likes of you. :hand: :whistle: :whistle:

Bill Gletsos
25-06-2004, 11:41 PM
In fact if anyone takes a close look at the figures they will see your arguement is cretinous in the extreme.

Of the 65 juniors 3 played no games of either normal or rapid.
Of the remainong 62 players 43 played less tha 30 games.
Of this group of 30 players 18 played 0 normal games, another 10 played less than 10 normal games, another 10 played less than 20 and the remaining 5 played 21-29 games.

And you have the stupidity to claim their rating problems are the fault of the system and not their lack of games.

Lets just face it David, when it comes to ratings and facts you are a complete joke.

frogmogdog
26-06-2004, 10:59 AM
you guys thrive on arguments. :)

bill, i've got a question.

regarding the inactive players' ratings - have you considered after 20 or 25 years of inactivity just dropping players off the list and making them officially unrated?

the lure of "easy" unrated prizes for one rating period might even encourage some to return to chess, and this is only a stunt that could be pulled once or twice in a lifetime.

(PS B.D. -- i think you're gnawing on some old grievance about the qld junior ratings. at least since 2000 that system has been good with no huge differential between briz and GC.)

Oepty
26-06-2004, 02:21 PM
Bill. This is an idea I haven't thought much about but I was wondering whether it might be worth droping the minimum time control for all junior events to be included on the normal list. I my experience that most of the games in all junior competition are finished in way less than the time control. It then becomes a problem of how to control 80-90% of the field while the last few games are finished. I can understand why a club running an all junior competition would want to run it as rapid.
Scott

Bill Gletsos
26-06-2004, 03:56 PM
Bill. This is an idea I haven't thought much about but I was wondering whether it might be worth droping the minimum time control for all junior events to be included on the normal list.
In my opinion not an option.
For the normal list you want to know how good the juniors are at normal time controls not rapid time controls. You have the rapid list for that.

Bill Gletsos
26-06-2004, 04:05 PM
you guys thrive on arguments. :)

bill, i've got a question.

regarding the inactive players' ratings - have you considered after 20 or 25 years of inactivity just dropping players off the list and making them officially unrated?
There was major uproar from around 1992 to April 1996 when the National Ratings Officer of the time removed over 8,000 players from the April 1993 list who he deemed inactive.

At the ACF's request we reinstated them in December 1997.

ursogr8
26-06-2004, 04:36 PM
Bill

A and B and probably important, but not time-urgent.
In the interim, why not just press the button?

starter

Bill, Bill, Bill

Quick, press the button mate.

I hear that jeffrei is considering forming a new Club based on the under-rated juniors. This is the end of civilisation as we know it. They will scoop the pool at Interclub next month.

Mate, mate. Re-rate them quick. Please.

chicken-little (aka starter)

Bill Gletsos
26-06-2004, 04:48 PM
Bill, Bill, Bill

Quick, press the button mate.

I hear that jeffrei is considering forming a new Club based on the under-rated juniors. This is the end of civilisation as we know it. They will scoop the pool at Interclub next month.

Mate, mate. Re-rate them quick. Please.

chicken-little (aka starter)
Sorry starter, but it isnt going to happen just yet.

Kevin Bonham
27-06-2004, 12:02 AM
regarding the inactive players' ratings - have you considered after 20 or 25 years of inactivity just dropping players off the list and making them officially unrated?

Glicko sort-of does this by placing so much emphasis on their new games. Probably the most contentious cases are when someone returns from hibernation and plays only one or two games. If they play a decent number they'll get a rating much closer to their PR than their old rating.


the lure of "easy" unrated prizes for one rating period might even encourage some to return to chess, and this is only a stunt that could be pulled once or twice in a lifetime.

How would that work - if they're going to be unrated then they'll have to get a new rating first to qualify for ratings prizes at all. I guess if their first event back is very rusty they might be a ratings prize threat (of sorts) next time.

Garvinator
27-06-2004, 06:30 AM
How would that work - if they're going to be unrated then they'll have to get a new rating first to qualify for ratings prizes at all. I guess if their first event back is very rusty they might be a ratings prize threat (of sorts) next time.
up here in qld, some tournaments have an unrated section for those players who dont have a rating.

Cat
01-07-2004, 10:09 PM
This has nothing to do with a NSW or any other form of centric thinking.
It has to do with your usual misrepresenttation of the facts.

If you want your juniors to have more accuarte normal ratings they need to play more normal rated games.
Even a moron like you should be able to see that.

I said they need to play more normal games.
A very significant number of them play less than 30 games in a 12 month period.

The GC runs many junior tournamennts it is just that most of them are rapids.

You keep making criticising the rating system for failures in your clubs tournament policy.
Why dont you try stopping complaining for once and ttake a leaf out of the ACT book and have juniors play more normal rated games.



Anyway I have more important things to do than waste my time on the likes of you. :hand: :whistle: :whistle:

Bill, you have absolutely no grasp of the problem. Your thinking becomes increasingly blinkered by the day. These are the problems facing Australia Chess;

Australian Chess is a minority sport struggling to achieve recognition, status and consequently funding. Part of the problem is undoubtedly a lack of tradition in this country. Chess is arriving here at the same time as the internet, DVD and The Simpsons, which greatly compounds the problem of increasing the profile of Chess. In a sunny country, a climate more suitable to beach sports than indoor pursuits, Chess faces an uphill battle.

However, there are some positives, both in what has been achieved in Australia and from experiences overseas. Chess has developed a formidable reputation as an educational tool. In part this comes from a number of small studies demonstrating possible educational benefits, in part from the incorporation of Chess in a number of American States and European Countries, also from the success of the junior coaching programs in Australia and finally reinforced by the perceptions of mothers who have involved their children in chess programs.

This offers the best prospect for future growth in Australia. Parental 'word-of-mouth' is a very powerful tool, and mothers & women in Australia are major consumers. Mums are not keen to see their kiddie-winkles being beaten up on the football field and are enthusiastic about finding 'wholesome' and safe pursuits for their charges. This may be contributing to the rise in popularity of soccer in Australia, which is seen as a less physical sport.

Parents are paying out greater & greater fees towards the education of their children. Education is being seen as an absolute necessity for the next generation - especially by their parents. Anything that is seen as beneficial to educational attainment is being embraced with open arms and virtually has a blank cheque.

To realise it's potential in Australia, Chess must be competitive, but more than anything else it must take every opportunity to integrate with the educational system -that's where the money is and where the prospects are brightest. It must feed from its competitive edge, but to do so requires a structure that can be integrated easily.

The problem with the rating system, the existing structure if you like, is that outside of Sydney (especially) children are handicapped. Even within Sydney there is no progressive structure paralllel to the education system, at which students can be guaranteed significant progress. The rating system is harsh and punative, and is essentially hopeless as an educational tool.

Bill, the children on the Gold Coast don't need chess, chess needs them. If chess is to enjoy a rhobust future in Australia it simply cannot afford to spurn it's opportunities to encourage juniors to develop a longer term affiliation with the sport. It is totally naive to believe that parents can continually fork-out to send their kids all over the country in order to achieve some improvement in their rating. And it doesn't matter how many games they play locally in an essentially closed rating pool, if they can't get games against stronger opposition from Sydney, etc to inject points into the pool

Previously I have argued that these goals could be achieved through a more junior responsive rating system allowing for development and maturation, which was easily achievable but there was an obvious lack of political foresight to engage these options whole-heartedly. But to me it makes much more sense to try to adapt our structure to the prevailing educational structures that exist - in essence a system of gradings with a national accreditation body. In this way a child from Gladstone (say) has essentially the same chances of success as a child from Sydney. Obviously, a lot will still depend on the available coaches, etc - but at least they're not being knee-capped by the system.

PHAT
01-07-2004, 10:41 PM
To realise it's potential in Australia, Chess must be competitive, but more than anything else it must take every opportunity to integrate with the educational system...

I must share this now. In NSW, high schools offen specialise in IT or the arts or sport et al. In the region south of metro Sydney, there is the Illawarra Sports High School. In resent weeks I have been teaching chess there to a/the class of "uncontrolables." There are ~15 in the class but only 3 to 6 ever there bacause the rest are on long/medium/short term suspension. I am pleased to say that last Monday, that as of next term, chess is an official sport at that school.

The sad bit is that the best player was expelled on Tuesday. I will miss him - I was the only teacher he would talk to sensibly.

Cat
01-07-2004, 10:51 PM
I must share this now. In NSW, high schools offen specialise in IT or the arts or sport et al. In the region south of metro Sydney, there is the Illawarra Sports High School. In resent weeks I have been teaching chess there to a/the class of "uncontrolables." There are ~15 in the class but only 3 to 6 ever there bacause the rest are on long/medium/short term suspension. I am pleased to say that last Monday, that as of next term, chess is an official sport at that school.

The sad bit is that the best player was expelled on Tuesday. I will miss him - I was the only teacher he would talk to sensibly.


Good luck Matt, where do you get the time?

PHAT
01-07-2004, 11:12 PM
Bill, you have absolutely no grasp of the problem. Your thinking becomes increasingly blinkered by the day...

...The rating system is harsh and punative, and is essentially hopeless as an educational tool.

...It is totally naive to believe that parents can continually fork-out to send their kids all over the country in order to achieve some improvement in their rating. And it doesn't matter how many games they play locally in an essentially closed rating pool, if they can't get games against stronger opposition from Sydney, etc to inject points into the pool


Here is The Billdo's total probelm. He ISan idiot, first class.

1. He thinks it is necessary to bring the isolated Australian rating pool into parity/synch with Europe by using a FIDE:ACF ratio. So, he does it.

2. He thinks that the same isolated pool phenominae don't never gonna happen with the heterogenous distribution of junior participation, because Glicko will handle it.

:wall:

There are none so blind as those who won't see.


The Glicko "works" by assuming that the rate and direction of improvement/deterioration in players strength is symetrical. Buit it ain't. There is a flow of players. They come in at a low rating, and leave at a higher rating. Thus, the cohort of new (junior) players are chronically under rated.

This is not rocket science.

David, more power to you - but how much energy does it take to move the immovable object.

PHAT
01-07-2004, 11:14 PM
Good luck Matt, where do you get the time?

I stopped caring about money.

Bill Gletsos
01-07-2004, 11:22 PM
Here is The Billdo's total probelm. He ISan idiot, first class.
The only idiots are you and your mate dumb and dumber.


1. He thinks it is necessary to bring the isolated Australian rating pool into parity/synch with Europe by using a FIDE:ACF ratio. So, he does it.
This isnt necessary its just desirable.


2. He thinks that the same isolated pool phenominae don't never gonna happen with the heterogenous distribution of junior participation, because Glicko will handle it.
Never said that.
What I said is that it is extremely difficult to automatically determine what and what does not constitute a pool.
If you cannot identify it you cannot handle it.


There are none so blind as those who won't see.
Sounds like you are talking about yourself and the doc.



The Glicko "works" by assuming that the rate and direction of improvement/deterioration in players strength is symetrical.
Provide a reference for this statement.



Buit it ain't. There is a flow of players. They come in at a low rating, and leave at a higher rating.
All ratings theory acknowledges this.
It has been known for years.


Thus, the cohort of new (junior) players are chronically under rated.
Rubbish. This does not automatically follow from the above.



This is not rocket science.
No but like ratings theory you know nothing about it.


David, more power to you - but how much energy does it take to move the immovable object.
More than you have got.

Bill Gletsos
01-07-2004, 11:25 PM
Bill, you have absolutely no grasp of the problem. Your thinking becomes increasingly blinkered by the day. These are the problems facing Australia Chess;

Australian Chess is a minority sport struggling to achieve recognition, status and consequently funding. Part of the problem is undoubtedly a lack of tradition in this country. Chess is arriving here at the same time as the internet, DVD and The Simpsons, which greatly compounds the problem of increasing the profile of Chess. In a sunny country, a climate more suitable to beach sports than indoor pursuits, Chess faces an uphill battle.

However, there are some positives, both in what has been achieved in Australia and from experiences overseas. Chess has developed a formidable reputation as an educational tool. In part this comes from a number of small studies demonstrating possible educational benefits, in part from the incorporation of Chess in a number of American States and European Countries, also from the success of the junior coaching programs in Australia and finally reinforced by the perceptions of mothers who have involved their children in chess programs.

This offers the best prospect for future growth in Australia. Parental 'word-of-mouth' is a very powerful tool, and mothers & women in Australia are major consumers. Mums are not keen to see their kiddie-winkles being beaten up on the football field and are enthusiastic about finding 'wholesome' and safe pursuits for their charges. This may be contributing to the rise in popularity of soccer in Australia, which is seen as a less physical sport.

Parents are paying out greater & greater fees towards the education of their children. Education is being seen as an absolute necessity for the next generation - especially by their parents. Anything that is seen as beneficial to educational attainment is being embraced with open arms and virtually has a blank cheque.

To realise it's potential in Australia, Chess must be competitive, but more than anything else it must take every opportunity to integrate with the educational system -that's where the money is and where the prospects are brightest. It must feed from its competitive edge, but to do so requires a structure that can be integrated easily.

The problem with the rating system, the existing structure if you like, is that outside of Sydney (especially) children are handicapped. Even within Sydney there is no progressive structure paralllel to the education system, at which students can be guaranteed significant progress. The rating system is harsh and punative, and is essentially hopeless as an educational tool.

Bill, the children on the Gold Coast don't need chess, chess needs them. If chess is to enjoy a rhobust future in Australia it simply cannot afford to spurn it's opportunities to encourage juniors to develop a longer term affiliation with the sport. It is totally naive to believe that parents can continually fork-out to send their kids all over the country in order to achieve some improvement in their rating. And it doesn't matter how many games they play locally in an essentially closed rating pool, if they can't get games against stronger opposition from Sydney, etc to inject points into the pool

Previously I have argued that these goals could be achieved through a more junior responsive rating system allowing for development and maturation, which was easily achievable but there was an obvious lack of political foresight to engage these options whole-heartedly. But to me it makes much more sense to try to adapt our structure to the prevailing educational structures that exist - in essence a system of gradings with a national accreditation body. In this way a child from Gladstone (say) has essentially the same chances of success as a child from Sydney. Obviously, a lot will still depend on the available coaches, etc - but at least they're not being knee-capped by the system.
The aim of the ratings sytem is to provide a indication of the relative strengths of players based on their results and not based on their age, education, ego or some airy fairy feel good factor.

Stop wasting my time.

Cat
01-07-2004, 11:35 PM
Here is The Billdo's total probelm. He ISan idiot, first class.

1. He thinks it is necessary to bring the isolated Australian rating pool into parity/synch with Europe by using a FIDE:ACF ratio. So, he does it.

2. He thinks that the same isolated pool phenominae don't never gonna happen with the heterogenous distribution of junior participation, because Glicko will handle it.

:wall:

There are none so blind as those who won't see.


The Glicko "works" by assuming that the rate and direction of improvement/deterioration in players strength is symetrical. Buit it ain't. There is a flow of players. They come in at a low rating, and leave at a higher rating. Thus, the cohort of new (junior) players are chronically under rated.

This is not rocket science.

David, more power to you - but how much energy does it take to move the immovable object.


Thanks, Matt, sometimes I feel I keep saying the bleedin' obvious. I guess I was perturbed by the way Bill foolishly dismissed constructive crticism last year and I felt this was not good. I followed on from what Ian had said. It was shocking to see how an alternative viewpoint could be so summarily dismissed and how Bill was seemingly able to establish an autocratic dogma that individuals failed to challenge for fear of ridicule. I thought this could be my small contribution to Australia Chess and my colleagues on the Gold Coast- to make sure their interests were served.

