PDA

View Full Version : Do you think the precise RD and Volatilities should be published with the ratings?



Rincewind
02-09-2005, 11:31 PM
Conversation from this evening's shoutbox (below) begs the question...

Is the system scientific?
Is it free of subjective influence ?

I think it is scientific..... It has a huge formula from memory.....

And an Rd factor that is not public...which makes the system unrepeatable

Why is it everytime someone doesnt like the way things work out they make scurrilous accustions

No accusations have been made Bill. Mischa is questioning the repeatability of the system as not all the information is published.

I am not accusing anyone
just asking
i wish to understand is all

Of course she is accusing someone. She just said "Is it free of subjective influence ?"

The Glicko calcs are (as you know) quite sensitive to variations within the precision of published RDs

a reasonable question

you have stipulated enough that the rating system is only required to be accurate, not explainable. Given that this is sufficient for you to feel happy with it, it seems that you would not be troubled by people wanting to understand the system, just tell them it is accurate
you = bill

macavity I suggest you see post http://www.chesschat.org/showthread.php?t=737&page=1 which was implemented last year

Bereaved
02-09-2005, 11:46 PM
Planned Rating Changes
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.

Hi everyone, here is the post that Bill referred to in the above thread reference.

The questions that arise for me;

Is it a truly easy thing to utilize the factors above, for say one of us?
Are they determined on a similiar mathematical formula as the glicko system proper?
If not, who is able to determine if someone is truly improving rapidly; surely there is not a light on the computer that starts to glow??

Welcome other people's ideas in this matter,

take care and God Bless, Macavity

Bill Gletsos
02-09-2005, 11:59 PM
The questions that arise for me;

Is it a truly easy thing to utilize the factors above, for say one of us?No.

Are they determined on a similiar mathematical formula as the glicko system proper?Yes.

Bereaved
03-09-2005, 12:13 AM
Bill, your proclivity for verbosity lets you down here. does no mean; few would have the mathematical ability; few could be bothered; there is immense possibilities for calculation errors, which could lead to arguments?
And as far as the Yes goes, where are the equations shown to calculate these values specific, or how to calculate these values in a generalised sense? Are they at Glicko's site? could you show us the url again?
Take care and God Bless, Macavity

Rincewind
03-09-2005, 12:18 AM
http://math.bu.edu/INDIVIDUAL/mg/glicko/index.html

Bill Gletsos
03-09-2005, 12:28 AM
Bill, your proclivity for verbosity lets you down here.I was just trying to save myself from repeating what had been discussed previously both here and on the old ACF board.

does no mean; few would have the mathematical ability; few could be bothered; there is immense possibilities for calculation errors, which could lead to arguments?Pretty much yes to all the above.
There is no way for anyone wishing to do the calculation to know what is the intermediate ratings of their opponents. Players also have no idea what rating is used for the unrated players. As such any calculation by someone is always only going to be an approximation.

And as far as the Yes goes, where are the equations shown to calculate these values specific, or how to calculate these values in a generalised sense? Are they at Glicko's site? could you show us the url again?www.glicko.com is down and has been for a few months however you can see the contents describing Glicko and Glicko2 at Professor Glickmans page at http://math.bu.edu/INDIVIDUAL/mg/glicko/

antichrist
03-09-2005, 12:33 AM
sorry slightly off topic but I thought NSWJCL had inaccuracies in their ratings a few years back. One player played a higher board all season, won all games, did brilliantly in big tourney as compared to a lower board player in same team who lost some games, and "failed" in big tourney - yet lower player overtook higher player. I could not think of other tourneys lower player may have entered into.

Bereaved
03-09-2005, 12:36 AM
Probably better to know about all the events played first, A/C,

Take care and God Bless, Malcolm

antichrist
03-09-2005, 12:39 AM
Probably better to know about all the events played first, A/C,

Take care and God Bless, Malcolm

of course yes, but the higher player had won his age group trophy beating 3 players about 300 points higher than himself, the lower player only won 2 low games I think. They were both my students so I knew about them, me going to maybe all NSWJCL comps back in those days.

Rincewind
03-09-2005, 12:41 AM
I was just trying to save myself from repeating what had been discussed previously both here and on the old ACF board.
Pretty much yes to all the above.
There is no way for anyone wishing to do the calculation to know what is the intermediate ratings of their opponents. Players also have no idea what rating is used for the unrated players. As such any calculation by someone is always only going to be an approximation.
www.glicko.com is down and has been for a few months however you can see the contents describing Glicko and Glicko2 at Professor Glickmans page at http://math.bu.edu/INDIVIDUAL/mg/glicko/

In theory the system is closed. All information comes from the either the master lists with the precise RDs and volatilities and the tournament results. Therefore, if someone was in possession of both of these sources of information, they could repeat your calculations and come up with exactly the same results, 100% of the time, guaranteed.

As the RDs are not published in a precise form and Volatilities are not published at all, someone in possession of the ACF rating master file (a misnomer) and all the tournament results could not reliably repeat the calculations as they do not have all the information. The system is not repeatable and therefore not scientific.

The advantages of a repeatable system is that it is transparent and defendable from "scurrilous accusations". The present system, due to its own opacity is not patently defendable in this way.

The fact that the intermediate ratings are not able to be generated easily is a tactical issue and yes reduces the effectiveness of having the precise RDs available. However, the system is, in principle, repeatable and further, more accurate estimates would be able to be made with precise RDs and Volatilities than would be possible without them.

PHAT
03-09-2005, 01:17 AM
Why is it everytime someone doesnt like the way things work out they make scurrilous accustions

:boohoo:

Two years ago I was saying a similar thing as Coxy is now. If the the system is too complicated ("opaque"), people dont' trust it.

:wall:

Garvinator
03-09-2005, 01:29 AM
I have voted no, the reason being is that I believe most chess players will just be more confused by more numbers. This will also give players something else to complain about with regards to ratings.

I think very few if any would actually bother to sit down and work out some ratings accurately if given the 'tools' to do so.

Publishing more numbers would just lead to more complaints and more supposedly unhappy people, regardless of the facts.

Rincewind
03-09-2005, 08:28 AM
:boohoo:

Two years ago I was saying a similar thing as Coxy is now. If the the system is too complicated ("opaque"), people dont' trust it.

:wall:

Actually my position is not that the system is too complicated, just that all the factors should be in the public domain. The complexity of the system is a necessary artifact of its design to model a complicated physical system.

To overcome the complexity issues, open source tools can be provided to ease the burden on the individual having to develop or even understand their function.

My vision for the ACF ratings page would be to have the current ratings available as well as an estimated new rating which is updated with every swiss perfect file which comes through to the ACF during the period. A report breaking down each rating change game by game and reports available on a tournament by tournament basis as well. In addition to that tools like my calculator which can provide hypothetical rating estimates or estimates for those for tournaments still not yet received. And finally a graphing feature like that developed by (I think) Lee Jones. In short, a one stop shop with everything a player, organisor, journalist, archivist or stats major might need to know about ratings.

Remember this is a vision statement and so I don't expect all this to be available anytime soon. But many of these items are available now and would just need to be incorporated (if possible) or equivalents developed under the ACF banner.

I've strayed from the point somewhat but the key thing is that complexity itself is not a bad thing and the risks of too much complexity can be mitigated. However transparency is a separate issue and I believe would strengthen the rating system rather than make it divisive.

Rincewind
03-09-2005, 08:38 AM
I have voted no, the reason being is that I believe most chess players will just be more confused by more numbers. This will also give players something else to complain about with regards to ratings.

I think very few if any would actually bother to sit down and work out some ratings accurately if given the 'tools' to do so.

Publishing more numbers would just lead to more complaints and more supposedly unhappy people, regardless of the facts.

I think you underestimate the intelligence of "most chess players". There would seem to be a correlation between chess and mathematical aptitude. Not surprising since both rely on pattern recognition and spacial visualision.

Secondly, the non-publication of RDs and Volatilities already gives players something to complain about. What is more, opacity provides the opportunity for players to suggest that fudgnig is being performed. A claim which would be patently defendable were the system entirely transparent.

The number of people who would bother to perform the calculation themselves is a non-issue. I don't suggest that most would. However, as I have already stated, the benefits of transparency remain.