I was also affected by the problems of a Queensland child who suffers from physical handicap, is very dedicated to his chess, but was finding it difficult to improve his rating due to the lack of local competition. It was very disheartening for him & his mum, and seemed so unnecessary.

I guess I'll continue to drop in from time to time.

Cat
01-07-2004, 11:43 PM
Provide a reference for this statement.

It was described by Glickman himself. His papers on on his web-site, you'll find the reference there




This does not automatically follow from the above.


That's true, if players play a large number of games throughout the entire pool, like essentially on the internet, then its may not flow. Thats precisely what Matt & I have been saying. Unfortunately that is not Australia.

Bill Gletsos
01-07-2004, 11:53 PM
Thanks, Matt, sometimes I feel I keep saying the bleedin' obvious. I guess I was perturbed by the way Bill foolishly dismissed constructive crticism last year and I felt this was not good. I followed on from what Ian had said. It was shocking to see how an alternative viewpoint could be so summarily dismissed and how Bill was seemingly able to establish an autocratic dogma that individuals failed to challenge for fear of ridicule.
If I recall both Graham Saint and Kevin also criticisied your ideas in the ACF bulletin at the time.
I would just say that Graham, Ian and I were involved in vigourous debate.
If others were afraid of ridicule then perhaps it was due to what they were saying being ridiculous. After all if what they were saying was true how could they be ridiculed.


I thought this could be my small contribution to Australia Chess and my colleagues on the Gold Coast- to make sure their interests were served.
Personally I found it better served by my discussions with Sue Kimura.


I was also affected by the problems of a Queensland child who suffers from physical handicap, is very dedicated to his chess, but was finding it difficult to improve his rating due to the lack of local competition.
If you are talking about who I suspect you are I have spoken personally to his mum long before your involvement in the ratings debate.

Bill Gletsos
02-07-2004, 12:04 AM
It was described by Glickman himself. His papers on on his web-site, you'll find the reference there
Given the fact you cannot seem to understand the difference between RD and volatility and the way they interact I'm not about to take your word for anything.
Provide the exact document and page reference.


That's true, if players play a large number of games throughout the entire pool, like essentially on the internet, then its may not flow. Thats precisely what Matt & I have been saying. Unfortunately that is not Australia.
Nice attempt to misrepresent what I said.
I'm just saying your assumptions and Matt's do not follow from Glickman's papers.

Ratings pools occur in all ratings systems.
They certainly occur in the FIDE list.
The point is that it is extremely difficult to automatically identify pools.

Show me an existing National rating system that attempts to handle pools and I'm happy to look at it.

PHAT
02-07-2004, 12:21 AM
Ratings pools occur in all ratings systems... The point is that it is extremely difficult to automatically identify pools.


YOU find them difficult to identify them, we do not. :rolleyes:

Bill Gletsos
02-07-2004, 12:25 AM
YOU find them difficult to identify them, we do not. :rolleyes:
You wouldnt have a clue.

Bill Gletsos
02-07-2004, 01:13 AM
Instead of wasting time on the BB and apparently doing nothing else perhaps you could assist John Mazzieri and provide him with the details of the U1600 St.George V the Gong match.
As I understand it he has been chasing you for them but you are "unreliable".
The NSWCA Council could have told him that :whistle:

Kevin Bonham
02-07-2004, 02:40 AM
Here is The Billdo's total probelm. He ISan idiot, first class.

1. He thinks it is necessary to bring the isolated Australian rating pool into parity/synch with Europe by using a FIDE:ACF ratio. So, he does it.

Even a third-class idiot could have paid enough attention to remember the numerous times this point has been refuted before. The divergence between ACF and FIDE ratings has nix to do with isolation, Europe, larger pools etc - it mostly has to do with FIDE's ratings system being inferior and inflationary. :wall:

PHAT
02-07-2004, 07:49 AM
Even a third-class idiot could have paid enough attention to remember the numerous times this point has been refuted before. The divergence between ACF and FIDE ratings has nix to do with isolation, Europe, larger pools etc - it mostly has to do with FIDE's ratings system being inferior and inflationary. :wall:

And it seems that you cannot make it as far as up the scale as idiot, moron, or imbecile. It matters not one whit why the divergence occurs, be it geographic isolation, interplay isilation or rating system isolation. There is a diververgence of number values, and so, BG corrected them for FIDE/ACF. Now he won't do it for improvers/stables.

PHAT
02-07-2004, 07:53 AM
Instead of wasting time on the BB and apparently doing nothing else perhaps you could assist John Mazzieri and provide him with the details of the U1600 St.George V the Gong match.
As I understand it he has been chasing you for them but you are "unreliable".
The NSWCA Council could have told him that :whistle:

For you info, CLOWN, both team cards were accidently picked up by the other St George captain and will be posted. So up yours.

Bill Gletsos
02-07-2004, 10:10 AM
For you info, CLOWN, both team cards were accidently picked up by the other St George captain and will be posted. So up yours.
Then perhaps you moron you should have informed him of that.

rob
02-07-2004, 02:54 PM
Bill, you regularly state that you are wasting your time explaining and repeating yourself to DR & MS, yet you continue. Don't you have more useful stuff to do?

BTW which one is dumb & which one dumber?

Perhaps we could have a pole as to which one we believe best understands how ACF ratings work.

Bill Gletsos
02-07-2004, 03:04 PM
Bill, you regularly state that you are wasting your time explaining and repeating yourself to DR & MS, yet you continue. Don't you have more useful stuff to do?
Yes, I certainly do.
The problem is that if I dont answer these jokers, people who are new who have not read all the previous threads here and on the old BB maybe inclined to think what they say has any merit.


BTW which one is dumb & which one dumber?
I'm not really sure.


Perhaps we could have a pole as to which one we believe best understands how ACF ratings work.
You mean in general or just between the two jokers.

PHAT
02-07-2004, 03:06 PM
Then perhaps you moron you should have informed him of that.

Actually you drongo, we have already spoken. So up yours, again.

Bill Gletsos
02-07-2004, 03:08 PM
Actually you drongo, we have already spoken. So up yours, again.
Well you clearly hadnt when he was complaining about not knowing the results, otherwise he wouldnt have been complaining to Steve about not being able to provide him with the results.

Given a choice of whom to believe, John gets my vote.

PHAT
02-07-2004, 03:20 PM
Yes, I certainly do.
The problem is that if I dont answer these jokers, people who are new who have not read all the previous threads here and on the old BB maybe inclined to think what they say has any merit.


If you had the ability, you might have taken 10% of the time you spend rebutting, on writing a chess players guide to Glicko. But you haven't got the ability to explain anything and you haven't got a clue how to model real world phenominae. As a researcher you are an imposter. Your skill is that of a technician. And like a plumber, you make a lot of money out of joining A and B, and having no understanding the architect designed it that way. You use Glicko as a tool, your only tool. Go and do a TAFE course on ... oh, hang on, they do technical training, nothing that requires broad knowledge and imagination.

And in case you forgot what it was that I wanted to say, I repeat; If you had the ability, you might have taken 10% of the time you spend rebutting, on writing a chess players guide to Glicko.

rob
02-07-2004, 03:20 PM
You mean in general or just between the two jokers.
I meant just the two jokers (a close well balanced match).

Cat
02-07-2004, 03:34 PM
Bill, you regularly state that you are wasting your time explaining and repeating yourself to DR & MS, yet you continue. Don't you have more useful stuff to do?

BTW which one is dumb & which one dumber?

Perhaps we could have a pole as to which one we believe best understands how ACF ratings work.

Rob, is it Bill that's dumb and you're the dumber one? I'm right, aren't I?

rob
02-07-2004, 03:45 PM
Rob, is it Bill that's dumb and you're the dumber one? I'm right, aren't I?
Of course you are.

I am in awe of your superier wisdom :)

Bill Gletsos
02-07-2004, 03:52 PM
If you had the ability, you might have taken 10% of the time you spend rebutting, on writing a chess players guide to Glicko. But you haven't got the ability to explain anything and you haven't got a clue how to model real world phenominae.
This is rich coming from a moron like you.



As a researcher you are an imposter. Your skill is that of a technician. And like a plumber, you make a lot of money out of joining A and B, and having no understanding the architect designed it that way.
That is were you are wrong. I'm a systems engineer. That means I hhave to know how processes work so as to advise customers the best way of implemnting solutions.
If I was just what you incorrectly believe, I wouldnt be paid the salary I am.



You use Glicko as a tool, your only tool. Go and do a TAFE course on ... oh, hang on, they do technical training, nothing that requires broad knowledge and imagination.
The only tool around here is you.
You have no clue about ratings theory.
It is clear you havent even read ELO's book let alone other articles.


And in case you forgot what it was that I wanted to say, I repeat; If you had the ability, you might have taken 10% of the time you spend rebutting, on writing a chess players guide to Glicko.
Why redo Glickman's work.
There are documents on his web site. With regards the Glicko system one is complicated and the other simple. like wise for the Glicko2 system.
People can read whichever one they choose.

Bill Gletsos
02-07-2004, 03:55 PM
Of course you are.

I am in awe of your superier wisdom :)
Careful Rob, he is silly enough to believe you are serious. ;)

arosar
02-07-2004, 03:56 PM
Man...we should just create an entire Forum dedicated solely to this ratings business. And youse can all abuse each other to your hearts' content.

AR

Bill Gletsos
02-07-2004, 04:03 PM
Man...we should just create an entire Forum dedicated solely to this ratings business. And youse can all abuse each other to your hearts' content.

AR
:lol:

Cat
02-07-2004, 04:15 PM
Of course you are.

I am in awe of your superier wisdom :)

Thanks Rob, but you need to develop a sense of self worth - no need to enter this dispute throwing self-inflicting insults around. If you keep working at it you might indeed find something intellegent to say! Don't take yourself into the gutter, develop a sense of confidence.

Bill Gletsos
02-07-2004, 04:24 PM
Thanks Rob, but you need to develop a sense of self worth - no need to enter this dispute throwing self-inflicting insults around. If you keep working at it you might indeed find something intellegent to say! Don't take yourself into the gutter, develop a sense of confidence.
It is obvious you dont know Rob.
He already has a sense of self worth, a sense of confidence and needless to say has already made an intelligent comment.

rob
02-07-2004, 06:32 PM
The aim of the ratings sytem is to provide a indication of the relative strengths of players based on their results and not based on their age, education, ego or some airy fairy feel good factor.

Stop wasting my time.

Yep I do understand what Bill means by relative strengths:

The ratings sytem does provide an indication of the strengths of my relatives: my dad, wife, brother-in-law, father-in-law all have ACF ratings :)

My self-confidence is rising, plus I've beaten 3 top players in the past 3 weeks: H Barber, A Haasse ,P Roza :)
But maybe they lost to me instead, where is that self-worth - someone call a doctor :(

Kevin Bonham
02-07-2004, 07:57 PM
And it seems that you cannot make it as far as up the scale as idiot, moron, or imbecile.

There are some ladders that only a fool would climb.

(ref previous Bill-Matt ladder climbing debate.)


It matters not one whit why the divergence occurs, be it geographic isolation, interplay isilation or rating system isolation. There is a diververgence of number values, and so, BG corrected them for FIDE/ACF. Now he won't do it for improvers/stables.

No predictively superior alternative that better identifies improvers from those who have simply had unusually good results has yet been identified. Your own attempt failed spectacularly.

Kevin Bonham
02-07-2004, 08:11 PM
Perhaps we could have a pole as to which one we believe best understands how ACF ratings work.

I think Matt understands the system slightly better than Dave but is much more prone to pretend that he doesn't, for the sake of trolling Bill.

Cat
03-07-2004, 04:44 AM
For supposedly rapidly improving players then anything less than 30 games in a 12 month period is pretty useless in determining if the players really are improving or just being lucky.

Of the 65 juniors 3 played no games of either normal or rapid.
Of the remainong 62 players 43 played less tha 30 games.
Of this group of 30 players 18 played 0 normal games, another 10 played less than 10 normal games, another 10 played less than 20 and the remaining 5 played 21-29 games.



OK Bill, so what you're saying is that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years, many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving. Now isn't that exactly what I've been saying, that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years, many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving because the Glicko system is too insensitive, too blunt an instrument.

The difference between us Bill, is that you think kids that play 25 games a year or 50 games in 2 years cannot expect to recieve a reasonably reliable rating, whereas I believe they have every right to believe their rating should be reasonably reliable, especially when you've been telling us for the last year how accurate the system is.

The other difference between us Bill, is that you believe to provide a rating system that can deliver this kind of technical sophistication is unobtainable, that the technical wizzardary is beyond mankind. Whereas I believe in an age when man has landed rockets on the moon and in the age of solid state engineering, that such a milestone is indeed within our grasp.

Cat
03-07-2004, 04:50 AM
I think Matt understands the system slightly better than Dave .

I wouldn't disagree with that KB, but I think we both understand the failings of the system better than either your good-self or Bill.

Rincewind
03-07-2004, 10:44 AM
I wouldn't disagree with that KB, but I think we both understand the failings of the system better than either your good-self or Bill.

That's because it is easier to criticise than be constructive.

The following post:


OK Bill, so what you're saying is that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years, many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving. Now isn't that exactly what I've been saying, that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years, many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving because the Glicko system is too insensitive, too blunt an instrument.

The difference between us Bill, is that you think kids that play 25 games a year or 50 games in 2 years cannot expect to recieve a reasonably reliable rating, whereas I believe they have every right to believe their rating should be reasonably reliable, especially when you've been telling us for the last year how accurate the system is.

The other difference between us Bill, is that you believe to provide a rating system that can deliver this kind of technical sophistication is unobtainable, that the technical wizzardary is beyond mankind. Whereas I believe in an age when man has landed rockets on the moon and in the age of solid state engineering, that such a milestone is indeed within our grasp.

THis would provide a useful exercise for 1st year philosphy tut - identify all the fallacies. I'll leave it as an exercise for the interested reader.

Bill Gletsos
03-07-2004, 05:07 PM
OK Bill, so what you're saying is that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years, many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving.
As usual you are speaking rubbish.
In fact you are totally misrepresenting the facts.

I said :

Of the 65 juniors 3 played no games of either normal or rapid.
Of the remaining 62 players 43 played less tha 30 games.
Of this group of 30 players 18 played 0 normal games, another 10 played less than 10 normal games, another 10 played less than 20 and the remaining 5 played 21-29 games.
I am criticising the fact that the majority of your juniors (45% of the active juniors) play very few normal rated games(less than 10). Therefore on those figures it is nigh impossible to accurately rate those juniors as you cannnnot determine if they are static, imroving slightly or rapidly improving or just plain lucky or unlucky.



Now isn't that exactly what I've been saying, that the average junior on the Gold Coast who plays around 25 games a year or 50 games in 2 years,
As shown above the majority of juniors play far less than 25 games.


many more than our adults, cannot expect to be reasonably rated by the Glicko system when they are rapidly improving because the Glicko system is too insensitive, too blunt an instrument.
No you moron.
What it means is that you have insufficient results to base any sort of judgemnet, whether the system is elo,glicko or reading tea leaves.
What you want is a statistically unsound, feel good load of crap to increase juniors ratings because well hey they are getting older so they must be rapidly improving.
As I keep saying you are a joke.



The difference between us Bill, is that you think kids that play 25 games a year or 50 games in 2 years cannot expect to recieve a reasonably reliable rating, whereas I believe they have every right to believe their rating should be reasonably reliable,
THis just shows your complete lack of understanding.
You cannnnot have reliable ratings with extremly small sets of results.



especially when you've been telling us for the last year how accurate the system is.
I have been telling you the system is accurate based on results. Not enough results means the ratings for those players are less accurate than for players with more results.
That said the system is still more accurate than other ratings systems.



The other difference between us Bill, is that you believe to provide a rating system that can deliver this kind of technical sophistication is unobtainable, that the technical wizzardary is beyond mankind.
What I believe is that it is extremely difficult to obtain and nigh impossible with insufficient data/results. Of course anyone with any brains would see that also.



Whereas I believe in an age when man has landed rockets on the moon and in the age of solid state engineering, that such a milestone is indeed within our grasp.
I suspect you believe in the tooth fairy also.

Bill Gletsos
03-07-2004, 05:11 PM
THis would provide a useful exercise for 1st year philosphy tut - identify all the fallacies. I'll leave it as an exercise for the interested reader.
Well said Barry.

skip to my lou
03-07-2004, 05:14 PM
David, what is your preferred rating system?

Kevin Bonham
04-07-2004, 04:59 AM
I wouldn't disagree with that KB, but I think we both understand the failings of the system better than either your good-self or Bill.

Not only do I know that isn't true, but furthermore I don't even think that you believe it.

ursogr8
04-07-2004, 08:27 AM
I wouldn't disagree with that KB, but I think we both understand the failings of the system better than either your good-self or Bill.

David, what is your preferred rating system?

starter

skip to my lou
04-07-2004, 08:32 AM
:pirate:

ursogr8
04-07-2004, 08:49 AM
:pirate:


K
I thought it was a good question of yours. :clap: :clap:
starter

Cat
05-07-2004, 08:46 AM
David, what is your preferred rating system?

I've already explained many times how the failings in the system have come about. If you read through previous posts you would understand my sentiments.

But essentially, there is no recoginition in the Glicko system that juniors, as a cohort (not all juniors) are improving as they mature. This is a dynamic population, whereas the adult population is static. Therefore the population mean for this cohort should increase with time, dependant on the rate of observed change.

This does not mean all juniors increase their ratings automatically, but simply greater dynamic change in an upward direction is given to the improvers, in effect skewing the RD. Juniors who do not improve in this way will not increase their rating.

If one thinks about learning, people learn and improve quickly, whereas they loose their understanding slowly. Skewing the RD in essence reflects this simple observation.

However, in my previous post on this thread, I suggested that juniors maybe better served by introducing a grading system similar to the music system - the ideas for this were mentioned in the thread.

Bill, I can't believe you are trying to defend your position. The majority of juniors on the Gold Coast are being poorly served by the system.

Cat
05-07-2004, 08:58 AM
As usual you are speaking rubbish.
In fact you are totally misrepresenting the facts.

I said :

I am criticising the fact that the majority of your juniors (45% of the active juniors) play very few normal rated games(less than 10). Therefore on those figures it is nigh impossible to accurately rate those juniors as you cannnnot determine if they are static, imroving slightly or rapidly improving or just plain lucky or unlucky.



Bill, again you are simply distorting the statistics. When you say 'active' you mean they are on the ACF active list. However, many of those individuals have not played any 'normal' games in the last 12 months, for whatever reason. If one looks at those that are 'active', in that they have played a normal game in the last 12 months, then the average junior has played 28.25 games during that period.

That is a hell of a commitment and they deserve much better than they are getting.

Bill Gletsos
05-07-2004, 03:00 PM
Bill, again you are simply distorting the statistics. When you say 'active' you mean they are on the ACF active list. However, many of those individuals have not played any 'normal' games in the last 12 months, for whatever reason. If one looks at those that are 'active', in that they have played a normal game in the last 12 months, then the average junior has played 28.25 games during that period.
The only person distorting the facts is you.
I made it clear the juniors were active on the rapid list(i.e.played in the 12mth period) but virtually inactive in the normal list.
As such if they want more reliable normal ratings they need to actually play normal games.


That is a hell of a commitment and they deserve much better than they are getting.
Actually what they deserve is for you to stop distoring the facts for your own ends.

Bill Gletsos
05-07-2004, 03:09 PM
I've already explained many times how the failings in the system have come about. If you read through previous posts you would understand my sentiments.
Actually I took it Jeo meant name an actual existing chess rating system somewhere in the world that meets your requirements.


But essentially, there is no recoginition in the Glicko system that juniors, as a cohort (not all juniors) are improving as they mature.
This has been pointed out before. Some juniors get better, some dont, some improve quickly some dont. However it isnt just because they mature. It might be because they are getting coaching.
However if they are playing virtually no normal rated games, then you cannot know what they are doing.


This is a dynamic population, whereas the adult population is static. Therefore the population mean for this cohort should increase with time, dependant on the rate of observed change.

This does not mean all juniors increase their ratings automatically, but simply greater dynamic change in an upward direction is given to the improvers, in effect skewing the RD. Juniors who do not improve in this way will not increase their rating.
The Glicko2 already identifies improving juniors and accelerates their rating increase. However if they dont play enough games then the system cannot tell if they are improving or just were lucky.


If one thinks about learning, people learn and improve quickly, whereas they loose their understanding slowly. Skewing the RD in essence reflects this simple observation.
The volatility factor of Glicko2 caters for this.
An improving player is an improving player whether they be adulst or juniors.
However what determines whether they are improving or just lucky depends on the number of games played. If they play very few games then that is their problem.


However, in my previous post on this thread, I suggested that juniors maybe better served by introducing a grading system similar to the music system - the ideas for this were mentioned in the thread.
Yes, well I doubt anyone will take that seriously.


Bill, I can't believe you are trying to defend your position. The majority of juniors on the Gold Coast are being poorly served by the system.
The majority of juniors on the Gold Coast play very few normal rated games.
That appears to be a problem caused by the Gold Coast. they should provode the juniors with more normal rated events.

arosar
05-07-2004, 03:22 PM
But essentially, there is no recoginition in the Glicko system that juniors, as a cohort (not all juniors) are improving as they mature. This is a dynamic population, whereas the adult population is static. Therefore the population mean for this cohort should increase with time, dependant on the rate of observed change.

Look, mate...I'm no mathematical person meself, but isn't this rating business about measuring people's OTB performance? I just don't see how the 'maturity' has anything to do with it. And maturity in what?

AR

Bill Gletsos
05-07-2004, 03:26 PM
Look, mate...I'm no mathematical person meself, but isn't this rating business about measuring people's OTB performance? I just don't see how the 'maturity' has anything to do with it. And maturity in what?
He has been on about this for over 12mths.
He originally raised his maturation factor back during Ian Rogers original Bulletin letters.
This maturity he refers to is just juniors getting older(maturing) especially around puberty.
Of course he has provided no evidence that this has any effect on chess ability. Kevin has already debated him at length over this.

Kevin Bonham
05-07-2004, 04:57 PM
Therefore the population mean for this cohort should increase with time, dependant on the rate of observed change.

When juniors are underrated and playing at least some games against adults, this is going to happen anyway.


If one thinks about learning, people learn and improve quickly

Some do. Others never learn. :lol:

Cat
05-07-2004, 05:50 PM
Look, mate...I'm no mathematical person meself, but isn't this rating business about measuring people's OTB performance? I just don't see how the 'maturity' has anything to do with it. And maturity in what?

AR

The average rating increase for a junior chess player is around 50-60pts year. This seems to be fairly consistent over most junior populations I have looked at. This means that a junior cohort with a popultion mean ratings of say 1200, will be performing at an average of 1250-60 1 year later.

Now in an open pool, as you say, this shouldn't matter, because points will diffuse through the pool to ensure the rating average increases. The problem in Australia is that we have geographically isolated areas and the junior population are congregated in distinct areas, such as the Gold Coast, ACT and Box Hill.

Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other. What we're experiencing is that juniors have a much higher chance of playing other juniors, which leads to distortions in the rating pool. In other words the junior pools are partially closed systems and unless the juniors leave their pools to bring in new rating points, which is happening too infrequently, then their ratings lag behind their performance ( as a group ).

However we know that the junior pool is naturally inflating at 50-60 pts/year. Therefore, if this dynamic is incorporated into the Glicko system it can compensate for the pooling effects and the distortions it causes. This can simply be achieved by skewing the RD in the junior cohort to ensure sufficient dyanmic effect to see an average rise in the junior cohort mean rating of 50-60 pts/yr.

Cat
05-07-2004, 05:54 PM
The only person distorting the facts is you.
I made it clear the juniors were active on the rapid list(i.e.played in the 12mth period) but virtually inactive in the normal list.
As such if they want more reliable normal ratings they need to actually play normal games.


Actually what they deserve is for you to stop distoring the facts for your own ends.

Bill face facts, the average active junior (who has played a normal rated game in the last 12 months) plays 28.25 rated games /year on the Gold Coast and in your own words thats not enough for the Glicko. Its not they don't play enough, it's the Glicko thats failing them.

skip to my lou
05-07-2004, 06:35 PM
Actually I took it Jeo meant name an actual existing chess rating system somewhere in the world that meets your requirements.

Yes. My question is: What system would David use instead of the current system?

He said something about the ? music ? system or something like that but I could not find any information on it.

Cat
05-07-2004, 06:40 PM
Yes. My question is: What system would David use instead of the current system?

He said something about the ? music ? system or something like that but I could not find any information on it.

There's no need to change the system, simply skew the variance as I described above. There are a number of ways this could be implemented, all fairly easy.

What I said about gradings relate to a desire to integrate chess more closely with the education system. Chess will always struggle in Australia unless finds ways of becoming more commercially sustainable. Read the post I palced on this thread earlier.

It is unaffordable for most of the parents of these children to pay the costs required in travelling to increase their children's participation. The children have other interest and commitments in school, etc, and can only further their chess though great sacrifice.

What is it all about, about the preservation of an abstract rating system, or getting kids and people involved in playing chess? It is so difficult for some of these kids to progress, the rating system is so punitive, it is little wonder they give up.

So why are the ratings so important to these kids? Because in an environment where kids are submerged in a culture of development, where they progress through musical grades, through higher levels of speech & drama, school achievement, etc, they quickly learn that their chess is on a road to nowhere. When they are rewarded to play instruments, for debating, etc, recieve achievement certificates for sport, all they have to measure their chess achievement is their bogged-down rating.

Now this is not important to the kids, because they've got other things to do - they're happy. But it is important to Australian chess, because chess looses out everytime.

If chess is to flourish in Australia, it has to find ways to compete. There are so many interests out there today, that chess will become squeezed more and more. It must become kid-friendly, encouraging, rewarding. The rating system is punitive and it is irrelevant to the needs of the majority of children.

Bill Gletsos
05-07-2004, 07:00 PM
The average rating increase for a junior chess player is around 50-60pts year. This seems to be fairly consistent over most junior populations I have looked at. This means that a junior cohort with a popultion mean ratings of say 1200, will be performing at an average of 1250-60 1 year later.
This was shown to be rubbish when you said it ages ago in another thread. It is still rubbish.


Now in an open pool, as you say, this shouldn't matter, because points will diffuse through the pool to ensure the rating average increases. The problem in Australia is that we have geographically isolated areas and the junior population are congregated in distinct areas, such as the Gold Coast, ACT and Box Hill.
There is no proof that this is currently a significant issue.
And as I have pointed out there is no automatic means of determing what does and does not constitute a pool.


Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other. What we're experiencing is that juniors have a much higher chance of playing other juniors, which leads to distortions in the rating pool. In other words the junior pools are partially closed systems and unless the juniors leave their pools to bring in new rating points, which is happening too infrequently, then their ratings lag behind their performance ( as a group ).
A nice set of words but a load of rubbish.


However we know that the junior pool is naturally inflating at 50-60 pts/year.
Unproven rubbish.


Therefore, if this dynamic is incorporated into the Glicko system it can compensate for the pooling effects and the distortions it causes.
More rubbish.



This can simply be achieved by skewing the RD in the junior cohort to ensure sufficient dyanmic effect to see an average rise in the junior cohort mean rating of 50-60 pts/yr.
And still more rubbish.

Bill Gletsos
05-07-2004, 07:33 PM
Bill face facts, the average active junior (who has played a normal rated game in the last 12 months) plays 28.25 rated games /year on the Gold Coast and in your own words thats not enough for the Glicko.
Face facts David the average GC junior who is active in rapid chess is significantly inactive in normal chess.



Its not they don't play enough, it's the Glicko thats failing them.
No its the fact they dont play enough normal rated games.
Their lack of games means you cannot tell if they are static, improving or just lucky and /or unlucky.

Garvinator
05-07-2004, 07:42 PM
No its the fact they dont play enough normal rated games.

arent you getting tired of sounding like a broken record :hmm: maybe you should dedicate some times to sending me the sp files for the caloundra open instead ;)

Bill Gletsos
05-07-2004, 07:53 PM
arent you getting tired of sounding like a broken record :hmm:
Well I can keep it up as long as David keeps playing the same old tune.


maybe you should dedicate some times to sending me the sp files for the caloundra open instead ;)
I currently dont have them.

Cat
05-07-2004, 10:35 PM
[QUOTE=Bill Gletsos]Face facts David the average GC junior who is active in rapid chess is significantly inactive in normal chess.

Bill, the juniors who are active only in rapid chess are generally the beginners. They are irrelevant to the whole debate. They play in rapid games at the Gardiner Chess Centre, and their ratings are primarily sent for Queensland Junior Ratings, which you accept as being reliable enough for an ACF rapid rating. This is a side issue, and you're only persisting with this to muddy the waters. The serious juniors play in normal ACF games regularly, on average 28.25 games/yr, and in your own words the Glicko can't cope.

Cat
05-07-2004, 10:42 PM
This was shown to be rubbish when you said it ages ago in another thread. It is still rubbish.



Bill we went over this on the old BB. We could see a clear trend albeit in a small group. I didn't persist with it then because I didn't want to humiliate you. If you seriously want me to bring up data to demonstrate this its fairly easy to do. Bill, just saying rubbish is no response at all. Your friends are right, if you have nothing of value to say its best to keep quiet and I can go away.

Bill Gletsos
05-07-2004, 10:47 PM
Bill, the juniors who are active only in rapid chess are generally the beginners. They are irrelevant to the whole debate. They play in rapid games at the Gardiner Chess Centre, and their ratings are primarily sent for Queensland Junior Ratings, which you accept as being reliable enough for an ACF rapid rating.
What do you mean I accept them as being reliable enough for an ACF rapid rating, as I'm not sure what you mean.



This is a side issue, and you're only persisting with this to muddy the waters. The serious juniors play in normal ACF games regularly, on average 28.25 games/yr, and in your own words the Glicko can't cope.
28.5 games per year is a small number of games when you consider it is spread over a 12 mth period.
There is not enough games for you to determine if the junior is static, improving slowly,improving rapidly, lucky or unlucky.
If you cannot see this then as I always suspected you are just a waste of time.

As for the Glicko not handling it, neither would ELO, FIDE or any other rating existing system. Of course it is to be expected because you cannot get accurate rating with insufficient results.

Bill Gletsos
05-07-2004, 10:53 PM
Bill we went over this on the old BB. We could see a clear trend albeit in a small group. I didn't persist with it then because I didn't want to humiliate you.
You based that figure on your flawed analysis of the past Gold Coast Open.
You mixed ratings from Elo, early Glicko and inaccurate Swiss Perfect performance ratings.