Finally, more information doesn't necessarily lead to more complaints. I covered this in para 2. Increased transparency would, I think, lead to a greater confidence in the fidelity of the system and while perhaps not less complaints, a greater general acceptance of the egality of the system as a whole.

PHAT
03-09-2005, 11:24 AM
Actually my position is not that the system is too complicated, just that all the factors should be in the public domain. The complexity of the system is a necessary artifact of its design to model a complicated physical system.

To overcome the complexity issues, open source tools can be provided to ease the burden on the individual having to develop or even understand their function.

My vision for the ACF ratings page would be to have the current ratings available as well as an estimated new rating which is updated with every swiss perfect file which comes through to the ACF during the period. A report breaking down each rating change game by game and reports available on a tournament by tournament basis as well. In addition to that tools like my calculator which can provide hypothetical rating estimates or estimates for those for tournaments still not yet received. And finally a graphing feature like that developed by (I think) Lee Jones. In short, a one stop shop with everything a player, organisor, journalist, archivist or stats major might need to know about ratings.

Remember this is a vision statement and so I don't expect all this to be available anytime soon. But many of these items are available now and would just need to be incorporated (if possible) or equivalents developed under the ACF banner.

I've strayed from the point somewhat but the key thing is that complexity itself is not a bad thing and the risks of too much complexity can be mitigated. However transparency is a separate issue and I believe would strengthen the rating system rather than make it divisive.

You must be wrong! Because I agree 100% with everything you have said above.

More later.

Garvinator
03-09-2005, 02:15 PM
I think you underestimate the intelligence of "most chess players". There would seem to be a correlation between chess and mathematical aptitude. Not surprising since both rely on pattern recognition and spacial visualision.
i dont think i underestimate them at all.

Rincewind
03-09-2005, 04:09 PM
i dont think i underestimate them at all.

Perhaps your opinion is skewed as you are more familiar with players from your own geographic location. Many chess players from the Sydney basin and surrounds have solid backgrounds in mathematics. Certainly the average would not be confused by a number to replace the !!, !, etc system and another to measure volatility. Neither are difficult concepts for the average person down here to grasp.

PHAT
03-09-2005, 05:15 PM
There would seem to be a correlation between chess and mathematical aptitude. Not surprising since both rely on pattern recognition and spacial visualision.

Personality traits matter just as much. A person who craves to win will make fewer mistakes because of fear and determination. Me, I have no fear and tend not to take games too seriously. Hense, it is likely that I will never go much past 1500 (current 1300). Also, I have met some rather mediocre minds who's strong drive helps them to belt me when I am feeling more casual than usual.

Rincewind
05-09-2005, 07:29 PM
From the shoutbox...


I voted no even though I thought should be yes, as long as not more work or questions for the ratings officer

Why would it be any more work or questions?

antichrist
05-09-2005, 07:45 PM
From the shoutbox...



Why would it be any more work or questions?

Post 3
Originally Posted by macavity
The questions that arise for me;

Is it a truly easy thing to utilize the factors above, for say one of us?

BG:No.

Bill's response is what I based my vote on.

You wil get questions regardless of necessity of such. People, e.g., some chess mothers who know virtually nothing about chess, putting their bib in with their fragile egos. Even some dads may be in the same boat. Not me but.

arosar
05-09-2005, 08:01 PM
I'm having me a nice shiraz at the moment. On this thread, I reckon I support RW just to piss Bill off. Anyway, it does make sense to have all that data available for the sake of scientificity.

AR

Rincewind
05-09-2005, 08:35 PM
You wil get questions regardless of necessity of such. People, e.g., some chess mothers who know virtually nothing about chess, putting their bib in with their fragile egos. Even some dads may be in the same boat. Not me but.

The point is questions are being asked now and they cannot be answered because the information is not freely available.

This thread is not about any specific question or any specific player. It is a discussion on current the policy of not publishing information which would make the processing of ratings transparant. Your unsubstatiated, attempted stirs just shows you to be the callow oaf that regular reader here know you to be.

antichrist
05-09-2005, 08:38 PM
The point is questions are being asked now and they cannot be answered because the information is not freely available.

This thread is not about any specific question or any specific player. It is a discussion on current the policy of not publishing information which would make the processing of ratings transparant. Your unsubstatiated, attempted stirs just shows you to be the callow oaf that regular reader here know you to be.

Any more compliments you can dish my way? As I said earlier, a bit of skirt and you become Shining White Knight!

Bill, come to my rescue, I was just trying to protect you from idiots.

Rincewind
05-09-2005, 08:41 PM
Bill, come to my rescue, I was just trying to protect you from idiots.

I could do that by banning your account. :D

antichrist
05-09-2005, 08:42 PM
I could do that by banning your account. :D

pay that one

Kevin Bonham
05-09-2005, 09:01 PM
I'd just like to know if there are any specific "scurrilous accusations" going about - eg does anyone really think there has been any "subjectivity" or is this all just completely hypothetical. Mischa: what was your motivation for asking; is there any particular rating you are concerned/curious about?

As far as I'm aware the only claim of tweaking that has been made was the claim that Nick Speck's rating was tweaked after he tied for first in the 2001-2 Aus Champs tournament (he later lost to Johansen in the playoff of course). That claim was based on a misunderstanding and was comprehensively debunked, despite which I still see it show up now and then.

Bill Gletsos
05-09-2005, 09:03 PM
Any more compliments you can dish my way? As I said earlier, a bit of skirt and you become Shining White Knight!

Bill, come to my rescue, I was just trying to protect you from idiots.Always a hard job A/C.

When the system was under Elo I used to get questions from people who couldnt even do the simple Elo maths which was for many years explained in the back of the ACF rating booklets. They also seemed to make simple mistakes like forgetting about whole tournaments that they played in (funny how it was always the tournaments they did badly in). The small number of players who did actually ever even bother to attempt to calculate their ratings changes under the old ACF Elo Sytem still could not do it accurately because they had no idea how an unrated player nor a provisionally rated player (one with less than 30 games) was likely to effect their ratings. As such even back then they could only ever do approximations of their rating changes.

As a final comment, the number of questions/queries I get from players with regards their ratings now is no greater than when the system was under Elo which was never more than 5-10 per year.

Rincewind
05-09-2005, 09:08 PM
I'd just like to know if there are any specific "scurrilous accusations" going about - eg does anyone really think there has been any "subjectivity" or is this all just completely hypothetical. Mischa: what was your motivation for asking; is there any particular rating you are concerned/curious about?

As far as I'm aware the only claim of tweaking that has been made was the claim that Nick Speck's rating was tweaked after he tied for first in the 2001-2 Aus Champs tournament (he later lost to Johansen in the playoff of course). That claim was based on a misunderstanding and was comprehensively debunked, despite which I still see it show up now and then.

I think Mischa's subsequent shouts (above) indicated she was not intending to make any "scurrilous accusation", specific or otherwise. However, my point is that by keeping the RDs and volatilities secret it does open the ACF up to accusations of rating tampering. The argument is, if there is nothing to hide, why not publish them? Increasing the transparency should increase the confidence that no one is tampering with the system. On the down side I can't see how publishing the data will lead to any extra work.

Rincewind
05-09-2005, 09:15 PM
When the system was under Elo I used to get questions from people who couldnt even do the simple Elo maths which was for many years explained in the back of the ACF rating booklets. They also seemed to make simple mistakes like forgetting about whole tournaments that they played in (funny how it was always the tournaments they did badly in). The small number of players who did actually ever even bother to attempt to calculate their ratings changes under the old ACF Elo Sytem still could not do it accurately because they had no idea how an unrated player nor a provisionally rated player (one with less than 30 games) was likely to effect their ratings. As such even back then they could only ever do approximations of their rating changes.

As a final comment, the number of questions/queries I get from players with regards their ratings now is no greater than when the system was under Elo which was never more than 5-10 per year.

Are you saying that even though the complexity has increased and now beyond the ken of the average player but miraculously the number of queries have not increased. I submit that this trend should follow (ie no increase in queries) if the RDs and Volatilities were released. If this is true, the question that remains is...

Is there any good reason not to publish the RDs and Volatilies?

antichrist
05-09-2005, 09:18 PM
Always a hard job A/C.