If you seriously want me to bring up data to demonstrate this its fairly easy to do.
Its no use you presenting flawed data, its just a waste of your time and mine.


Bill, just saying rubbish is no response at all. Your friends are right, if you have nothing of value to say its best to keep quiet and I can go away.
Not only I but others showed this was rubbish when you originally presented it since your method was flawed.
Save us all the trouble stop wasting our time.

Cat
05-07-2004, 11:09 PM
You based that figure on your flawed analysis of the past Gold Coast Open.
You mixed ratings from Elo, early Glicko and inaccurate Swiss Perfect performance ratings.
.

Bill I used Swissperfect as you know to make a comparison between junior and adult rating performance. Its true that if I had simply measured adults and juniors, the adults being near the top end of the rating groups, and the juniors at the bottom, then the results would have been erroneous. To avoid this I divided the pool into rating groups so that adults and juniors were compared to each other within rating groups.

Across virtually all rating groups there was about rating differential of about 200 rating points, the juniors being rated around 200 points less than the adults with the same performance. This was a consistent finiding through all rating groups over 4 years. Even if Swissperfect was as crude as you suggest, it wouldn't have created a massive junior/adult bias. The results were so striking that there was only one conclusion to draw.

Perhaps I'll use Barry's calculator to assess the Gold Coast Open. I suspect the picture has improved substantially, you might be pleasantly surprised.

Cat
05-07-2004, 11:13 PM
Save us all the trouble stop wasting our time.

I'm not wasting your time Bill. I'm sure my persistent pestering you has made you address issue you would otherwise not have regarded.

Kevin Bonham
05-07-2004, 11:16 PM
I didn't persist with it then because I didn't want to humiliate you.

What a joke. You've made heaps of very clearcut attempts to humiliate Bill but the only people you have humiliated are yourself and Matthew Sweeney (on the rare occasions he has unwisely decided to play #2 in your tag-team).


If you seriously want me to bring up data to demonstrate this its fairly easy to do.

We've seen this before. You've got data and you've got things you want to say with them but you don't know how to join the dots, mainly because they're not even on the same sheet of paper.


Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other.

Got a reference for that claim? The system clearly does not require any such assumption in order to function.


So why are the ratings so important to these kids? Because in an environment where kids are submerged in a culture of development, where they progress through musical grades, through higher levels of speech & drama, school achievement, etc, they quickly learn that their chess is on a road to nowhere. When they are rewarded to play instruments, for debating, etc, recieve achievement certificates for sport, all they have to measure their chess achievement is their bogged-down rating.

There are many ways such a system could exist alongside a ratings system. State associations could award junior class levels along similar lines to the awards of title norms. If you think it is worth doing, why don't you write to your state association offering to administer it? You could probably have come up with a pretty decent system in the time that you spend spouting nonsense about ratings here.


The rating system is punitive and it is irrelevant to the needs of the majority of children.

Try organising unrated tournaments for juniors - advertised as such - and let us know how the entry numbers stack up compared with the rated ones.

Bill Gletsos
05-07-2004, 11:17 PM
Bill I used Swissperfect as you know to make a comparison between junior and adult rating performance. Its true that if I had simply measured adults and juniors, the adults being near the top end of the rating groups, and the juniors at the bottom, then the results would have been erroneous. To avoid this I divided the pool into rating groups so that adults and juniors were compared to each other within rating groups.

Across virtually all rating groups there was about rating differential of about 200 rating points, the juniors being rated around 200 points less than the adults with the same performance. This was a consistent finiding through all rating groups over 4 years. Even if Swissperfect was as crude as you suggest, it wouldn't have created a massive junior/adult bias. The results were so striking that there was only one conclusion to draw.
All of your conclusions were shown to be flawed at the time.


Perhaps I'll use Barry's calculator to assess the Gold Coast Open. I suspect the picture has improved substantially, you might be pleasantly surprised.
I wont hold my breath if you are doing the calcs.

Bill Gletsos
05-07-2004, 11:29 PM
I'm not wasting your time Bill. I'm sure my persistent pestering you has made you address issue you would otherwise not have regarded.
The only effect you had was when we included the GC in our corrections when we did the ACT junior correction.

As for actual changes to the rating system calculation changes then no change that has been implemented by Graham and I have has had anything to do with your pestering.
Virtually all of them were based on ideas we had prior to Ian's original letter to the ACF bulletin.

Alan Shore
05-07-2004, 11:35 PM
Try organising unrated tournaments for juniors - advertised as such - and let us know how the entry numbers stack up compared with the rated ones.

I agree, rating is a big motivator for juniors to play tournaments. Abolishing it would destroy the purpose of seedings and affect the outcome of results.


Perhaps I'll use Barry's calculator to assess the Gold Coast Open. I suspect the picture has improved substantially, you might be pleasantly surprised.

How come you didn't play?

Garvinator
05-07-2004, 11:37 PM
How come you didn't play?
too many under rated juniors for DR's liking :lol: :lol: :whistle: ;)

Cat
05-07-2004, 11:37 PM
Got a reference for that claim? The system clearly does not require any such assumption in order to function.

Try Glickmans paper


There are many ways such a system could exist alongside a ratings system. State associations could award junior class levels along similar lines to the awards of title norms. If you think it is worth doing, why don't you write to your state association offering to administer it? You could probably have come up with a pretty decent system in the time that you spend spouting nonsense about ratings here.

Yes the QJR is well run, but needs corrections like the ACF from time to time.


Try organising unrated tournaments for juniors - advertised as such - and let us know how the entry numbers stack up compared with the rated ones.

It's not so simple. It's the structure I'm talking about. It exists because its what was inherited. The question is, how well does it work? Graeme's done a great job with the juniors on the Coast, at the primary school level, things couldn't be rosier. The problem is the drop out rate. High School Chess is struggling here, and I think it's probably a world-wide problem. We could continue to accept this as a fact of life, but if so it's likely things are going to get worse.

Competition for children's time is increasing. If we sit on our hands, chess will likely loose out. We should ask ourselves is this the best way to get kids playing chess. This is the kind of proper debate Australian Chess needs. It's sad that there isn't the good-will to discuss these things without abuse being delivered from all quarters.

Cat
05-07-2004, 11:41 PM
[QUOTE=Bruce Dickinson


How come you didn't play?

I was on holiday - Frazer Island, fishing & relaxing. The GC Open seems to have co-incided with the school holidays the last few years.

Cat
05-07-2004, 11:42 PM
too many under rated juniors for DR's liking :lol: :lol: :whistle: ;)

I would have been in the Open Section, not too many juniors there, but 2 days of probable misery for me, for sure.

Bill Gletsos
05-07-2004, 11:45 PM
Got a reference for that claim? The system clearly does not require any such assumption in order to function.
He also claimed in the Underated junior thread that:

Professor Glickman in his original paper on the Glicko system describes how the 'bedrock adult population' was essential for the proper operation of the Glciko system - that the large adult population would act as a stabilising effect in the system and that developing juniors would move more rapidly to this much larger population mass.

Now Glickmans original paper on the Glicko is called "Parameter Estimation in Large Dynamic Paired Comparison Experiments".
At no point in that document does the words "adult bedrock population" occur. (I say this because I did individual searches for the words adult and bedrock and neither appear to be in it)

Bill Gletsos
05-07-2004, 11:47 PM
Try Glickmans paper
Not good enough.
Based on your previous claim(see above) provide the actual page document name and page.

Kevin Bonham
06-07-2004, 12:32 AM
Try Glickmans paper

Like Bill, I would like to see you explain exactly which bit backs up what you are saying about the system assuming all players are equally likely to play each other.


Yes the QJR is well run, but needs corrections like the ACF from time to time.

I wasn't even talking about the QJR. I was talking about whether there should be a chess equivalent of grading in music as you suggested.


It's not so simple. It's the structure I'm talking about. It exists because its what was inherited. The question is, how well does it work?

That is just another evasion. Though I don't blame you for running away from the point; it would certainly bite you if it didn't. If, as you maintain, ratings are letting juniors down and making them drop out of chess, then if you organise an event without ratings as competition to the rated events, and publicise its unrated nature, the disenchanted juniors will vote with their feet.


We should ask ourselves is this the best way to get kids playing chess. This is the kind of proper debate Australian Chess needs. It's sad that there isn't the good-will to discuss these things without abuse being delivered from all quarters.

What's sad is that you keep carrying on as if the ratings system is the be-all-and-end-all of children's interest in chess when (a) it isn't, (b) you can't even prove that its net effect is negative. If "ways to get kids playing more chess" is your real motive on the ratings threads, then why not start seperate threads on that issue and cover the whole gamut of relevant points?

Of course, this is not really your prime reason for getting involved in the ratings debate - it is simply a high and noble sentiment that no-one could quibble with in and of itself, which you use to attempt to legitimise your beat-ups. :lol:

A question about this whole skewing-the-RD idea you propose. What makes you think it will work at all where Matthew's trending experiment - based on very similar ideas - failed spectacularly?

Cat
06-07-2004, 09:49 AM
Like Bill, I would like to see you explain exactly which bit backs up what you are saying about the system assuming all players are equally likely to play each other.

I'll dig out the quotes when I get a chance



What's sad is that you keep carrying on as if the ratings system is the be-all-and-end-all of children's interest in chess when (a) it isn't, (b) you can't even prove that its net effect is negative. If "ways to get kids playing more chess" is your real motive on the ratings threads, then why not start seperate threads on that issue and cover the whole gamut of relevant points?

Precisely, its not the be-all and end-all. It's the system that makes it so. Its necessary for qualifications at all levels.

As for proving negative effects, there is a wealth of information on learning processes, what works, what doesn't. The systems in place in the education system haven't arisen by accident, it's a distillation of accumulated understanding of what works and what doesn't - i.e. best practice. We should learn from their experience, not deride it. There is certainly nothing out there that supports our rating system as a developmental tool.


Of course, this is not really your prime reason for getting involved in the ratings debate - it is simply a high and noble sentiment that no-one could quibble with in and of itself, which you use to attempt to legitimise your beat-ups. :lol:

No, well what is my motivation exactly Dr Freud?


A question about this whole skewing-the-RD idea you propose. What makes you think it will work at all where Matthew's trending experiment - based on very similar ideas - failed spectacularly?

Because the kind of modelling I'm talking about is widely used to express population dynamics within the Medical Profession. Again there is a wealth of information available of which you're obviously unaware.

Look, the average (truely) active junior on the Gold Coast plays 28.25 games /yr. How many of these games are 'internal'(between players SE Queensland/Gold Coast/etc)? Juniors represent 70% of the Gold Coast playing population, and the ratio of !! ratings between juniors & adults is about 29/6. From that you should be able to gauge the effect of population pooling. Publish the figures and there can be no quibbling.

Kevin Bonham
06-07-2004, 04:46 PM
Precisely, its not the be-all and end-all. It's the system that makes it so. Its necessary for qualifications at all levels.

What qualifications? If you're talking about selections, these only concern a tiny %age of all junior players. If you're talking about ratings groups, being underrated gives a junior an unfair advantage. Most likely, you're simply talking nonsense. :hand:


As for proving negative effects, there is a wealth of information on learning processes, what works, what doesn't.

None of which has been applied specifically to chess. You cannot extrapolate from what motivates a child to do well in school to what motivates them to do well on the chessboard as, for starters, the purposes of the child in pursuing each activity are different.


There is certainly nothing out there that supports our rating system as a developmental tool.

On the line you have been arguing, doesn't assessment of any kind (eg exams) where the child is not thrown bonuses for no reason start to look harsh and discouraging?


No, well what is my motivation exactly Dr Freud?

Don't know, don't care, all I know is that you are engaging in a beat-up.


Because the kind of modelling I'm talking about is widely used to express population dynamics within the Medical Profession. Again there is a wealth of information available of which you're obviously unaware.

So? Just because a model works in predicting one thing doesn't mean it will work in predicting something different. Glicko has its own assumption, which is that volatile performance can be predicted from past volatile performance. If you think that change in performance can be better predicted by something else, such as age, put up a list of juniors whose ages you know and a model for how you would alter the system to "improve" it. Then we can track it and see how it goes.


Look, the average (truely) active junior on the Gold Coast plays 28.25 games /yr. How many of these games are 'internal'(between players SE Queensland/Gold Coast/etc)? Juniors represent 70% of the Gold Coast playing population, and the ratio of !! ratings between juniors & adults is about 29/6. From that you should be able to gauge the effect of population pooling. Publish the figures and there can be no quibbling.

Completely irrelevant to the question that I asked. When you argued that there was a Gold Coast ratings problem you used this line of argument to explain it and it turned out that while there was a problem, your explanation of it was incorrect. I suspect that it still is.

Since you didn't answer my last question at all usefully, here's another one: How do the rating dynamics of juniors playing mostly other juniors in an area with high chess activity vary from the rating dynamics of juniors playing mostly adults in a low chess activity area? This question is not difficult but there are two relevant points to grasp.

Cat
06-07-2004, 11:18 PM
These Statements are taken from the USCF Ratings Committee Report 1996 (Glickman & Matthews), when Glickman first described the plan for the introduction of the Glicko system, and a Glickman paper 1999;

Hobby players who are satisfied to improve slowly, or not at all, have fairly certain ratings. These players are the bedrock of the system.

Only 6 of our bedrock adults have !! rating.

Implementing this two-parameter system, allowing for subjective estimates of playing strength, was accomplished using standard Bayesian methodology.

You're not even aware of the mathematical origins of the system!

The Glicko System works best when the number of games in a rating period is moderate, say an average of 5-10 games per player in a rating period.

Our active juniors play 28.25 games/yr, or 7 games per rating period. These are Glickman's words Bill. It's not the juniors fault, matey!

One practical problem with the Glicko system is that when a player competes very frequently, his/her rating stops changing appreciably which reflects that the RD is very small. This may sometimes prevent a player's rating from changing substantially when the player is truly improving.

Ergo, the problem. Now before you tell me volatility takes care of all of that, I say it's not enough. Things are better, I grant you, but bury your prejudices and tackle the problem objectively.

Kevin Bonham
07-07-2004, 12:48 AM
Implementing this two-parameter system, allowing for subjective estimates of playing strength, was accomplished using standard Bayesian methodology.

You're not even aware of the mathematical origins of the system!

You're not even aware of what I actually wrote. :rolleyes: I disputed your claim that the system made the assumption that all players are equally likely to play each other, not the claim that the system was Bayesian. You have still not provided any quote from Glickman or any other source to show that this is so. You are wasting both of our time.


Hobby players who are satisfied to improve slowly, or not at all, have fairly certain ratings. These players are the bedrock of the system.

Only 6 of our bedrock adults have !! rating.

How many have ! ratings? Is it such an issue if an adult is ! rather than !! anyway ? Just looking at the Tassie list (which has 10 !!s and 22 !s) I would say that nearly all the adults with ! ratings are of fairly certain strength. Their ratings appear accurate within, say, 100 points, anyway.


One practical problem with the Glicko system is that when a player competes very frequently, his/her rating stops changing appreciably which reflects that the RD is very small. This may sometimes prevent a player's rating from changing substantially when the player is truly improving.

Ergo, the problem. Now before you tell me volatility takes care of all of that, I say it's not enough. Things are better, I grant you, but bury your prejudices and tackle the problem objectively.

Oh dear, I see that someone wasn't paying attention when Bill explained that the Australian versions of both Glicko and Glicko 2 were fitted with RD floors to circumvent exactly this problem. :hand:

Better luck next time.