When the system was under Elo I used to get questions from people who couldnt even do the simple Elo maths which was for many years explained in the back of the ACF rating booklets. They also seemed to make simple mistakes like forgetting about whole tournaments that they played in (funny how it was always the tournaments they did badly in). The small number of players who did actually ever even bother to attempt to calculate their ratings changes under the old ACF Elo Sytem still could not do it accurately because they had no idea how an unrated player nor a provisionally rated player (one with less than 30 games) was likely to effect their ratings. As such even back then they could only ever do approximations of their rating changes.

As a final comment, the number of questions/queries I get from players with regards their ratings now is no greater than when the system was under Elo which was never more than 5-10 per year.

But now with a hyperactive BB, zealots could drive ratings officers crazy. And you have a few posters who will argue the toss just for the sake of arguing with you.

Bill's objection in post 3 still holds despite my callowness.

Rincewind
05-09-2005, 09:22 PM
But now with a hyperactive BB, zealots could drive ratings officers crazy. And you have a few posters who will argue the toss just for the sake of arguing with you.

BB has been running in one form or another for several years. I assume Bill includes these figures in his 5-10/year estimate. (PS you are the biggest offender in the argumentative stakes).

antichrist
05-09-2005, 09:31 PM
BB has been running in one form or another for several years. I assume Bill includes these figures in his 5-10/year estimate. (PS you are the biggest offender in the argumentative stakes).

Nope, can't agree there. Bill describes me as only a nuisance, FG and Matt are the big offenders - so there.

How do you rate in the insulting stakes? A bit of skirt...

I might do a Brodgen if you keep this up.

Rincewind
05-09-2005, 09:35 PM
Nope, can't agree there. Bill describes me as only a nuisance, FG and Matt are the big offenders - so there.

I think you have misapplied the word "offender". While you think about that perhaps you would like to run a poll on "who is the most argumentative poster here?" and stay out of chess related threads until you can add something relevant.

antichrist
05-09-2005, 09:37 PM
I think you have misapplied the word "offender". While you think about that perhaps you would like to run a poll on "who is the most argumentative poster here?" and stay out of chess related threads until you can add something relevant.

In post 20 I brought an approp point which you appeared to have missed - then you started the slander.

The whole weekend I did not get a decent argument, you have made me happy now. Thanks

Rincewind
05-09-2005, 09:44 PM
In post 20 I brought an approp point which you appeared to have missed - then you started the slander.

On the contrary, the point you parroted in post #20 was covered by the final paragraph of my post #10. I can't be held to account if you don't read and/or understand all the posts. How is your ADD going, by the way?

antichrist
05-09-2005, 09:48 PM
On the contrary, the point you parroted in post #20 was covered by the final paragraph of my post #10. I can't be held to account if you don't read and/or understand all the posts. How is your ADD going, by the way?

I managed to watch all of 4 'Corners and now watching Enough Rope, even concentrating on 2 things at once - surprise, surprise (as Jim Nabors says it in Sgt Bilko)

Kevin Bonham
05-09-2005, 10:28 PM
I think Mischa's subsequent shouts (above) indicated she was not intending to make any "scurrilous accusation", specific or otherwise.

I think a full history of the shoutbox including all deleted shouts would cast far more light upon the matter. Even from those still there I think I can read between the lines and the writing between the lines spells "**l* **r**t ***dr***" Who wants to buy a vowel? :P :P


However, my point is that by keeping the RDs and volatilities secret it does open the ACF up to accusations of rating tampering. The argument is, if there is nothing to hide, why not publish them? Increasing the transparency should increase the confidence that no one is tampering with the system. On the down side I can't see how publishing the data will lead to any extra work.

I don't see why the ACF should give a damn about "accusations of rating tampering" unless there is actually reason to believe there is substance to them - and any tampering large-scale enough to justify the effort would be detectable using an approximate calculator.

Because of the use of intermediate ratings and the way unrated players are rated, to check a published rating with complete accuracy you need:

* The ratings, volatilities and RDs for every player active in the period, including the unpublished internal ratings for previously active unrated players.

* The correct histories of every rated game played in the country in the ratings period, as finalised by the ACF ratings officers in the final days before ratings are issued.

* The full set of ACF formulae used in the process.

So any transparency-checking that can be done will be after the fact, and will be only able to be done by someone with the ability and desire to program calculators for all the above completely accurately - this for a trivial gain in accuracy in a circumstance where no evidence of corruption has ever been advanced.

IMO this is a beat-up.

arosar
05-09-2005, 10:38 PM
IMO this is a beat-up.

We cannot just have you blurt out your mantra Kevo. It's a legit question. The bureaucracy does not need a reason to believe accusations of tampering. First and foremost, it need only to be transparent to its constituents.

Further, that the process of checking might be an onerous one is no excuse either.

AR

Rincewind
05-09-2005, 10:40 PM
I'll buy a vowel.

I believe the ACF definitely should give a damn about "accusations of rating tampering", even if such accusations are only made between players not forward publicly or through proper channels. Such accusations only weaken the confidence chess players have in the objectivity of the rating system.

To maintain the position we will continue to withhold information and those who think we are cheating can go and get stuffed is nothing short of arrogance and only helps fuel any speculation.

PHAT
05-09-2005, 11:12 PM
BazzCox,

You have my sympathy and empathy. When I wanted info from the ACF rating orfficer, I got nothing. I know that you want the info for good intentions., too. I suspect that you NOW know that I know that you know that I know that you know that I know there IS a chess Mafia and we are not members.

Kevin Bonham
05-09-2005, 11:33 PM
We cannot just have you blurt out your mantra Kevo.

You're missing the story on this one, AR. You seem to be confused about who is doing all the hiding here.


It's a legit question. The bureaucracy does not need a reason to believe accusations of tampering. First and foremost, it need only to be transparent to its constituents.

How "transparent" can it be when 99+% of those constituents would be incapable of accurately assessing a claim about an error in the ratings and determining whether that claim was accurate?

I've already explained how a fairly accurate degree of transparency is already possible - enough to detect systematic corruption on a serious level. What part of my argument on that point did you wilfully fail to comprehend?


Further, that the process of checking might be an onerous one is no excuse either.

Yes it is - if it suggests that virtually nobody will bother to do it properly.


I'll buy a vowel.

I believe the ACF definitely should give a damn about "accusations of rating tampering", even if such accusations are only made between players not forward publicly or through proper channels. Such accusations only weaken the confidence chess players have in the objectivity of the rating system.

But that suggests that accusations are being made between players. Are they? Are there really people out there who believe that their ratings or the ratings of their friends, relatives, or other connections are being deliberately diddled? Has anyone here heard anything of that sort?


To maintain the position we will continue to withhold information and those who think we are cheating can go and get stuffed is nothing short of arrogance and only helps fuel any speculation.

But is there any speculation? Or is this just speculation about speculation? Or is it a speculation that dare not speak its name after its initial outburst for fear of being made to look as totally silly as it is?

I'm a semi-generous man, there's an o after the second r. If you want any more vowels you'll have to do much better than that. :whistle:

arosar
06-09-2005, 12:01 AM
You're missing the story on this one, AR. You seem to be confused about who is doing all the hiding here.

Err...no.

AR

Rincewind
06-09-2005, 12:17 AM
But that suggests that accusations are being made between players. Are they? Are there really people out there who believe that their ratings or the ratings of their friends, relatives, or other connections are being deliberately diddled? Has anyone here heard anything of that sort?

I would say everyone has heard someone suggesting that at one time or another.


But is there any speculation? Or is this just speculation about speculation? Or is it a speculation that dare not speak its name after its initial outburst for fear of being made to look as totally silly as it is?

I believe there is speculation. Further the position of distributing information on a 'needs to know' basis serves what purpose?

Kevin Bonham
06-09-2005, 03:32 AM
I would say everyone has heard someone suggesting that at one time or another.

Well, I haven't heard anything of the sort for at least 15 years, except for the Speck case (which was an accusation about tweaking formulae not results or calculations, and was comprehensively discredited anyway.)

Until this ... whatever this may be ...


I believe there is speculation.

I believe you know there is, or was, speculation but don't want to embarrass the speculator. (Come on AR, this really is right up your alley, pay attention. :P )


Further the position of distributing information on a 'needs to know' basis serves what purpose?

I think there are good reasons for keeping the internal ratings of unrated players secret - you've probably seen how much players whine if their early and fairly inaccurate ratings go down a stack as it is. And there's no point from a transparency perspective releasing the other stuff without that information.

(I also make the ambit claim that the more data you release the more ammo you give people engaged in mistaken beatups to create confusion, data-drift and the false impression that they have a clue. If there's no evidence of corruption in a situation where it would be easily detectable there's no point making it easier for the DRs of this world to waste even more of people's time and energy responding to their rubbish.)

antichrist
06-09-2005, 07:22 AM
...(I also make the ambit claim that the more data you release the more ammo you give people engaged in mistaken beatups to create confusion, data-drift and the false impression that they have a clue. If there's no evidence of corruption in a situation where it would be easily detectable there's no point making it easier for the DRs of this world to waste even more of people's time and energy responding to their rubbish.)

Combine that with overzealous parents - some with fragile egos, teenagers finding their voice and shooting off half -cocked and adults who whinge about everything as well as maybe having a grudge against authority in general and the ratings officer.

As I earier stated: a little bit of info is dangerous, and in this case a lot of info is more dangerous. Like with atomic bomb secrets.

We have already seen here how flame wars have taken whole weekends over nothing of significance at all.

(if there is anyone I have missed out in insulting I sincerely apologise)

Rincewind
06-09-2005, 08:03 AM
As I earier stated: a little bit of info is dangerous, and in this case a lot of info is more dangerous. Like with atomic bomb secrets.

As far as I can tell your position onthis issue keeps shifting like a flag in the wind. Now you are saying that informatino can be dangerous. Please explain how knowledge of the RDs and Volatilities can dangerous.

Rincewind
06-09-2005, 08:16 AM
Well, I haven't heard anything of the sort for at least 15 years, except for the Speck case (which was an accusation about tweaking formulae not results or calculations, and was comprehensively discredited anyway.)

Until this ... whatever this may be ...

Except the Speck case. I have also heard other less well publicised cases of varying degrees. Even things as innocent as Ronald's comment on the board a few days ago. Any time the rating system does not behave as expected people will question its validity and much of this questioning could be prevented by releasing all the information and then people (if they so desired) could see the results for themselves. I repeat the policy of secrecy only fuels speculation.


I believe you know there is, or was, speculation but don't want to embarrass the speculator. (Come on AR, this really is right up your alley, pay attention. :P )

There were legitimate questions which have been answered by Bill releasing the information on a small scale in a case by case way. However, much of this angst could be avoid by making the information freely available.


I think there are good reasons for keeping the internal ratings of unrated players secret - you've probably seen how much players whine if their early and fairly inaccurate ratings go down a stack as it is. And there's no point from a transparency perspective releasing the other stuff without that information.

That is a false point. Even if the internal ratings of unrated players is not released (and I never said it should be) then more accurate RDs and volatilities will still lead to the individuals being able to calculate more accurate rating changes. More ambitious individuals might even source the results of their unrated oppositions and estimate their own internal provisional ratings as well.


(I also make the ambit claim that the more data you release the more ammo you give people engaged in mistaken beatups to create confusion, data-drift and the false impression that they have a clue. If there's no evidence of corruption in a situation where it would be easily detectable there's no point making it easier for the DRs of this world to waste even more of people's time and energy responding to their rubbish.)

The only beat-up here is you Kevin. You are deperately trying to make this thread about a individual query which has been answered. It is not and was never intended to be. It is a discussion on the ACF policy of not releasing all the rating information regarding individuals (RDs and Volatilites) to those individuals. Even though there is no good reason not to. Even though the majority of those people want this information to be released. And even though, ostensibly, this is funded by the individual via the collection of the per game rating fee.

My position has always been that the information should be made available if there is not a good reason to not do so. Extraplating on this poll I could claim that most of the ACFs paying customers agree with me.

Kevin Bonham
06-09-2005, 04:24 PM
Even things as innocent as Ronald's comment on the board a few days ago.

... which was based on a misreading of a shout above it and included no accusations or even hints of improprietry whatsoever and is therefore completely irrelevant.


Any time the rating system does not behave as expected people will question its validity and much of this questioning could be prevented by releasing all the information and then people (if they so desired) could see the results for themselves. I repeat the policy of secrecy only fuels speculation.

Yes, they will often question the system's adequacy as a model, or even the accuracy of the calculations, but these are quite different things from "scurrilous accusations" about so-called "subjective influences" (ie corruption). I just wanted to know if any accusations of corruption had been made or if any was suspected, because if virtually no-one really suspects fiddling then why is it even being raised as an issue?

Why didn't those of you getting involved in this campaign just abandon the line about "scurrilous accusations" and instead argue that transparency could protect the system from innocent errors?


There were legitimate questions which have been answered by Bill releasing the information on a small scale in a case by case way. However, much of this angst could be avoid by making the information freely available.

Angst? I don't see any angst on my side of this, so I assume it must be somewhere else. :lol: Sounds more like initially legitimate questions became less so at some point.


That is a false point. Even if the internal ratings of unrated players is not released (and I never said it should be) then more accurate RDs and volatilities will still lead to the individuals being able to calculate more accurate rating changes.

True, but how much more accurate, considering the issues with unrated players and the use of RPs? Also, are you talking about calculation of ratings changes in advance (in which case the accuracy increase would be limited) or post-hoc, and if the latter, to what end?


More ambitious individuals might even source the results of their unrated oppositions and estimate their own internal provisional ratings as well.

Even more ambitious individuals might research the entire history of rated chess in Australia from scratch and do the whole thing themselves, assuming they could obtain the 11 secret herbs and spices from the ratings officers. :P


The only beat-up here is you Kevin. You are deperately trying to make this thread about a individual query which has been answered. It is not and was never intended to be. It is a discussion on the ACF policy of not releasing all the rating information regarding individuals (RDs and Volatilites) to those individuals. Even though there is no good reason not to. Even though the majority of those people want this information to be released. And even though, ostensibly, this is funded by the individual via the collection of the per game rating fee.

My position has always been that the information should be made available if there is not a good reason to not do so. Extraplating on this poll I could claim that most of the ACFs paying customers agree with me.

If I was so obsessed with the "individual query" that I am supposedly beating up then I wouldn't be bothering to attempt to engage with points directly relevant to your topic. (Although, as it happens I am interested in the history of this query, because I sense a degree of hedging about just what may have been said in the heat of the moment and what it meant.)

You are saying there is no good reason not to release the data - I have just given one that I consider valid. Another is that I simply think a more convincing argument is needed to justify any increase in effort, however small, for ratings officers. The BB is not necessarily a representative sample of the Australian chess community and the sample size of voters is too small anyway. As for the per-game fee, I suppose you are going to claim the fact that virtually every chessplayer in the country continues paying it in spite of the lack of the data provided as evidence of some kind of market failure? That should be interesting for us all.

Hmmmm. For my next exciting adventure I might start a thread called "Vicarious rating angst" and insist that it has nothing to do with anything that was discussed on this one whatsoever. :whistle:

Rincewind
06-09-2005, 07:01 PM
... which was based on a misreading of a shout above it and included no accusations or even hints of improprietry whatsoever and is therefore completely irrelevant.

Not so as it reflected a perception that the systems operation in some way favours a certain rating group. My hope is that if more of the information is available, more accurate predictions would be possible and there will be a greater awareness of the operation of the rating system ni the wider chess-playing community. Were that to come to pass, such misconceptions would naturally be reduced.

My point is that it is currently easy for people to allocate blame to the rating system as a whole, either innocently or by inferring tampering as it is for the wider public a black box which the ACF ratings officers turn the handle and inside some miracle occurs.


Yes, they will often question the system's adequacy as a model, or even the accuracy of the calculations, but these are quite different things from "scurrilous accusations" about so-called "subjective influences" (ie corruption). I just wanted to know if any accusations of corruption had been made or if any was suspected, because if virtually no-one really suspects fiddling then why is it even being raised as an issue?

Why didn't those of you getting involved in this campaign just abandon the line about "scurrilous accusations" and instead argue that transparency could protect the system from innocent errors?