How about you "tackle the problem objectively" by providing empirical evidence (not airy-fairy theoretical waffle from other fields) that Glicko-2 is still not working for your region post-fix?

Bill Gletsos
07-07-2004, 12:50 AM
These Statements are taken from the USCF Ratings Committee Report 1996 (Glickman & Matthews), when Glickman first described the plan for the introduction of the Glicko system, and a Glickman paper 1999;
Nice try but you dont know what you are talking about.
Firstly this is a report by the USCF Ratings Committee of which Glickman just happens to be the chairman.
Its for this reason that only his name appears on the title page.
This is to be expected as the views in the report aew those of the committee and not particular members unless explicitly stated.
Chapter 2 is by Ken Sloan and refers to rating floors.
Paul Matthews is the author of chapter 6 of the report which describes the American Go Association rating system.

The Glicko system existed well before this document was published.
In fact it notes that the "Glicko" system was already in place on the FICS server. The report makes this comment whilst mentioning the committee was looking at the eventual transition to a two paramter rating system.

So far the USCF has not yet implemented a two parameter rating sysytem.



Hobby players who are satisfied to improve slowly, or not at all, have fairly certain ratings. These players are the bedrock of the system.
Ha ha ha. Firstly this quote isnt even by Glickman. Its by Paul Matthews on chapter 6 of the report where he is describing the American Go Assocoation system.
In fact the above quote is referring to go not chess.


Secondly you said in a previous post that Glickman refered to the "bedrock adult population". As can be seen no such comment exists by anyone.
In fact if this is your justification for the the term "bedrock adult population" then you are way off the mark.
The reference to "hobby" players does not necessarily make them adults. Also it refers to players who are prepared to improve slowly.
No although he wasnt referring to chess but go lets assume for a second this statement was applicable to chess. There are many adults who put a lot of effort into studying chess who could hardly refer to as "hobby" players.

You are just trying to read support for your viewpoint from this statement.
You are a joke.



Implementing this two-parameter system, allowing for subjective estimates of playing strength, was accomplished using standard Bayesian methodology.
You cretin.
This is Matthews referring to the AGA system not the Glicko system.


You're not even aware of the mathematical origins of the system!
The only one not aware of the origins of the Glikco system is you.
Glickman had used Bayesian methodology long before 1996.
In fact Glickman investigated aspects of what became the Glicko system as far back as 1991.


The Glicko System works best when the number of games in a rating period is moderate, say an average of 5-10 games per player in a rating period.
You cannot even given decent references.
This does not appear in the above rating committees report.
It is actually in the simple 5 page document describing the Glicko system.

As Glickman notes in the actual complcated 30 page document describing the derivation of the Glicko system he states:
"The choice of length of a rating period involves a variance/bias tradeoff. For short rating periods, little data may be available to estimate players' strengths, and the analytic approximations used in the algorithm in Section 3.3 may not be dependable. Conversely, if long rating periods are used, a player's ability may have changed substantially over a rating period, but this would not be detectable. The best compromise seems to be rating periods that are as short as possible, but where enough data is available to have some indication of players' strengths, perhaps at least 5--10 games per player on average."
However this is not a problem in the Glicko2 system, as its very reason for being is to detect substaintial rating ability changes.

In fact it isnt even a problem in the Glicko system if a good rating period length, number of games and c-squared value are chosen.

All I can say is that Graham and I know far better than you what constitutes an acceptable number of games.


Our active juniors play 28.25 games/yr, or 7 games per rating period. These are Glickman's words Bill. It's not the juniors fault, matey!
Firstly Graham and I know Glickmans words far better than you as we have corresponded with him on a number of occasions in regards to both the Glicko and Glicko2 systems.



One practical problem with the Glicko system is that when a player competes very frequently, his/her rating stops changing appreciably which reflects that the RD is very small. This may sometimes prevent a player's rating from changing substantially when the player is truly improving.
You really are a moron.
Thats just another quote out of context.
Immediately after that statement Glickman then says:
"I would therefore recommend that an RD never drop below a threshold value,
such as 30, so that ratings can change appreciably even in a relatively short time."

At no stage yet since the implementation of the Glicko/Glicko2 ACF system has that minimum value had to come into play.


Ergo, the problem. Now before you tell me volatility takes care of all of that, I say it's not enough. Things are better, I grant you, but bury your prejudices and tackle the problem objectively.
Ergo the only problem is you have no clue what you are talking about.

Cat
07-07-2004, 01:03 AM
[QUOTE=Kevin Bonham]You're not even aware of what I actually wrote. :rolleyes: I disputed your claim that the system made the assumption that all players are equally likely to play each other, not the claim that the system was Bayesian. You have still not provided any quote from Glickman or any other source to show that this is so. You are wasting both of our time.

I didn't say it was necessary that all players were equally likely to play each other, it's not necessary in a large pool like the internet because there's enough mix. But in small regional groups it becomes important.



How many have ! ratings? Is it such an issue if an adult is ! rather than !! anyway ? Just looking at the Tassie list (which has 10 !!s and 22 !s) I would say that nearly all the adults with ! ratings are of fairly certain strength. Their ratings appear accurate within, say, 100 points, anyway.

Give me a break! You're really scratching now.




Oh dear, I see that someone wasn't paying attention when Bill explained that the Australian versions of both Glicko and Glicko 2 were fitted with RD floors to circumvent exactly this problem. :hand:

Yes Glickman recommended 30. Why 30? The flooring doesn't solve the problem - it ameliorates the problem.


How about you "tackle the problem objectively" by providing empirical evidence (not airy-fairy theoretical waffle from other fields) that Glicko-2 is still not working for your region post-fix?

Post which fix KB, there have been so many? As I say, I believe things are a lot better, that was indicated from the last Gold Coast Classic, but it was a small sample. I'm looking at the Gold Coast Open, as I said.

Bill Gletsos
07-07-2004, 01:09 AM
I didn't say it was necessary that all players were equally likely to play each other, it's not necessary in a large pool like the internet because there's enough mix. But in small regional groups it becomes important.
You cannot even distort the truth accurately.
You said :
"Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other."

You cant actually provide any reference by Glickman where he says this now can you? :whistle:



Give me a break! You're really scratching now.
If anyones been scratching it is you.


Yes Glickman recommended 30. Why 30? The flooring doesn't solve the problem - it ameliorates the problem.
This is just more unsubstaintaited rubbish on your part.
You have absolutely no clue what you are talking about.



Post which fix KB, there have been so many? As I say, I believe things are a lot better, that was indicated from the last Gold Coast Classic, but it was a small sample. I'm looking at the Gold Coast Open, as I said.
Dont act so dumb and be a complete cretin. (ah I get it its no act :whistle: )
It is obvious KB is referring to the GC correction made at the same time as the ACT junior correction.

Kevin Bonham
07-07-2004, 01:25 AM
I didn't say it was necessary that all players were equally likely to play each other, it's not necessary in a large pool like the internet because there's enough mix. But in small regional groups it becomes important.

You said:

Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other.

You have still not provided any evidence that the system assumes this. It's very simple - put up evidence that the system assumes this, or retract it.


Give me a break! You're really scratching now.

No, you just can't be bothered answering, or can't answer, the point so you tread water with random mindless drivel rather than admit it. Another standard beat-up tactic, as is using other people's comments completely out of context (of which your post now demolished by Bill was one of the saddest cases I have ever seen. :lol: ) Go back and answer the question I asked.


Yes Glickman recommended 30. Why 30? The flooring doesn't solve the problem - it ameliorates the problem.

You still haven't proven that the problem you are talking about, with the mechanism that you claim, even exists. Maybe had you actually answered my previous question about different ratings pools we would have made some progress here?


Post which fix KB, there have been so many?

As Bill points out I was referring to the Gold Coast fix, as should be fairly obvious since I don't think any of the others except the G - G2 transition would have affected the dynamics of your region significantly.

Cat
07-07-2004, 01:41 AM
[
QUOTE=Bill Gletsos]Nice try but you dont know what you are talking about.
Firstly this is a report by the USCF Ratings Committee of which Glickman just happens to be the chairman.
Its for this reason that only his name appears on the title page.
This is to be expected as the views in the report aew those of the committee and not particular members unless explicitly stated.
Chapter 2 is by Ken Sloan and refers to rating floors.
Paul Matthews is the author of chapter 6 of the report which describes the American Go Association rating system.

What a load of rubbish. I acknowledged Matthews above - Glickman has put his name on the paper, it was off his web-site - it is entirely legitimate to believe Glickman fully supports these opinions. It's tragic you find consolation in making such facile points.



Secondly you said in a previous post that Glickman refered to the "bedrock adult population". As can be seen no such comment exists by anyone.
In fact if this is your justification for the the term "bedrock adult population" then you are way off the mark.
The reference to "hobby" players does not necessarily make them adults. Also it refers to players who are prepared to improve slowly.
No although he wasnt referring to chess but go lets assume for a second this statement was applicable to chess. There are many adults who put a lot of effort into studying chess who could hardly refer to as "hobby" players.

Bedrock adult population were my words. You suggested there was no evidence for my statement, and I responded that it was in Glickman's Paper to the USCF 1996 as above. Whether it's bedrock adult or bedrock hobby player, most of our adults certainly fit that discription. This is the kind of semantic distinction one would expect a child to make, it is entirely irrelevant. Unless there is a bedrock stable population, then improving players at the lower end of the rating scale will fail to progress. Grow up!





Glickman had used Bayesian methodology long before 1996.
In fact Glickman investigated aspects of what became the Glicko system as far back as 1991.

You were the one disputing Bayesian methodology, not me.


You cannot even given decent references.
This does not appear in the above rating committees report.
It is actually in the simple 5 page document describing the Glicko system.

As Glickman notes in the actual complcated 30 page document describing the derivation of the Glicko system he states:
"The choice of length of a rating period involves a variance/bias tradeoff. For short rating periods, little data may be available to estimate players' strengths, and the analytic approximations used in the algorithm in Section 3.3 may not be dependable. Conversely, if long rating periods are used, a player's ability may have changed substantially over a rating period, but this would not be detectable. The best compromise seems to be rating periods that are as short as possible, but where enough data is available to have some indication of players' strengths, perhaps at least 5--10 games per player on average." However this is not a problem in the Glicko2 system, as its very reason for being is to detect substaintial rating ability changes.

In fact it isnt even a problem in the Glicko system if a good rating period length, number of games and c-squared value are chosen.

I said some of the quotes come from the 1999 paper. You've missed the point entirely, the point of the quote was to demonstrate that our juniors play more than enough games to satisfy the data requirements Glickman advocates. Stop making excuses.






"I would therefore recommend that an RD never drop below a threshold value,
such as 30, so that ratings can change appreciably even in a relatively short time."

At no stage yet since the implementation of the Glicko/Glicko2 ACF system has that minimum value had to come into play.



We've debated this ad nauseum. I've always argued that reducing RD when players play frequently reduces dynamic change in the junior population unfairly, you've denied there was ever any support for this. I included this statement to demonstrate that Glickman himself recognises the potential problem. You believe the floorings protect enough - we disagree.

Cat
07-07-2004, 01:55 AM
[
QUOTE=Bill Gletsos]You cannot even distort the truth accurately.

True Bill, you're much better at distorting truth than I.


You said :
"Now the Glicko system uses Bayesian methodology which essentially assumes that any one player has approximately the same chance of playing any other."

You cant actually provide any reference by Glickman where he says this now can you? :whistle:

Yes, Bayesian methodology assumes randomness, but it doesn't mean that in a large population it must be a necessity, it's only in small populations that non-random associations become significant.





This is just more unsubstaintaited rubbish on your part.
You have absolutely no clue what you are talking about.

Bill, underrating of juniors has been so endemic that people assume its part of the natural order. You demanding, or professor suggesting that the floors are adequate does not mean it is, at least not in Australia. When juniors are competing on close to equal terms, then you can gain comfort. Maybe we're already there, but recent history is not convincing.

Cat
07-07-2004, 02:01 AM
[QUOTE=Kevin Bonham]You said:

You have still not provided any evidence that the system assumes this. It's very simple - put up evidence that the system assumes this, or retract it.

Well, why don't you try reading Bayes' theorum?


Bill was one of the saddest cases I have ever seen. :lol: ) Go back and answer the question I asked.

Yes, very sad. KB, I'm not interested in your silly games. I'm interested to see if the corrections have been adequate, and that the improvements can be maintained.

Bill Gletsos
07-07-2004, 10:50 AM
What a load of rubbish. I acknowledged Matthews above - Glickman has put his name on the paper, it was off his web-site - it is entirely legitimate to believe Glickman fully supports these opinions. It's tragic you find consolation in making such facile points.
What is tragic is that it is apparent you have no idea how committees work.
To assume that all the views in the report are endorsed by Glickman is sheer folly. Neither of us can tell what he does and does not support from the report because it is a COMMITTEE report.



Bedrock adult population were my words.
Yes it is obvious now they are your words. However you claimed they were Glickman's when you said:

Professor Glickman in his original paper on the Glicko system describes how the 'bedrock adult population' was essential for the proper

This clearly is not the case.



You suggested there was no evidence for my statement, and I responded that it was in Glickman's Paper to the USCF 1996 as above. Whether it's bedrock adult or bedrock hobby player, most of our adults certainly fit that discription. This is the kind of semantic distinction one would expect a child to make, it is entirely irrelevant. Unless there is a bedrock stable population, then improving players at the lower end of the rating scale will fail to progress. Grow up!
As I pointed out it is Matthews words not Glickmans.
Also Matthews was referring to Go not chess.
You are just clutching at straws trying to avoid looking like the goose you are.



You were the one disputing Bayesian methodology, not me.
No, I was asking you to provide a reference by Glickman to backup your claim that his methodology for the Glicko system assumes that all players are equally likely to play each other.

So far you have not provided it, mainly because he has never said it.




I said some of the quotes come from the 1999 paper.
Completely incorrect.
You said:

These Statements are taken from the USCF Ratings Committee Report 1996
There is no some mentioned there.
It implies all the following statements were from that report.
To now deny it makes you look like a complete joke.



You've missed the point entirely, the point of the quote was to demonstrate that our juniors play more than enough games to satisfy the data requirements Glickman advocates.
As I pointed out you have no clue what you are talking about.



We've debated this ad nauseum. I've always argued that reducing RD when players play frequently reduces dynamic change in the junior population unfairly, you've denied there was ever any support for this.
Yes because there is no support for this assertion.


I included this statement to demonstrate that Glickman himself recognises the potential problem.
He does not recognise is at you portray it.



You believe the floorings protect enough - we disagree.
Disagree all you like.
What is clear is that you have totally no understanding how the Glicko and Glicko2 systems work and the differences between them.

Here is a clue.
In the Glicko every game played causes a reduction in the RD and the more games played the more the RD will decrease, however how much it decreases is dependent not on the number of games but the ratings and RD's of ones opponents.

With regards the Glicko2, this is not the case as the RD can not only decrease but more importantly increase even if many games are played.

Of course all of this is obvious to anyone who has actually carried out Glicko and Glicko2 calculations.
You of course have done neither.

Bill Gletsos
07-07-2004, 11:00 AM
True Bill, you're much better at distorting truth than I.
You are a complete joke.
You make claims that were never said and you continually take quotes out of context.
You are as always a waste of time.



Yes, Bayesian methodology assumes randomness, but it doesn't mean that in a large population it must be a necessity, it's only in small populations that non-random associations become significant.
I would ask you to provide a reference, but your previous attempts have been so feeble, i wont bother as it will only waste my time easily pulling it apart.