Subjective influences is not the same thing as corruption. For example, Bill and Graham have made a number of changes to the rating system over time. For example the way FIDE rated players are seeded with provisional ACF ratnigs or the RD factor applied to Rapid rated players who begin playing in standard rated events. I've seen Bill release information as to the rationale and some statement as to the degree of testing of these changes prior to implementation, however, I've not seen a scientifically rigourous statement of method and results. As such I would dispassionately argue that these changes could well have been coloured with subjective influences without inferring of any impropriety at all.

To my mind, anything increase in transparency is a good thing. It allows others to reproduce calculations and even assess the affect and effectiveness of rating system changes themselves.


Angst? I don't see any angst on my side of this, so I assume it must be somewhere else. :lol: Sounds more like initially legitimate questions became less so at some point.

Perhaps I mingle in more colourful circles than you. You really must get out more.


True, but how much more accurate, considering the issues with unrated players and the use of RPs? Also, are you talking about calculation of ratings changes in advance (in which case the accuracy increase would be limited) or post-hoc, and if the latter, to what end?

In the particular case in point I believe Bill sets the figure at approximately 30% of the error was due to inaccurate RD information and 70% of the error was due to intermediate rating/FIDE rated player issues. Of course I must take Bill at his word on this as I cannnot make my own assessment. This too is not an accusation of impropriety, I just like to check calculations myself as mistakes happen, even (occasionally) by me.


Even more ambitious individuals might research the entire history of rated chess in Australia from scratch and do the whole thing themselves, assuming they could obtain the 11 secret herbs and spices from the ratings officers. :P

Yes that is true but most of that knowledge has been filtered into a much more practical form by the ACF ratings officers over the history of that position's existence. The information is accessible, collated and could be published for practically zero cost to the ACF.


If I was so obsessed with the "individual query" that I am supposedly beating up then I wouldn't be bothering to attempt to engage with points directly relevant to your topic. (Although, as it happens I am interested in the history of this query, because I sense a degree of hedging about just what may have been said in the heat of the moment and what it meant.)

I'm encouraged by you engaging in the broader scope of the thread. Your early comments did seem to be very much centred around a specific query which (as I pointed out) was answered prior to the shouts I copied above. The resolution of that query may not have been entirely satisfactory to all parties but it is really not what this poll or thread it about. As I said in the first post, the comments in the shoutbox highlighted the issue of the transparency of the system.


You are saying there is no good reason not to release the data - I have just given one that I consider valid. Another is that I simply think a more convincing argument is needed to justify any increase in effort, however small, for ratings officers. The BB is not necessarily a representative sample of the Australian chess community and the sample size of voters is too small anyway. As for the per-game fee, I suppose you are going to claim the fact that virtually every chessplayer in the country continues paying it in spite of the lack of the data provided as evidence of some kind of market failure? That should be interesting for us all.

My belief is that the BB is probably skewed the other way and the wider community would probably more aligned to Don't Care and Yes than No. This position is based on my perception that ACF and State Assoc committee members are overrepresented here and they could well have a more jaundiced view than Mr/Ms Average who simply want access to all his/her information.

Regarding the policy of increased effort. No one wants to make the Ratings Officers' job any more difficult and I don't see that the increased effort is really anything more than changing the report that produces the master file and active lists once and all existing practices continue as per normal. Certainly it is significantly less effort than the increase from 3 to 4 rating periods, or the introduction of Glicko-1 or the move to Glicko-2 or even the other "small" modifications mentioned in post #2.

Oepty
06-09-2005, 07:19 PM
I see two significant instances where an error in ratings might have some real effect on a player.
We have a rating cut off for entry to the Australian Championships and a mistake of even a few points could cause a player to miss out on playing on the Championships. I know there are other avenues for a player to get into the championships.
Ratings and in particular the order of players rating might make a significant difference to the selection of places in a team. It would be a pity if the 6th place in an Olympiad team was decided based on an incorrect publised rating. These are perhaps not some much arguments for the RDs etc to be released but for any corrections to the ratings to be published as soon as possible after the error has been found.
Scott

PS I voted yes, but only just. There is quite a reasonable arguement for saying no as well.

Kevin Bonham
06-09-2005, 11:23 PM
Not so as it reflected a perception that the systems operation in some way favours a certain rating group.

I didn't personally perceive any suggestion of a deliberate fiddle in favour of high rated players in what he was saying though - though granted, having lengthy discussions on the precise "meaning" of shoutbox posts is always tricky territory. Actually it's an irony because the usual flow of criticism has been that Glicko is more harmful to high ratings.


My hope is that if more of the information is available, more accurate predictions would be possible and there will be a greater awareness of the operation of the rating system ni the wider chess-playing community. Were that to come to pass, such misconceptions would naturally be reduced.

No - in this case Duff simply misunderstood a comment by Mischa as referring to a ratings gain per period when it actually referred to one tourney.


My point is that it is currently easy for people to allocate blame to the rating system as a whole, either innocently or by inferring tampering as it is for the wider public a black box which the ACF ratings officers turn the handle and inside some miracle occurs.

True but I suspect publication of greater information would simply transfer this perception of the "black box" to other aspects of the system, eg the formulae used to calculate the ratings.


Subjective influences is not the same thing as corruption. For example, Bill and Graham have made a number of changes to the rating system over time. For example the way FIDE rated players are seeded with provisional ACF ratnigs or the RD factor applied to Rapid rated players who begin playing in standard rated events. I've seen Bill release information as to the rationale and some statement as to the degree of testing of these changes prior to implementation, however, I've not seen a scientifically rigourous statement of method and results. As such I would dispassionately argue that these changes could well have been coloured with subjective influences without inferring of any impropriety at all.

Yes, but here we do come back to the whole issue of where this thread came from (which I will admit I'm starting to feel like I've said enough on and am getting a little bored with). You might argue this dispassionately in isolation but what we were witnessing was a call about "subjective influences" in response to a rating not having gone up as much as someone expected it should.


Perhaps I mingle in more colourful circles than you. You really must get out more.

???????????? Am I missing something here? Damn. Must be the wine. :lol:


My belief is that the BB is probably skewed the other way and the wider community would probably more aligned to Don't Care and Yes than No. This position is based on my perception that ACF and State Assoc committee members are overrepresented here and they could well have a more jaundiced view than Mr/Ms Average who simply want access to all his/her information.

ACF and SA commitariat may be overrepresented here but opinionated critics of the ACF also are. Mr/Ms Average found even the previous rating system difficult to follow and would find this one "so far beyond my comprehension that I cannot comment further" (to quote one experienced administrator's view of the math of the Glicko system) no matter how much detail you did or did not give them. Of course, free software tools would then exist but I think there is some danger in giving people much more info for only a small increase in accuracy. Expectations may become unrealistic, leading to an increased number of complaints for the RO to deal with.

Why any normal (sub-2100) player should care overmuch if their rating or the rating of anyone they care about is <20 points higher or lower than what they expected is beyond me.


Certainly it is significantly less effort than the increase from 3 to 4 rating periods, or the introduction of Glicko-1 or the move to Glicko-2 or even the other "small" modifications mentioned in post #2.

True but these resulted in significant benefits to the accuracy of the ratings - unless there are lots of mistakes being made (which I certainly don't believe) greater transparency will not.

Kevin Bonham
06-09-2005, 11:30 PM
I see two significant instances where an error in ratings might have some real effect on a player.
We have a rating cut off for entry to the Australian Championships and a mistake of even a few points could cause a player to miss out on playing on the Championships. I know there are other avenues for a player to get into the championships.

Good on you Scott - you're getting at a point that Barry for some reason didn't pick up on when I kept repeating the line about "why should anyone care about ratings fiddles when any large scale fiddle would be detectable and any small scale one not worth the bother"?

There are contexts where small rating differences can be crucial. The Aus Champs is not really one of them because a player just below the cutoff who applies and has ever shown form at that level is almost always let in (and a player who only just gets in is hardly going to win the thing anyway).

Also typically selectors ignore very small rating differences unless there is nothing else to split the candidates at all, in which case the selection is pretty much a lottery anyway.

However one context in which a small rating difference can be decisive is an emergency selection where time is limited based on the average of the applicant's ACF and FIDE ratings.

PHAT
06-09-2005, 11:41 PM
This whole convo is going in circles. Kevin, stop arguing for the sake of arguing, and say what you realy think. Coxrinse, stop wasteing time with Kevin and go after the Gletsos. Gletsos has the info, go to the source.

Kevin Bonham
07-09-2005, 12:06 AM
This whole convo is going in circles. Kevin, stop arguing for the sake of arguing, and say what you realy think.

I have been doing the latter actually, it's just that sometimes I've seen potential counter-points for the other side after saying things and then been amused that Barry didn't grab them.

Since you insist I wouldn't mind slightly more breakdown of RDs if it made predictions players could do of their ratings much more accurate - but from the above it sounds like the difference really isn't that great because the RP thing (a truly excellent innovation) is more significant and requires full knowledge of player results for the period.


stop wasteing time with Kevin and go after the Gletsos. Gletsos has the info, go to the source.

Note Bill's vivid interest in this thread. :lol:

Rincewind
07-09-2005, 12:23 AM
I didn't personally perceive any suggestion of a deliberate fiddle in favour of high rated players in what he was saying though - though granted, having lengthy discussions on the precise "meaning" of shoutbox posts is always tricky territory. Actually it's an irony because the usual flow of criticism has been that Glicko is more harmful to high ratings.

No - in this case Duff simply misunderstood a comment by Mischa as referring to a ratings gain per period when it actually referred to one tourney.

The soruce of the confusion mostly irrelevant. What is relevant was that Duff was quick to allocate "blame" at the doorstep of the rating system. Having a blackbox system is more open to this sort of thing, and based no my limited sampling, occurs every 3 months in chess clubs all around the country.


True but I suspect publication of greater information would simply transfer this perception of the "black box" to other aspects of the system, eg the formulae used to calculate the ratings.

If the formulae are available they are not black boxes, just knowledge beyond current comprehension. If anyone has a query on their rating adjustment and was the system repeatable, then there are avenues people can take to overcome this hurdle without having to make an official approach to the ACF which they may be reticent to do for a number of reasons.


Yes, but here we do come back to the whole issue of where this thread came from (which I will admit I'm starting to feel like I've said enough on and am getting a little bored with). You might argue this dispassionately in isolation but what we were witnessing was a call about "subjective influences" in response to a rating not having gone up as much as someone expected it should.

Just pointing out that the phrase "subjective influences" may be interpreted in ways which doesn't carry any air of misconduct. People may say and interpret things said in one way. But all in all it is a non-issue. In fact I would argue tat subjective influences are an unavoidable consequence of having an opaque system. If the goal is to avoid subjectivity then transparency at some level is a must.


ACF and SA commitariat may be overrepresented here but opinionated critics of the ACF also are. Mr/Ms Average found even the previous rating system difficult to follow and would find this one "so far beyond my comprehension that I cannot comment further" (to quote one experienced administrator's view of the math of the Glicko system) no matter how much detail you did or did not give them. Of course, free software tools would then exist but I think there is some danger in giving people much more info for only a small increase in accuracy. Expectations may become unrealistic, leading to an increased number of complaints for the RO to deal with.

Why any normal (sub-2100) player should care overmuch if their rating or the rating of anyone they care about is <20 points higher or lower than what they expected is beyond me.

Is a 6 year old rated at 1500 "normal"? For juniors and female players selection issues could typically involve ratings below 2100. However I believe all players have an equal right to a transparent system, regardless of whether you think they should care or not.


True but these resulted in significant benefits to the accuracy of the ratings - unless there are lots of mistakes being made (which I certainly don't believe) greater transparency will not.

There is more to the system than just producing more accurate ratings. Perception of the fidelity is another issue. For example, references to information already in the public domain to substantiate rating change queries provides a much stronger response that releasing data which was previously held in confidence. Also the data being available will mean that people will be able to perform more accurate calculations which will in turn lead to a greater level of expecation alignment and therefore less dissatisfied players when the rating lists are published.

As I said the cost of publication is practically zero. The arguments against releasing it are flimsy at best and the majority of paying customers seem to want it. Releasing the data will make the system more robust in the face of scrutiny, lead to fewer queries via better expectation management and be a big step toward removing subjective influences from the rating process.

Kevin Bonham
07-09-2005, 12:46 AM
The soruce of the confusion mostly irrelevant. What is relevant was that Duff was quick to allocate "blame" at the doorstep of the rating system.

The system, you say? Sure. But check the context in which you originally raised this example and what I was asking in reply to which you raised it.


If the formulae are available they are not black boxes, just knowledge beyond current comprehension.

Same difference because if people cannot comprehend the rationale for the formulae they can easily imagine that the ROs have tweaked the system by rigging the formulae against them.


Just pointing out that the phrase "subjective influences" may be interpreted in ways which doesn't carry any air of misconduct.

Hmmm. Maybe if the user could explain what she meant by the term in her own words, that might help? :lol:


Is a 6 year old rated at 1500 "normal"?

Show me one and I'll tell you. :whistle:


For juniors and female players selection issues could typically involve ratings below 2100.

See reply to Scott.


However I believe all players have an equal right to a transparent system, regardless of whether you think they should care or not.

I'm concerned with who has a demonstrable need for it, rather than the fictitious concept of "right", which, as my favourite philosopher pointed out, is a "wheel in the head, put there by a spook".


There is more to the system than just producing more accurate ratings. Perception of the fidelity is another issue.

It is but I do not believe the nonpublication of precise information impacts upon it - as demonstrated by your failure to produce any example of an accusation of improper tweaking or corruption apart from the discredited Speck beatup since the Glicko started.


Releasing the data will make the system more robust in the face of scrutiny, lead to fewer queries via better expectation management and be a big step toward removing subjective influences from the rating process.

Do you have empirical examples of other federations where this has been done and proved to be the case or is all this just total supposition based perhaps on a belief that others would react to having more info available in the way that you would?

Rincewind
07-09-2005, 08:03 AM
The system, you say? Sure. But check the context in which you originally raised this example and what I was asking in reply to which you raised it.

I think you just misunderstood me originally. My context has always been regarding the wider question. Duff's case was just an example of the equality of the rating system being called into question.


Same difference because if people cannot comprehend the rationale for the formulae they can easily imagine that the ROs have tweaked the system by rigging the formulae against them.

No it is not the same difference for the reasons I stated in my previous reply.


Hmmm. Maybe if the user could explain what she meant by the term in her own words, that might help? :lol:

Like Bill you will not Mischa's disinterest in this thread.


Show me one and I'll tell you. :whistle:

The word is "hypothetical". I'll take your response as a no. And your argument that ratings are only important for 2100+ players as successfully rebutted.


See reply to Scott.

Seen but irrelevant. It was one example of where the rating system is the butt of blame due to its inscrutibility.


I'm concerned with who has a demonstrable need for it, rather than the fictitious concept of "right", which, as my favourite philosopher pointed out, is a "wheel in the head, put there by a spook".

There are reasons of right as well as utility. Your favourtite philosopher may think that rights are bunk but that is hardly the view of the wider community as is evidenced by the existence of legislation and other legal paraphanlia to protect them.


It is but I do not believe the nonpublication of precise information impacts upon it - as demonstrated by your failure to produce any example of an accusation of improper tweaking or corruption apart from the discredited Speck beatup since the Glicko started.

There are examples, just no examples which have not been credited. This is either evidence of very little or no corruption, or else very good tweaking. The fact remains that people are suspicious and the policy of secrecy fuels such speculation.


Do you have empirical examples of other federations where this has been done and proved to be the case or is all this just total supposition based perhaps on a belief that others would react to having more info available in the way that you would?

I have little experience of other federations, even less so with other federations which use the Glicko-2 rating system. FICS use (I believe) Glicko-1 and publish ratings and RDs (there are no volatilities in Glicko-1).

antichrist
07-09-2005, 10:52 AM
As far as I can tell your position on this issue keeps shifting like a flag in the wind. Now you are saying that informatino can be dangerous. Please explain how knowledge of the RDs and Volatilities can dangerous.

As stated earlier, dangereous in the manner that people will never stop arguing - about interpretations, influences and any other little thing they can grab at. And it is left up to the ratings officer to answer such babble. Unless you volunteer to take on such Q & A Dorothy Dixers.

And a perfect example is this thread - come from nowhere (babble) gone nowhere and going nowhere. Only a great waste of time because someone's Littly Johnny didn't become a GM overnight.