Bill, underrating of juniors has been so endemic that people assume its part of the natural order. You demanding, or professor suggesting that the floors are adequate does not mean it is, at least not in Australia.
Yes I'm really going to take your viewpoint you clown over that of Glickman's.
Dont change your day job as stats sure isnt your strong point.



When juniors are competing on close to equal terms, then you can gain comfort. Maybe we're already there, but recent history is not convincing.
Yeah yeah doc. :whistle:

arosar
07-07-2004, 11:10 AM
Alrighty youse blokes...I've decided to form me own independent committee of mathematicians and sort this all out. We don't wanna see a ratings debate between youse for the next bloody 50 years you mongrels!

I know a genius who knows more maths than youse all put together: the 1minute_grandmaster. He can tell me who's making sense. And I got me mate DS who will surely receive the Nobel Prize for Chemistry one day - very poor chess player though but FMD(!) - what a freakin' genius!! I'll get back to youse soon.

Good day boys!

AR

Cat
07-07-2004, 11:20 AM
=Bill Gletsos]What is tragic is that it is apparent you have no idea how committees work.
To assume that all the views in the report are endorsed by Glickman is sheer folly. Neither of us can tell what he does and does not support from the report because it is a COMMITTEE report.



Yes it is obvious now they are your words. However you claimed they were Glickman's when you said:


This clearly is not the case.



As I pointed out it is Matthews words not Glickmans.
Also Matthews was referring to Go not chess.
You are just clutching at straws trying to avoid looking like the goose you are.



No, I was asking you to provide a reference by Glickman to backup your claim that his methodology for the Glicko system assumes that all players are equally likely to play each other.

So far you have not provided it, mainly because he has never said it.




Completely incorrect.
You said:

There is no some mentioned there.
It implies all the following statements were from that report.
To now deny it makes you look like a complete joke.



As I pointed out you have no clue what you are talking about.



Yes because there is no support for this assertion.


He does not recognise is at you portray it.


All this is semantic nonsense - it doesn't deserve any response, just gives an impression of a mind of a child.


Here is a clue.
In the Glicko every game played causes a reduction in the RD and the more games played the more the RD will decrease, however how much it decreases is dependent not on the number of games but the ratings and RD's of ones opponents. With regards the Glicko2, this is not the case as the RD can not only decrease but more importantly increase even if many games are played.

Of course all of this is obvious to anyone who has actually carried out Glicko and Glicko2 calculations.
You of course have done neither.


As I say the details of Glicko 2 are copyrighted, not available and the times I have invited you make them available you've declined (BJC's web-site has details of Glicko- 1). Make them available and then lets open the debate by all means.

Bill, the points I am making is clear for all to see. Our juniors play 28.25 games/yr. That's well within the 5-10/rating period range optimal for the operation of the Glicko system.

If your fixes don't solve the problem, you can't blame anybody else, the responsiblity rest with you. It's not our juniors fault, it's not our organisers fault, you can't keep pointing the finger. If the fixes solve the problem I will be the first to thank you. If not then I will be asking you for a solution.

Bill Gletsos
07-07-2004, 11:33 AM
All this is semantic nonsense - it doesn't deserve any response, just gives an impression of a mind of a child.
Ha ha this is rich coming from someone who cannot even use the BB quote function correctly.
Of course it isnt semantic nonsense.
You cannot answer it so you just try and dismiss it.
You are a joke.



As I say the details of Glicko 2 are copyrighted, not available and the times I have invited you make them available you've declined (BJC's web-site has details of Glicko- 1). Make them available and then lets open the debate by all means.
Ha ha ha ha.
This is even better.
You are a complete moron.
As far as I can recall you have never asked me for the details of glicko2, but if you had I most certainly would have pointed you to Glickmans web pages.
The Glicko2 system is described on Glickmans web site. In fact it has been in the public domain and available since late 2001.

Also here is a maths example he provides on his web site.
http://www.glicko.com/glicko2.doc/example.html



Bill, the points I am making is clear for all to see. Our juniors play 28.25 games/yr. That's well within the 5-10/rating period range optimal for the operation of the Glicko system.
You can make that claim as much as you like it isnt true.


If your fixes don't solve the problem, you can't blame anybody else, the responsiblity rest with you. It's not our juniors fault, it's not our organisers fault, you can't keep pointing the finger.
I'll keep pointing the finger as long as clowns like you keep making false statements.



If the fixes solve the problem I will be the first to thank you. If not then I will be asking you for a solution.
Simple play more rated normal games.
Until then stop wasting my time.

Bill Gletsos
07-07-2004, 11:37 AM
Alrighty youse blokes...I've decided to form me own independent committee of mathematicians and sort this all out. We don't wanna see a ratings debate between youse for the next bloody 50 years you mongrels!
Look at it this way AR.

Kevin, Barry and I all disagree with David.


I know a genius who knows more maths than youse all put together: the 1minute_grandmaster.
Perhaps, but I suspect he knows less about all this than Glickman.

Garvinator
07-07-2004, 11:45 AM
Perhaps, but I suspect he knows less about all this than Glickman.
i would think that the person who knows most about the glicko system is the creator :whistle:

arosar
07-07-2004, 11:52 AM
Ha ha ha ha.
This is even better.
You are a complete moron.

I hate the way you behave sometimes Bill. It's quite unbecoming - especially for one in such high office.

Clearly, Dr Richards is no moron; he is an MD after all. At the very least statistically incompetent - according to you anyway.

Just a bit of upward feedback.

AR

Bill Gletsos
07-07-2004, 12:02 PM
I hate the way you behave sometimes Bill. It's quite unbecoming - especially for one in such high office.
I'm posting as the ratings officer not the NSWCA President or as an ACF Councillor. As such i'm not sure I would describe the ratings officer as "one in such a high office".
I just dont tolerate fools lightly but I do take your point.


Clearly, Dr Richards is no moron; he is an MD after all.
One can be intelligent in some areas and a moron in others. ;)



At the very least statistically incompetent - according to you anyway.
Its not just me who questions his stats ability but appears to be Kevin and Barry as well.

Anyaway it appears he is incompetent in other areas as well.
Although he has apparently been reading Glickmans web site he has not read Glickmans paper on the Glicko2 system even though its been on there for over 2 years.

And of course any google search on Glicko2 would have revealed the mathematical example document I mentioned.


Just a bit of upward feedback.
Thanks, I've noted it.

Garvinator
07-07-2004, 12:09 PM
I'm posting as the ratings officer not the NSWCA President or as an ACF Councillor. As such i'm not sure I would describe the ratings officer as "one in such a high office". in this matter i would say that the national rating officers are the highest position in the land ;)


Its not just me who questions his stats ability but appears to be Kevin and Barry as well.i am questioning his stats ability and the assumptions/claims david makes and i will be doing similiar stat analysis to david after the queensland open is finished, with about four tournaments data to use.

Cat
07-07-2004, 12:25 PM
I hate the way you behave sometimes Bill. It's quite unbecoming - especially for one in such high office.

Clearly, Dr Richards is no moron; he is an MD after all. At the very least statistically incompetent - according to you anyway.

Just a bit of upward feedback.

AR

I'm afraid it's an indictment of the way the ACF handles it's affairs. I do feel I've been a little harsh with Bill a times, but it's his total failure to acknowledge weaknesses in the system which leads to these wreckless, destructive dialogues. He argued the system was perfect 12/12 ago and he argues its perfect today, yet the ratings look very different, both can't be correct.

AR, none of us are morons as you well know, we are advocates for our communities, and it's his total lack of respect for the constuency he serves that is the most concerning aspect.

As for my statistical abilities, I managed A level Maths in the UK (including Statistics), equivalent to your OP 1. I've some experience in population dynamics with my Honours degree in Human Genetics. Bill's abilities are in boldest when he attempts to misrepresent statistical information, as he does repeatedly. I doubt he has any experience with population dynamics. His duplicity in presenting information undermines my confidence in his veracity.

Bill Gletsos
07-07-2004, 12:43 PM
I'm afraid it's an indictment of the way the ACF handles it's affairs. I do feel I've been a little harsh with Bill a times, but it's his total failure to acknowledge weaknesses in the system which leads to these wreckless, destructive dialogues. He argued the system was perfect 12/12 ago and he argues its perfect today, yet the ratings look very different, both can't be correct.
You keep saying this, but it isnt true.
As I have repeatedly said Graham and I dont claim it is perfect.
We just demand that any proposed change be proven to be beneficial prior to implementation and its potential impacts know.
None of yours have been proven to be based on anything other than your beliefs, beliefs not supported by others like Kevin and Barry.


AR, none of us are morons as you well know, we are advocates for our communities, and it's his total lack of respect for the constuency he serves that is the most concerning aspect.
With regards to the ratings area specifically I show respect to jenni and starter. I showed repect to Sue Kimura from the GC when I discussed the GC issues with her last year prior to the correction.
I show respect to those who deserve it.


As for my statistical abilities, I managed A level Maths in the UK (including Statistics), equivalent to your OP 1. I've some experience in population dynamics with my Honours degree in Human Genetics. Bill's abilities are in boldest when he attempts to misrepresent statistical information, as he does repeatedly. I doubt he has any experience with population dynamics. His duplicity in presenting information undermines my confidence in his veracity.
If you can get Kevin and Barry onside then I'll consider listening to you.
Until then all you are espousing are unsubstantiated pet theories.

Cat
07-07-2004, 01:12 PM
You keep saying this, but it isnt true.
As I have repeatedly said Graham abd I dont claim it is perfect.
We just demand that any proposed change be proven to be beneficial prior to implementation and its potential impacts know.
None of yours have been proven to be based on anything other than your beliefs, beliefs not supported by others like Kevin and Barry.

No Bill, not belief, but knowledge from outside your own narrow field. The things I describe are well known to professionals in the medical and child development fields. That you have no knowledge doesn't mean that the information doesn't exist.

There is nothing unique about chess, the essential learning process are the same as any other discpline, hence the interest in the educational benefits of chess. Isolationist approaches are ridiculous and counter-productive, stupid even.


With regards to the ratings area specifically I show respect to jenni and starter. I showed repect to Sue Kimura from the GC when I discussed the GC issues with her last year prior to the correction.
I show respect to those who deserve it.

You show respect to those who don't challenge your prejudices, Bill. You have major problems accepting information you don't like.


If you can get Kevin and Barry onside then I'll consider listening to you.
Until then all you are espousing are unsubstantiated pet theories.

You can hardly think I'm interested in winning a popularity contest. Things have been wrong, very wrong - they're getting better. Your excuses and denials have been a poor response to legitimate criticism. You litter the place with abuse and profanity, deliberately distort statistical information and you expect all this to be done unquestioningly. It won't wash, the only source of satisfaction will come from a demonstration that the position has improved and that the 'fixes' lead to longlasting improvement.

Garvinator
07-07-2004, 01:21 PM
There is nothing unique about chess, the essential learning process are the same as any other discpline, hence the interest in the educational benefits of chess.

David, chess does have one thing that is unique when talking about learning process and improvement as a person gets older and goes through puberty, no physical skill component.

Usually a persons skill level and knowledge improves cause they become more co-ordinated as they get older. Also concentration levels increase as they become more dedicated to their chosen sport.

In terms of chess, as i said, there is no physical skill component,so that is taken away. that leaves the increased concentration abilities and dedication levels. This increased potential is very much an individual characteristic and can not be measured in an across the board rating scheme/system.

Cat
07-07-2004, 01:48 PM
David, chess does have one thing that is unique when talking about learning process and improvement as a person gets older and goes through puberty, no physical skill component.

Usually a persons skill level and knowledge improves cause they become more co-ordinated as they get older. Also concentration levels increase as they become more dedicated to their chosen sport.

In terms of chess, as i said, there is no physical skill component,so that is taken away. that leaves the increased concentration abilities and dedication levels. This increased potential is very much an individual characteristic and can not be measured in an across the board rating scheme/system.

What about maths, english, reading, abstract thinking, visual-spacial skills, languages, need I go on? Garvin, please be careful.

Bill Gletsos
07-07-2004, 02:22 PM
No Bill, not belief, but knowledge from outside your own narrow field. The things I describe are well known to professionals in the medical and child development fields. That you have no knowledge doesn't mean that the information doesn't exist.

There is nothing unique about chess, the essential learning process are the same as any other discpline, hence the interest in the educational benefits of chess. Isolationist approaches are ridiculous and counter-productive, stupid even.
Kevin has already debated you at length on this. He has asked you to provide evidence that chess ability falls into this area. So far you have provided nothing to show that it does.




You show respect to those who don't challenge your prejudices, Bill. You have major problems accepting information you don't like.
No I have problems accepting unsupported theories.


You can hardly think I'm interested in winning a popularity contest.
Nice attempt at deflection but it has nothing to do with popularity. Its to do with proven ideas. So far you have provided no evidence to support your theories that they have any relevance to chess skill.


Things have been wrong, very wrong - they're getting better. Your excuses and denials have been a poor response to legitimate criticism.
You have provided no evidence what soever to support your theories regarding your maturation factor having any basis with regards chess playing ability.



You litter the place with abuse and profanity,
I would hardly call calling someone a moron, cretin, goose, joker or clown as profanity. If was wante dto be profane I'd take a leaf out of your mate Matt's repetoire.


deliberately distort statistical information and you expect all this to be done unquestioningly.
I have not distorted any statistical information.
I have simply provided the figures to show that your own selective figures were invalid.



It won't wash, the only source of satisfaction will come from a demonstration that the position has improved and that the 'fixes' lead to longlasting improvement.
The situation wont improve as long as you continue to rate minimal number of games in the normal system.
Until you address that issue, you can whistle dixie.

Bill Gletsos
07-07-2004, 02:24 PM
What about maths, english, reading, abstract thinking, visual-spacial skills, languages, need I go on? Garvin, please be careful.
The rating system works on game results.
If a junior is improving then this will be evidenced by his results.

If however the junior plays insufficient games then there is no means to tell if they are improving, lucky or unlucky.

Cat
07-07-2004, 03:23 PM
Kevin has already debated you at length on this. He has asked you to provide evidence that chess ability falls into this area. So far you have provided nothing to show that it does.




No I have problems accepting unsupported theories.

With the greatest respect to KB, he is hardly an authority on child development. Stop looking to Uncle K to answer your difficulties.

These are not unsupported theories, they are mainstream opinions about learning and development, of which you clearly know nothing. I suggest you
learn a little more before you continue with your ignorant comments.



Nice attempt at deflection but it has nothing to do with popularity. Its to do with proven ideas. So far you have provided no evidence to support your theories that they have any relevance to chess skill.


You have provided no evidence what soever to support your theories regarding your maturation factor having any basis with regards chess playing ability.

They are not my theories, its is mainstream understanding of learning an your ignorance of it that we're talking about.


I would hardly call calling someone a moron, cretin, goose, joker or clown as profanity

No sadly you wouldn't





The situation wont improve as long as you continue to rate minimal number of games in the normal system.


So you admit it, the system is failing despite the juniors playing an average of 28.25 games/yr!

Lucena
07-07-2004, 04:06 PM
I would hardly call calling someone a moron, cretin, goose, joker or clown as profanity.

He's right, profanity is words like **** and **** that have to be censored :D

Bill Gletsos
07-07-2004, 04:06 PM
With the greatest respect to KB, he is hardly an authority on child development. Stop looking to Uncle K to answer your difficulties.