No one has come up with any glaring wrong rating in ACF ratings, I mentioned a possible one in JCL which should have been investigated. So in the total of junior and adult chess we have one only that maybe should have been investigated. And when I mentioned this case earlier you weren't the slightest interested in it. You only wanted to harp on about been able to double check on ratings that have no genuine gripe.

We do not live in a perfect world and ratings officers are not paid overtime to handle queries as far as I know. We get what and even better than what we pay for.

Another example of this is the SEC, one guy only pays ten dollar entry (pensioner?) gets equal with 2 others in the minorest of prizes and whinges that his share is only 20 or thirty bucks. He he had a hot meal thrown in as well. He also got better than what he paid for.

I won't try and make the same comparison with you as Admin of this BB.

Now didn't you miss me yesterday?

Rincewind
07-09-2005, 11:48 AM
As stated earlier, dangereous in the manner that people will never stop arguing - about interpretations, influences and any other little thing they can grab at. And it is left up to the ratings officer to answer such babble. Unless you volunteer to take on such Q & A Dorothy Dixers.

Bill previously advised 5-10/year and no increase since moving from Elo to Glicko to Glicko-2.


And a perfect example is this thread - come from nowhere (babble) gone nowhere and going nowhere. Only a great waste of time because someone's Littly Johnny didn't become a GM overnight.

I'll thank you to keep your personal issues in the non-chess forums. (And try to not make them offensive next time you do that).


No one has come up with any glaring wrong rating in ACF ratings, I mentioned a possible one in JCL which should have been investigated. So in the total of junior and adult chess we have one only that maybe should have been investigated. And when I mentioned this case earlier you weren't the slightest interested in it. You only wanted to harp on about been able to double check on ratings that have no genuine gripe.

Several issues have been addressed over time but this is not about wrong ratings but more repeatablity and removal of subjective influences. Increase of transparency. Giving the customers the information to more reliably calculations and access to the data they are paying for.


We do not live in a perfect world and ratings officers are not paid overtime to handle queries as far as I know. We get what and even better than what we pay for.

Nearly all ACF positions are voluntary how is this significant unless you can successfully argue that it will lead more work.


Another example of this is the SEC, one guy only pays ten dollar entry (pensioner?) gets equal with 2 others in the minorest of prizes and whinges that his share is only 20 or thirty bucks. He he had a hot meal thrown in as well. He also got better than what he paid for.

AC, it is not all about you and the SEC. Your analogy is in no way relevant.


I won't try and make the same comparison with you as Admin of this BB.

Now didn't you miss me yesterday?

You were gone?

Kevin Bonham
07-09-2005, 01:53 PM
I think you just misunderstood me originally. My context has always been regarding the wider question. Duff's case was just an example of the equality of the rating system being called into question.

In #41 I asked "Are there really people out there who believe that their ratings or the ratings of their friends, relatives, or other connections are being deliberately diddled? Has anyone here heard anything of that sort?" In #43 you responded "I would say everyone has heard someone suggesting that at one time or another." In #44 I said "Well, I haven't heard anything of the sort for at least 15 years [..]" and then in #47 you wrote "Except the Speck case. I have also heard other less well publicised cases of varying degrees. Even things as innocent as Ronald's comment on the board a few days ago." Therefore I assumed you were referring to "Ronald's comment" as an example of "people out ther who believe that ... ratings ... are being deliberately diddled".


No it is not the same difference for the reasons I stated in my previous reply.

I'll assume you are alluding to the possibility that independent ratings checking services will be developed and that people might feel happier if someone else gets the same answer even though they still have no clue how those ratings were derived. If so fair enough.


Like Bill you will not Mischa's disinterest in this thread.

Yes but I am betting that the reasons are totally different.


The word is "hypothetical". I'll take your response as a no. And your argument that ratings are only important for 2100+ players as successfully rebutted.


You can misperceive my comments how you like, it doesn't make your perception correct. My response was flippant because the example was far-fetched, but for what it is worth it is very unlikely it should make a great difference to anything or anyone whether a 6 year old prodigy was rated 1600 or 1584 especially as said prodigy would probably be playing at 1800+ strength anyway.


Seen but irrelevant. It was one example of where the rating system is the butt of blame due to its inscrutibility.

If it is the butt of blame then what is a case of it being blamed for that reason? I would think I had more familiarity with selection controversies than most and since Glicko was brought in these are all the cases of recent selection controversies involving ratings I can recall (some may actually predate Glicko):

* Various players have been claimed to be underrated or overrated, including Speck, Wohl, Ronald Yu and Depasquale. In all cases (except the Speck one which was slightly different, see above) the suggestion has been that the maths of the system have treated the player incorrectly, not that there have been calculation errors, corruption or other process transparency issues.

* In one case it was discovered that a result for Rogers had been incorrectly coded as a loss when he had won. This resulted in the ACF deciding to release player game results to the applicants as well as the selectors so they could check for correct coding. This is an issue about transparency of coding, not transparency of calculation.

* Several years ago there were numerous accusations about a 10-game private match as a result of which one player's rating rose substantially at a time when that player could benefit from the rating rise (I think as an Olympiad applicant). Nothing was proven but the ACF closed the potential loophole by banning the rating of private matches.


There are reasons of right as well as utility. Your favourtite philosopher may think that rights are bunk but that is hardly the view of the wider community as is evidenced by the existence of legislation and other legal paraphanlia to protect them.

The existence of legal rights is no evidence for the existence of rights as matters of objective philosophy - rather it is evidence for the strength of desire (whether originating from belief, pragmatic motive or whatever) and ability of individuals to enforce those rights for whatever varying reasons they wish to have them. Your lucky word of the day is "intersubjectivity" and I would be more than happy to explore (or perhaps in my case "explain" is a better word) this further in non-chess. There must be an old objective/subjective morality thread there somewhere.


There are examples, just no examples which have not been credited. This is either evidence of very little or no corruption, or else very good tweaking. The fact remains that people are suspicious and the policy of secrecy fuels such speculation.

If this is a fact, then why can't you prove it? Or even if you can prove that people are "suspicious", why should we assume their suspicion is reasonable and amenable to rational evidence, rather than paranoid, gletsophobic or otherwise silly?

antichrist
07-09-2005, 02:10 PM
Bill previously advised 5-10/year and no increase since moving from Elo to Glicko to Glicko-2.

A/C
What has this to do with the issue and your quote of mine?


RW
I'll thank you to keep your personal issues in the non-chess forums. (And try to not make them offensive next time you do that).

A/C
That was meant at anyone who complains when there is not a glaring inaccuracy


RW
Several issues have been addressed over time but this is not about wrong ratings but more repeatablity and removal of subjective influences. Increase of transparency. Giving the customers the information to more reliably calculations and access to the data they are paying for.

AC
They are not customers, and they only pay about two bob for ratings anyway. Let them pay a fee for checking ($10 at least) and only refundable if there is a mistake in the rating.


RW
Nearly all ACF positions are voluntary how is this significant unless you can successfully argue that it will lead more work.

AC
It will definitely be more work for Bill answering all the silly questions from people with fragile egos - not being specific here. Otherwise they would not be asking the questions in the first place.


RW
AC, it is not all about you and the SEC. Your analogy is in no way relevant.

AC
It was relevant and more relevant since you mention customers having expectations. As that SEC whinger also had expectations. We are all members of a voluntary assoc and should not bear too heavily on voluntary workers. Just as KB got sick of the extra monitoring of Matt. But I know you like monitoring me - a S & M relationship.


RW
You were gone?

AC
But not forgotten - this is a tale of Johnny Rotten

PHAT
07-09-2005, 05:02 PM
An old aunty of mine would use an expression to show how mean a person might be - even with stuff they could give away, at no cost whatsoever to themselves.

"Joe Bloggs is so mean he wouldn't give you the steam off his piss."

I reckon KB is going along with this ACF meanness. He argues as if a full environmental impact study must be undertaken before you can even see the steam off ACF piss.


Kevin, shut up. You are sand in the gearbox and an imediment to progress.

Oepty
07-09-2005, 05:37 PM
Good on you Scott - you're getting at a point that Barry for some reason didn't pick up on when I kept repeating the line about "why should anyone care about ratings fiddles when any large scale fiddle would be detectable and any small scale one not worth the bother"?