These are not unsupported theories, they are mainstream opinions about learning and development, of which you clearly know nothing. I suggest you
learn a little more before you continue with your ignorant comments.

They are not my theories, its is mainstream understanding of learning an your ignorance of it that we're talking about.
What is obvious is that you can provide no evidence that this is true for chess ability.

Even if you could, the child's improvement in chess ability should be evidenced by their results.



So you admit it, the system is failing despite the juniors playing an average of 28.25 games/yr!
It should be obvious to anyone that 28.25 games/year(I make it 27.55 with a standard deviation of 19.28) is clearly inadequate to determine if a junior is rapidly improving or not.

Cat
07-07-2004, 05:31 PM
What is obvious is that you can provide no evidence that this is true for chess ability.

Even if you could, the child's improvement in chess ability should be evidenced by their results.



It should be obvious to anyone that 28.25 games/year(I make it 27.55 with a standard deviation of 19.28) is clearly inadequate to determine if a junior is rapidly improving or not.

OK, we'll see. As I say things have improved enormously. My main concern was to dispell the myth that our juniors play an inadequate number of rating games and I'm happy to settle on 27.55. You may feel thats inadequate - I say its substantial, more than enough to earn the right, to expect fair treatment. It's no good crying poor, this is your constituency. Maybe things will go well, lets hope so.

Bill Gletsos
07-07-2004, 05:55 PM
OK, we'll see. As I say things have improved enormously. My main concern was to dispell the myth that our juniors play an inadequate number of rating games and I'm happy to settle on 27.55. You may feel thats inadequate - I say its substantial, more than enough to earn the right, to expect fair treatment.
Ok dont listen, but if you are basing this on Glickman's comment re 5-10 games per rating period then you are wasting your time as what is a good figure is dependant on a number of variables.
Btw in his Glicko2 document he recommends 10-15 games per period.
Based on 4 periods in a year thats 40-60 games. Your average is not in the ball park.

Of course as Glickman explained to us what constitutes good numbers are dependant on the system parameters.

Anyway dont come complaining to me until they lift there number of normal rated games.


It's no good crying poor, this is your constituency. Maybe things will go well, lets hope so.
If its my constituency then you should heed my recommendations.
Ignore it if you wish as I will then ignore any of your complaints.

Kevin Bonham
07-07-2004, 06:42 PM
Well, why don't you try reading Bayes' theorum?

Why don't you try understanding it? (You ask as if you assume I've never heard of it, what a joke.)

Bayes' Theorem is a well-known approach to probability theory. One form it is given in is as follows:

The probability of a hypothesis H conditional on a given body of data E is the ratio of the unconditional probability of the conjunction of the hypothesis with the data to the unconditional probability of the data alone.

Now, precisely where in there do you find any assumption about players being equally likely to play each other? It is simply not required at all. The whole edifice of probability theory as applied to chess ratings does not collapse into a smouldering ruin if a 6 year old beginner isn't allowed to play Ian Rogers.

I don't think you even understand the Glicko system well enough to even explain how Bayesian methods are used in it.


KB, I'm not interested in your silly games.

... because you're far too busy even playing your own to admit that mine are not silly. I am just pointing out that your case is a mess of rubbish taken out of context or extended to fields where it has not been shown to apply, in which what little data you are willing to put up is typically outdated and misinterpreted. As usual this concoction comes with completely irrelevant topic-drifts, and whenever you are asked a question that would give you an opportunity to show some understanding, you don't answer it.

In short, you have no clue. :hand:


I'm interested to see if the corrections have been adequate, and that the improvements can be maintained.

Well just bring on the raw, dry data and stop all the hot air and unsound theoretical waffle. :wall:


He argued the system was perfect 12/12 ago and he argues its perfect today, yet the ratings look very different, both can't be correct.

Where is your evidence that he argued the system was perfect 12 months ago (I assume that's the meaning of that peculiar American notation)? I doubt he said any such thing.


No Bill, not belief, but knowledge from outside your own narrow field. The things I describe are well known to professionals in the medical and child development fields. That you have no knowledge doesn't mean that the information doesn't exist.

There is nothing unique about chess, the essential learning process are the same as any other discpline, hence the interest in the educational benefits of chess. Isolationist approaches are ridiculous and counter-productive, stupid even.

I mentioned one difference before - different motivations. You simply cannot compare education in the schoolyard (something that children generally don't enjoy and only do because they are compelled) with education in a recreational field such as chess. You also haven't answered my comment about how your comments about chess ratings would be applicable to things like assessment methods if chess and other forms of learning were truly compatible.


With the greatest respect to KB, he is hardly an authority on child development.

Indeed, but since you supposedly know far more than me about this field you should be able to dispose of my methodical concerns easily. Instead you have been repeatedly unable to, simply hand-waving the same nonsense over and over again without advancing the argument. It sounds like you are the one who is trying to apply knowledge from an isolated field without having any idea of the issues involved in generalising your findings, and I am the one who is making use of a particular generalist skill one builds up over long enough in the university system - the ability to detect and dispose of obvious BS in any field.

Kevin Bonham
07-07-2004, 07:00 PM
Clearly, Dr Richards is no moron; he is an MD after all. At the very least statistically incompetent - according to you anyway.

Just a bit of upward feedback.

Sure he's an MD but that doesn't guarantee he knows what he's talking about in any field other than medicine. He also frequently hasn't had a clue on non-medicine topics in the off-topic section. Unfortunately he's typical of a small minority of MDs who seem to think their community status means they have to be respected no matter what they speak out about - like the Doctors for Forests type groups who are always speaking out on environmental issues they know nothing about.

David has long subjected Bill to some pretty patronising criticism even though his (David's) arguments have been total nonsense at least 90% of the time. If he can't take it he shouldn't dish it out, and if he can take it, what's the problem?

David has also been quite tiresome in his habit of continuing to repeat nonsense after it has been repeatedly discredited.

This message was brought to you by everyone's favourite lobby group, Ecologists for Euthanasia. :cool:

Cat
08-07-2004, 10:54 PM
Ok dont listen, but if you are basing this on Glickman's comment re 5-10 games per rating period then you are wasting your time as what is a good figure is dependant on a number of variables.
Btw in his Glicko2 document he recommends 10-15 games per period.
Based on 4 periods in a year thats 40-60 games. Your average is not in the ball park.

Of course as Glickman explained to us what constitutes good numbers are dependant on the system parameters.

Anyway dont come complaining to me until they lift there number of normal rated games.


If its my constituency then you should heed my recommendations.
Ignore it if you wish as I will then ignore any of your complaints.

Great, after 15 months of denial and disagreement, at last some consensus and acknowledgement.

We can agree;

1. That the average junior on the Gold Coast plays rated 27.55 games/year.

2. That the Glicko 2 system cannot adequately describe their rating change, given their level of participation.

Your solution is that our juniors should simply play more rating games and it's our responsibility as organisers to ensure this happens.

However, our juniors already play a rating game more than once a fortnight - that's one hell of a commitment. Their chess interest has to compete with schoolwork, other sports, music commitments, family commitment and one or two of them may just have a social life - you never know.

Bill, your suggestion is impractical, it will never happen in a month of Sundays. It is entirely unreasonable that we demand any more commitment from our juniors. We can encourage, yes, but those kinds of demands to sustain a rating are totally ridiculous.

The Gold Coast Chess Club pays good money to the ACF for our rating service. You are telling us that the Glicko 2 system cannot adequately support movements in our junior ratings, given the level of participation, and we're just going to have to lump it. You are the service provider, you know its inadequate, but you have absolutely no intention of doing anything about it.

Well I'm telling you that is unacceptable. It's an insult to the efforts of all our juniors. We have every right to demand better service, and the ACF has a responsibility to ensure the service is properly delivered.

I know you cannot find a way of solving the problem in real time, I understand that. In that case what you must do, what your responsibility demands, is that you present us with an undertaking that our ratings will be monitored on a regular basis, that they will be cross-referenced with other regions, to ensure the mess that has occurred doesn't happen again.


For now the ratings seem fine, though I will be checking the information from the recent tournaments. I will await news of your strategy to provide a reliable rating service for our region into the future.

Bill Gletsos
09-07-2004, 12:41 AM
Great, after 15 months of denial and disagreement, at last some consensus and acknowledgement.
No concensus at all.
And certainly no acknowledgement.
I've been telling you to play more normal games.
I'm still telling you that.


We can agree;

1. That the average junior on the Gold Coast plays rated 27.55 games/year.
An irrelevant stat on your part.


2. That the Glicko 2 system cannot adequately describe their rating change, given their level of participation.
I'm not saying that at all you clown.
I'm saying based on the number of games they play their rating is as accurate as possible. In fact I can find no evidence that any existing rating system would handle it better.
I've been telling you since last year they need to play more games.
You have been stubborn and refused to do anything about it.


Your solution is that our juniors should simply play more rating games and it's our responsibility as organisers to ensure this happens.
Finally you got something right.


However, our juniors already play a rating game more than once a fortnight - that's one hell of a commitment. Their chess interest has to compete with schoolwork, other sports, music commitments, family commitment and one or two of them may just have a social life - you never know.
Rubbish.
You were the dope who hung is hat on Glickman's 5-10 games per period for the Glicko, even though I tried to tell you it depended on other factors.

I originally complained that the majority of your juniors play less than 30 normal games in a 12mth period.
In fact 57% of your active juniors played less than 30 normal games in the past 12mths.
You only have 6 players(13.6%) who play 50 or more and 2 (0.04%) who played 60 or more.
Compare this to the ACT.
Only 47% of the ACT juniors played less than 30 normal games in the past 12mths.
They had 17(23%) play 50 or more games and 15(20%) who play 60 or more.



Bill, your suggestion is impractical, it will never happen in a month of Sundays. It is entirely unreasonable that we demand any more commitment from our juniors. We can encourage, yes, but those kinds of demands to sustain a rating are totally ridiculous.
It isnt totally ridiculous for the ACT juniors.
Should I therefore assume from the figures and your statement that obviously the ACT juniors are just more committed than the GC juniors.


The Gold Coast Chess Club pays good money to the ACF for our rating service.
Yes and you get the service you pay for.


You are telling us that the Glicko 2 system cannot adequately support movements in our junior ratings, given the level of participation, and we're just going to have to lump it.
No you idiot.
I'm telling you the Glicko2 system handles your juniors perfectly based on their results.
Its not the systems fault.
Its your juniors lack of normal rated games.



You are the service provider, you know its inadequate, but you have absolutely no intention of doing anything about it.
The only thing inadequate is you.
The problem is not with the system.
The problem is with their lack of games.


Well I'm telling you that is unacceptable. It's an insult to the efforts of all our juniors. We have every right to demand better service, and the ACF has a responsibility to ensure the service is properly delivered.
Well I'm telling you, the service is fine.
Your lack of normal rated games is the problem not the system.


I know you cannot find a way of solving the problem in real time, I understand that.
You have repeatedly demonstrated you understand nothing.


In that case what you must do, what your responsibility demands, is that you present us with an undertaking that our ratings will be monitored on a regular basis, that they will be cross-referenced with other regions, to ensure the mess that has occurred doesn't happen again.
No it is you who have a responsability to ensure your juniots have the opportunity to play more normal rated games.



For now the ratings seem fine, though I will be checking the information from the recent tournaments. I will await news of your strategy to provide a reliable rating service for our region into the future.
You can wait till hell freezes over for all I care about what you want. :whistle:

Kevin Bonham
09-07-2004, 01:04 AM
It isnt totally ridiculous for the ACT juniors.
Should I therefore assume from the figures and your statement that obviously the ACT juniors are just more committed than the GC juniors.

Maybe David's real problem is the common one elsewhere - not enough organisers, less supply of chess than player demand for it.

Maybe he could do something about it with some of the time he wastes repeating the same old claims here.

Garvinator
09-07-2004, 01:47 AM
im starting to wonder, how many different ways can you say, play more normal rated games to get more accurate ratings :wall:

Alan Shore
09-07-2004, 02:34 AM
No concensus at all.
And certainly no acknowledgement.
I've been telling you to play more normal games.
I'm still telling you that.

But they won't and you telling them to for the sake of the rating system won't change that fact.



No it is you who have a responsability to ensure your juniots have the opportunity to play more normal rated games.


The opportunity certainly exists. What you can't ignore is the reluctance of juniors to participate given such opportunity.



The only thing inadequate is you.
The problem is not with the system.
The problem is with their lack of games.

Not fair calling David inadequate.. what we'd like is a straight answer. I will maintain there is no problem with the rating system and be on your side completely Bill, if you can show David that your rating system can sufficiently handle juniors who do not have the propensity to play as many normal games as you personally would like them to. I have enough faith in Glicko but your continual use of the argument 'more normal games are required' is concerning. It would be more comforting for you to simply say 'There is no problem, despite a tendency for juniors to not play as many normal rated games, yet their ratings are indeed accurate enough to describe their past performances and give some indication of their playing strength'. If you can admit this, perfect. However, if you continue to squarely place the blame on juniors for not playing many normal games and say 'this does adversely affect their rating insofar it is not a good enough indication of their playing strength' then you have a problem. So, is it column A, or B?

Alan Shore
09-07-2004, 02:35 AM
im starting to wonder, how many different ways can you say, play more normal rated games to get more accurate ratings :wall:

Tell it to the juniors who don't want to spend an hour playing a game of chess. Not everyone has such a penchant for slow-thinking activities particularly in this day and age. :hand:

Garvinator
09-07-2004, 02:38 AM
Tell it to the juniors who don't want to spend an hour playing a game of chess. Not everyone has such a penchant for slow-thinking activities particularly in this day and age. :hand:
and all that means is that they will have accurate rapid ratings. The only time normal ratings matter are for normal rated tournaments.

Alan Shore
09-07-2004, 02:47 AM
and all that means is that they will have accurate rapid ratings. The only time normal ratings matter are for normal rated tournaments.

Too right but then how accurate are the rapid ratings of say, Victorians, Taswegians and those from WA? Not very, is my guess if you're going by Bill's insistence that more games need to be played.

Garvinator
09-07-2004, 02:52 AM
Too right but then how accurate are the rapid ratings of say, Victorians, Taswegians and those from WA? Not very, is my guess if you're going by Bill's insistence that more games need to be played.
well i do think bill is right on the fact that more games need to be played. More games more accurate results.

Alan Shore
09-07-2004, 02:57 AM
well i do think bill is right on the fact that more games need to be played. More games more accurate results.

That should be painfully obvious, yet this isn't a perfect world and the ratings need to work in real conditions not just laboratory ones. Therefore I put the question to Bill of whether the ratings do work with juniors not playing as many games as he'd like. Now no more replies gg, or my original long post will get buried and I want Bill and David to see it :p

Bill Gletsos
09-07-2004, 09:46 AM
Not fair calling David inadequate.. what we'd like is a straight answer.
I keep giving him straight answers, he keeps ignoringthem.


I will maintain there is no problem with the rating system and be on your side completely Bill, if you can show David that your rating system can sufficiently handle juniors who do not have the propensity to play as many normal games as you personally would like them to. I have enough faith in Glicko but your continual use of the argument 'more normal games are required' is concerning. It would be more comforting for you to simply say 'There is no problem, despite a tendency for juniors to not play as many normal rated games, yet their ratings are indeed accurate enough to describe their past performances and give some indication of their playing strength'. If you can admit this, perfect. However, if you continue to squarely place the blame on juniors for not playing many normal games and say 'this does adversely affect their rating insofar it is not a good enough indication of their playing strength' then you have a problem. So, is it column A, or B?
It is fairly simple.
Assume a maths exam to determine how well a child knows maths.
You can ask them 3 questions or 20 questions.
The accuracy of your assessment of their knowledge is much greater with 20 than 3.
With 20 questions you can determine how well they know their maths.
With 3 questions you cannot with any degree of certainty as the error margin is far greater.
The same is true with regards the number of rated games.