There are contexts where small rating differences can be crucial. The Aus Champs is not really one of them because a player just below the cutoff who applies and has ever shown form at that level is almost always let in (and a player who only just gets in is hardly going to win the thing anyway).

All very true, but it might not be the case where you are talking about instead of a national championship, a state championships or a club championship. Not as significant as Australian Champs overall but extremely important to individuals.



Also typically selectors ignore very small rating differences unless there is nothing else to split the candidates at all, in which case the selection is pretty much a lottery anyway.

However one context in which a small rating difference can be decisive is an emergency selection where time is limited based on the average of the applicant's ACF and FIDE ratings.

Kevin. I have been told in the recent past, 3 or 4 olympiads ago one selector submitted there selections in the rating order of applicants. This is only a rumour, and even if true may not mean the selector did just list players in ratings order without considering other factors. If what I was told is true it did effect the team selected. I am not sure whether the ratings order was ACF or FIDE ratings order.
Scott

rob
07-09-2005, 06:31 PM
Kevin, shut up.
Just because you don't agree with Kevin's opinion on this matter it does you no credit resorting to try to suppress his right to comment. I didnít think you liked dictators, yet you appear to be dictating to Kevin.

You are sand in the gearbox and an imediment to progress.
Not all changes lead to progress.

PHAT
07-09-2005, 08:07 PM
Just because you don't agree with Kevin's opinion on this matter it does you no credit resorting to try to suppress his right to comment. I didnít think you liked dictators, yet you appear to be dictating to Kevin.

Furthermore, I am not trying to suppress his right to comment, I am trying to suppress his wish to comment any more on this topic. :cool: Nevertheless, the chances of me actually dictating anythink to KB is close to zero. :rolleyes:



Not all changes lead to progress.

But, there is no progress without change. :doh:

eclectic
07-09-2005, 08:48 PM
they ought to have eeg wires connected from your head to the dgt to record your mood during a game so that can also be factored into your rating figure ... :whistle:

eclectic

eclectic
08-09-2005, 01:46 PM
kegless

congratulations on your new appointment!

(for others: ... this week's acf newsletter)

;)

eclectic

Kevin Bonham
08-09-2005, 02:51 PM
All very true, but it might not be the case where you are talking about instead of a national championship, a state championships or a club championship. Not as significant as Australian Champs overall but extremely important to individuals.

True - I'd be interested to know how widespread the use of rigid ratings cutoffs for championships like this was.


Kevin. I have been told in the recent past, 3 or 4 olympiads ago one selector submitted there selections in the rating order of applicants. This is only a rumour, and even if true may not mean the selector did just list players in ratings order without considering other factors. If what I was told is true it did effect the team selected. I am not sure whether the ratings order was ACF or FIDE ratings order.

Probably before my involvement with selections so I can't really comment but if a selector actually said something like "oh I just listed the players in rating order without thinking about anything else" that would be concerning. However if a selector just happened to come up with selections that matched the rating order by coincidence, that's not so unusual. Also, some selectors might use the rating order as a starting point and then say "based on everything else I know about the candidates, is there any reason to change the order given by the ratings?"


I reckon KB is going along with this ACF meanness. He argues as if a full environmental impact study must be undertaken before you can even see the steam off ACF piss.

Hey, I'm a consultant ecologist; arguing for the precautionary principle even though we don't believe in it is a survival mechanism for our kind. :owned:

I'll admit it: I really don't have such strong feelings on this issue and I mainly got involved in this debate because I smelt a Rattus rattus in the leadup to the thread getting started and wanted to lure it out of its burrow. That said, I do think any issue to do with supposed corruption is one where people need to tread carefully in how they express themselves - I do care about that.

Oepty
08-09-2005, 06:17 PM
True - I'd be interested to know how widespread the use of rigid ratings cutoffs for championships like this was.

The entry to the SA Champs in the winner of the previous years reserves plus the seven highest rated players who have entered. I don't think this has not had to be used for a while in SA unfortunately, the number of entrants is ussaully lower than eight. Note the swiss tournament held this year was only held because there was not the required number of entries.

Just one further point on the Australian Championship qualification. A player that might actually be selected to played in the champs may not actually apply because they don't want to go through the alternative process.

Scott

Kevin Bonham
08-09-2005, 10:31 PM
Just one further point on the Australian Championship qualification. A player that might actually be selected to played in the champs may not actually apply because they don't want to go through the alternative process.

I'd tend to consider that their loss, if so. It's by no means burdensome, you just say you want to play and the committee does the rest.

PHAT
08-09-2005, 11:09 PM
I'll admit it: I really don't have such strong feelings on this issue and I mainly got involved in this debate because I smelt a Rattus rattus in the leadup to the thread getting started and wanted to lure it out of its burrow. That said, I do think any issue to do with supposed corruption is one where people need to tread carefully in how they express themselves - I do care about that.

[I declare that I have not got inside info on this from my erstwhile Wollongong clubmate.]

Here is my bet. Bazzcox wants the RDs so that he can improve his rating calculator. Nothing more, nothing less. I reckon he would be happy to receive the RDs on a confidentially arrangement. I also reckon that if he was given the algorithm for the generation of RDs, he could be trusted to be the independant arbitor. Players could then rest easy knowing that the black box has a viewing hole and someone's eye is at it.

Everyone would be h,a,p,p,y.
Bill glets to keep his secrets.
Players have more confidence that their confidence level [RD] is correct.
And we can all get some bleep sleep.

Rincewind
08-09-2005, 11:15 PM
[I declare that I have not got inside info on this from my erstwhile Wollongong clubmate.]

Erstwhile!!! What do you mean by bleeping erstwhile???

Mischa
08-09-2005, 11:16 PM
what's with the incorrect spelling of woollygolightly

PHAT
08-09-2005, 11:37 PM
Erstwhile!!! What do you mean by bleeping erstwhile???

Were you not M.I.A last Tuesday night?

Mischa
08-09-2005, 11:41 PM
hey if he stood you up then this is not the place to be discussing it .
I have it on secret that he was actually with Mick on Wednesday 'cos On Tuesday he had another date.....
Perhaps you should re think your relationship

PHAT
09-09-2005, 12:09 AM
hey if he stood you up then this is not the place to be discussing it .
I have it on secret that he was actually with Mick on Wednesday 'cos On Tuesday he had another date.....
Perhaps you should re think your relationship


No no no. Tuesdays is our club night and the C/ships are in full swing - I didn't see him there, that is all.

And for your information I don't date married men. :snooty:

Rincewind
09-09-2005, 12:26 AM
No no no. Tuesdays is our club night and the C/ships are in full swing - I didn't see him there, that is all.

Mischa has it right. My opponent (Mick) couldn't make it Tuesday so we agreed last week to play at my place Wednesday. The result was 1/2-1/2.


And for your information I don't date married men. :snooty:

Hey, I don't date men, full-stop. But single guys are ok with you?

PHAT
09-09-2005, 12:40 AM
Hey, I don't date men, full-stop*. But single guys are ok with you?

No, it is the thought married men that makes me gag.

Of married women? They have as much at stake as their secret married man. Singles are just too phkn dangerous.

Rincewind
09-09-2005, 07:42 AM
And the thread drifts on...

I will be getting back and responding to some of the more on-topic posts (I hope) over the weekend.

Alan Shore
14-09-2005, 05:22 PM
Um, why aren't they published? They're available on FICS for all to see.

Because they're the secrets that allow Bill to keep his executive powers as ratings officer? Conspiracy theories? Bill is surprisingly quiet on the matter.

PHAT
14-09-2005, 08:06 PM
Um, why aren't they published? They're available on FICS for all to see.

Because they're the secrets that allow Bill to keep his executive powers as ratings officer? Conspiracy theories? Bill is surprisingly quiet on the matter.

:eek: Gee wizz. We hadn't noticed ;) :lol:

I wonder if the RDs are available through FOI. Does anyone know if FOI applies to, say, QCA asking for rating info from the ACF, of which it is a member. :hmm:

Kevin Bonham
14-09-2005, 10:16 PM
Bill is surprisingly quiet on the matter.

I would advise strongly against reading anything specific into that. :P

Alan Shore
15-09-2005, 12:35 AM
I would advise strongly against reading anything specific into that. :P

Ah yeah, I see you've been TCB for him ;)