As with the maths example the answer is to ask more questions, with chess its play more games.

Cat
09-07-2004, 01:16 PM
I originally complained that the majority of your juniors play less than 30 normal games in a 12mth period.
In fact 57% of your active juniors played less than 30 normal games in the past 12mths.
You only have 6 players(13.6%) who play 50 or more and 2 (0.04%) who played 60 or more.
Compare this to the ACT.
Only 47% of the ACT juniors played less than 30 normal games in the past 12mths.
They had 17(23%) play 50 or more games and 15(20%) who play 60 or more.







want. :whistle:

What's perfectly clear to me from what you've said, Bill, is that you've known there's been a problem with the junior ratings in the Glicko system all along, despite your endless denials. 15 months ago you could have said there was a systemic problem in the junior ratings, because the Glicko system requires far higher participation rates than we are achieving, yet you choose to be less than frank with the Australian chess public.

If you had openly admitted participation rates were inadequate for the proper operation of the Glicko system, then the problem could have started to have been addressed there & then, but you choose instead to live in denial.

Bill, what percentage of the junior chess players nationally are reaching the required participation rates? By you own figures, even in the ACT around 70% of juniors are not reaching the required participation rates - this is the percentage of juniors who are being potentially under-rated. But the situation is worse than that, because if they are chronically under-rated then they are likely to be corrupting the ratings of those that are reaching the reuqired participation rate.

Bill what is the the national junior participation rate?

Bill, 15 months ago when you said there wasn't a problem with the junior ratings, you knew that wasn't true becuase you knew the national participation rates were woefully inadequate for the successful operation of the Glicko system in Australia.

When you told me 'the Glicko can handle it', you knew that wasn't true, for the same reason.

When you said that there were no regional problems in our area cause by the Glicko system, you knew that wasn't true, because you knew that our participation rates were far too low.

For the last 15 months you have been decieving the Australian Chess public that all was well in the Glicko garden, when you knew full well that juniors were being systematically underrated in the Glicko system because they were not achieiving, nor will they ever likely achieve the required participation rates.

Bill, most of the Chess public live in the real world, not the Glicko world. We want real solutions to our very real problems. We are never going to get the required participation rates while we continually let our junior players down. This admission after 15 months of denial and lies is nothing short of a disaster for Australia Chess, and says everything about how the ACF chooses to conduct itself. No more lies, no more denial, its time the Australian public were treated to the truth!

rob
09-07-2004, 03:53 PM
Bill, most of the Chess public live in the real world, not the Glicko world. We want real solutions to our very real problems. We are never going to get the required participation rates while we continually let our junior players down. This admission after 15 months of denial and lies is nothing short of a disaster for Australia Chess, and says everything about how the ACF chooses to conduct itself. No more lies, no more denial, its time the Australian public were treated to the truth!

David, you are concerned about a few juniors in one state. There are seven states/territories with thousands of juniors - the GC juniors are a tiny minority (not the only tiny thing around here). Why are you the only one complaining? Why should we change the rating system to fit in with a few juniors who play a small number of 1 hour rated games against each other? It is obvious to everyone that the more rated games against a variety of players leads to more accurate ratings. Some states have separate junior rating systems, plus rapid rating, you could even create your own rating system if it will make you and the kids 'feel good'. The more effort that you put into something (playing more rated games) the more reward (more accurate) that you get out of it. More information in (rated games) more certain results (accurate rating). How many times, in how many different ways, can so many ppl tell you stuff before you get it?

Perhaps if I suggested that each junior receives 2 rating points for every rated game you would be in favour (incentive to play, would correspond with natural improvement)?

Garvinator
09-07-2004, 04:22 PM
Perhaps if I suggested that each junior receives 2 rating points for every rated game you would be in favour (incentive to play, would correspond with natural improvement)?
please dont :hand:

rob
09-07-2004, 04:27 PM
Too right but then how accurate are the rapid ratings of say, Victorians, Taswegians and those from WA? Not very, is my guess if you're going by Bill's insistence that more games need to be played.

Not at all is my guess as there are no WA rapid ratings. Bad example :)

Garvinator
09-07-2004, 04:28 PM
Not at all is my guess as there are no WA rapid ratings. Bad example :)
well when there are no rapid ratings, david cant complain that those ratings are wrong :whistle: ;)

rob
09-07-2004, 04:32 PM
please dont :hand:
Absolutely not :)

Bill Gletsos
09-07-2004, 07:12 PM
What's perfectly clear to me from what you've said, Bill, is that you've known there's been a problem with the junior ratings in the Glicko system all along, despite your endless denials. 15 months ago you could have said there was a systemic problem in the junior ratings, because the Glicko system requires far higher participation rates than we are achieving, yet you choose to be less than frank with the Australian chess public.
Total misrepresentation of the facts as usual.
There is no systemic problem at all.
The ratings of players are as accurate as can be based on their results.
Any fool should be able to see that the more results you have the more statistically valid you are.



If you had openly admitted participation rates were inadequate for the proper operation of the Glicko system, then the problem could have started to have been addressed there & then, but you choose instead to live in denial.
The Glicko system works fine.
All I am saying is if you want more accurate ratings play more games.
If you cannot see that then you are dumber than you have led me to believe. :hand:


Bill, what percentage of the junior chess players nationally are reaching the required participation rates? By you own figures, even in the ACT around 70% of juniors are not reaching the required participation rates
You are a fool.
53% of the ACT juniors play more than 30 games a year which is the only rate I have been mentioning.



- this is the percentage of juniors who are being potentially under-rated. But the situation is worse than that, because if they are chronically under-rated then they are likely to be corrupting the ratings of those that are reaching the reuqired participation rate.
Thats just rubbish.
I keep telling you that the number of games required is based on a number of factors.



Bill what is the the national junior participation rate?
I'm not wasting my time any further with you on this.
Play more games or dont play more games it up to you.
Just stop wasting my time.


Bill, 15 months ago when you said there wasn't a problem with the junior ratings, you knew that wasn't true becuase you knew the national participation rates were woefully inadequate for the successful operation of the Glicko system in Australia.
Incorrect.
As far as I am concerned the system is working fine for the majority of dedicated juniors.
If they cannot play 30 games a year they can hardly be called dedicated.


When you told me 'the Glicko can handle it', you knew that wasn't true, for the same reason.
I stand by what I said, not what you are trying to twist it into.
I said the Glciko2 can handle juniors.
For the majority this is true.
However if juniors play very few games then that isnt a failure of the system.


When you said that there were no regional problems in our area cause by the Glicko system, you knew that wasn't true, because you knew that our participation rates were far too low.
More rubbish.
Your juniors lack of games is not a Glicko system issue or even a ELO issue.
It is about accuracy.
I have been telling you since last year that you needed to increase the number of normal games.


For the last 15 months you have been decieving the Australian Chess public that all was well in the Glicko garden, when you knew full well that juniors were being systematically underrated in the Glicko system because they were not achieiving, nor will they ever likely achieve the required participation rates.
Completely incorrect.

A player is underrated or in fact overrated when their rating does not reflect their ability based on their results.
This is not the case.
The players are correctly rated based on their results.


Bill, most of the Chess public live in the real world, not the Glicko world. We want real solutions to our very real problems. We are never going to get the required participation rates while we continually let our junior players down.{/quote]
Rubbish.
The ACT juniors clearly dont have your problem.


[QUOTE=David_Richards]This admission after 15 months of denial and lies is nothing short of a disaster for Australia Chess, and says everything about how the ACF chooses to conduct itself.
The only lies I have seen is your continual distortion of the facts.


No more lies, no more denial, its time the Australian public were treated to the truth!
You have been getting the truth.
The problem is YOU cant handle the truth.

But just for you I'll repeat it one more time.

A player is underrated or in fact overrated when their rating does not reflect their ability based on their results.
This is not the case.
As far as we are concerned the players are correctly rated based on their results.

So my advice to you is simple Doc.

Play more normal games
Play more normal games
Play more normal games.

Consider it a mantra.
Eventually it should even sink into your thick skull. :whistle:

Bill Gletsos
09-07-2004, 07:16 PM
David, you are concerned about a few juniors in one state. There are seven states/territories with thousands of juniors - the GC juniors are a tiny minority (not the only tiny thing around here). Why are you the only one complaining? Why should we change the rating system to fit in with a few juniors who play a small number of 1 hour rated games against each other? It is obvious to everyone that the more rated games against a variety of players leads to more accurate ratings. Some states have separate junior rating systems, plus rapid rating, you could even create your own rating system if it will make you and the kids 'feel good'. The more effort that you put into something (playing more rated games) the more reward (more accurate) that you get out of it. More information in (rated games) more certain results (accurate rating). How many times, in how many different ways, can so many ppl tell you stuff before you get it?
4 million 3 thousand 2 hundred and twenty perhaps?


Perhaps if I suggested that each junior receives 2 rating points for every rated game you would be in favour (incentive to play, would correspond with natural improvement)?
Ah yes the infamous USCF fiddle points debacle. :whistle:

Bill Gletsos
09-07-2004, 07:19 PM
Tell it to the juniors who don't want to spend an hour playing a game of chess. Not everyone has such a penchant for slow-thinking activities particularly in this day and age. :hand:
GG was right.
However he should have ended it with "to get more accurate normal ratings".

Bill Gletsos
09-07-2004, 07:22 PM
Too right but then how accurate are the rapid ratings of say, Victorians, Taswegians and those from WA? Not very, is my guess if you're going by Bill's insistence that more games need to be played.
Well the majority of adults as far as I can tell could not care less about their rapid ratings.

QLD is the only state that is rating lots of rapid events and even there they are mostly junior events.

Kevin Bonham
09-07-2004, 07:31 PM
David,

I have some questions for you.

1. What are your criteria for determining whether the ratings of juniors in your area are "working" or not?

2. A player plays ten games in a three month period and gets a rating of 900. In their next rating period they play another ten games, in which they perform at 1100. (You can assume the reliabilities and volatilities of each group of opponents are the same). Assuming you don't know their age, what do you think their new rating should be? Assuming you do know their age, how would that affect your answer?

(Bonus points to anyone who guesses why I am asking David these specific questions.)

skip to my lou
09-07-2004, 07:33 PM
(Bonus points to anyone who guesses why I am asking David these specific questions.)

To end this endless argument.

Apparently, Glicko is fine with Dave. He said there needs to be some changes to the current system. What I would be interested in is some real details and instructions to modify the current system, not endless waffle.

What is happening now:


for ($i = 1; $i <= ∞; $i++) {
daveComments();
}

function daveComments(){
$davesComment = "Rating system sucks";
billDenies();
}

function billDenies($davesComment){
echo "HELL NO.";
}

Garvinator
09-07-2004, 07:35 PM
GG was right.
However he should have ended it with "to get more accurate normal ratings".
umm ill make a small assumption that you are saying that david is a dimwit in this area :whistle: cause my comment was accurate in terms of whether you are talking about rapid or normal ratings. But still point taken, david may not recognise the difference, so ill adjust my comment:

ggrayggray said:


well i do think bill is right on the fact that more normal games need to be played. More normal games more accurate normal results.

arosar
09-07-2004, 07:42 PM
To end this endless argument.

Apparently, Glicko is fine with Dave. He said there needs to be some changes to the current system. What I would be interested in is some real details and instructions to modify the current system, not endless waffle.

What is happening now:


for ($i = 1; $i <= ∞; $i++) {
daveComments();
}

function daveComments(){
$davesComment = "Rating system sucks";
billDenies();
}

function billDenies($davesComment){
echo "HELL NO.";
}

//code updated - Jeo

So what is that...like an infinite loop or something?

AR

skip to my lou
09-07-2004, 07:46 PM
Pretty much, ∞ = Infinity. It is not processing it in the PHP Code box for code display reasons :)

Kevin Bonham
09-07-2004, 07:49 PM
To end this endless argument.

Where David's arguments are concerned, defeat has never been the end before. :lol:

skip to my lou
09-07-2004, 07:56 PM
Damn, I wrote another 10 lines of code, stepped away from the comp for 10 seconds and it dissappeared!! :( I made an array of possible output from Bill and then put a random generator to pick one and display it. Too lazy to write it again.

Bill Gletsos
09-07-2004, 08:10 PM
umm ill make a small assumption that you are saying that david is a dimwit in this area :whistle: cause my comment was accurate in terms of whether you are talking about rapid or normal ratings. But still point taken, david may not recognise the difference, so ill adjust my comment:

ggrayggray said:
Actually I wasnt saying David was a dimwit in that particular circumstance because my reply was to Bruce who was commenting on your statement.
Also before you now claim I implied Bruce was a dimwit that also isnt true. :hand:

skip to my lou
09-07-2004, 08:12 PM
What is the difference between a dimwit and a fool?

Bill Gletsos
09-07-2004, 08:14 PM
Damn, I wrote another 10 lines of code, stepped away from the comp for 10 seconds and it dissappeared!! :( I made an array of possible output from Bill and then put a random generator to pick one and display it.
I hope you included the words rubbish, you have no clue and the nouns moron, cretin, clown, joker and of course goose. ;)

Bill Gletsos
09-07-2004, 08:15 PM
What is the difference between a dimwit and a fool?
Not sure but perhaps a dimwit cannot help it where as a fool actually works at it till he perfects it. :owned:

skip to my lou
09-07-2004, 08:17 PM
I included everything except joker. Instead of joker I used fool, dipstick,.,.

Bill Gletsos
09-07-2004, 08:21 PM
I included everything except joker. Instead of joker I used fool, dipstick,.,.
Good choice.
I've only used joker a couple of times lately.

Alan Shore
09-07-2004, 08:32 PM
It is fairly simple.
Assume a maths exam to determine how well a child knows maths.
You can ask them 3 questions or 20 questions.
The accuracy of your assessment of their knowledge is much greater with 20 than 3.
With 20 questions you can determine how well they know their maths.
With 3 questions you cannot with any degree of certainty as the error margin is far greater.
The same is true with regards the number of rated games.

As with the maths example the answer is to ask more questions, with chess its play more games.

Naturally, this should be obvious to anyone. I've completed third-year research methodology and statistics so no need to dumb it down. 3 questions (games) will not be statistically significant and have poor validity. However, if there is no room for any more questions (normal rated games) it will remain the only method of determination we have. The question should then be posed, is the Glicko the system that best handles rated games even at this low number of games. This was my real question. I would like to think it is, with David as of yet failing to provide a superior alternative. Therefore Bill, do you agree Glicko is the best way, even with such a low number of games?

Alan Shore
09-07-2004, 08:34 PM
Not at all is my guess as there are no WA rapid ratings. Bad example :)

Shocking stuff... it appears dinosaurs still do exist over in desert country! :D

Alan Shore
09-07-2004, 08:36 PM
Well the majority of adults as far as I can tell could not care less about their rapid ratings.

QLD is the only state that is rating lots of rapid events and even there they are mostly junior events.

I'm the opposite, I couldn't care less about my normal rating, yet my rapid rating more clearly reflects my strength at a time control I am well suited to and play the overwhelming proportion of. A large number of my fellow Queenslanders seem to follow suit.