PDA

View Full Version : ACF Dec 2003 V FIDE Jan 2004

Bill Gletsos
01-01-2004, 05:47 PM
Starter,

I have done a comparison between the Dec 2003 ACF ratings and the Jan 2004 FIDE ratings.

Back when we did the comparison between the ACf Dec 1999 ratings and the Jan 2000 FIDe ratings we ignored anyone who had an ACF rating below 1900. This resulted in an average difference of 150 points. Hence the 150 point uplift added to players ratings prior to the calculation of the ACF April 2000 ratings.

This time we have ignored anyone with a ACF rating below 2050. This time the difference is 68 points.

Now someone is surely going to argue that players lost most of the the 150 points that were added to virtually all players Dec 1999 ratings. If they did this they would be incorrect.

There are currently 212 FIDE rated players who have played games in the ACF System since the Dec 1999 rating period. The average ACF rating of these 212 players back in Dec 1999 after the 150 points was added was 1976.038.
The average rating for this same group of 212 players in December 2003 is 1975.775. A difference of only 0.283.

Therefore this would indicate that there is no loss of 150 points but simply a redistribution of the points amongst the players, a situation that is to be expected unless someone is silly enough to argue that all 212 players are static and have neither improved or declined during this 4 year period.

Since there is no loss of rating points within the ACF pool, it therefore seems that the FIDE pool has inflated by 68 points during this same 4 year period.

Bill Gletsos
13-01-2004, 09:13 PM
Whilst everyone has been doing barry's marble puzzle, I have been busy collating the following data.

The following shows the figures for the FIDE January rating list for particular years.

The data shown is in the following order:
a) Year
b) Highest rated player on list
c) The rating of player 100 on the list
d) Number of players over 2600
e) Number of players over 2700
f) Average rating of the top 20 players
g) Average rating of the top 50 players
h) Average rating of the top 100 players

2004 2831 2606 117 18 2727.85 2687.28 2654.21
2003 2847 2598 98 14 2723.55 2683.68 2650.42
2002 2838 2596 90 13 2722.10 2681.74 2647.00
2001 2849 2595 92 12 2721.35 2681.16 2646.67
2000 2851 2595 89 11 2715.55 2675.46 2642.55
1999 2812 2587 83 10 2708.10 2667.34 2634.90
1998 2825 2584 76 8 2705.50 2664.00 2634.80
1997 2795 2580 66 8 2696.00 2657.30 2625.20
1996 2775 2580 65 8 2694.25 2654.40 2623.40
1995 2805 2575 60 9 2687.50 2648.10 2618.55
1994 2740 2565 56 5 2674.50 2638.70 2611.25
1993 2805 2565 54 4 2670.75 2638.60 2609.30
1992 2780 2555 41 3 2657.25 2625.70 2598.45
1991 2800 2545 33 3 2655.25 2620.00 2589.20
1990 2800 2540 29 2 2641.75 2607.80 2579.75
1989 2775 2530 28 2 2634.50 2603.80 2573.00
1988 2750 2520 18 2 2633.50 2597.40 2567.05
1987 2735 2515 13 2 2618.25 2585.70 2557.70
1986 2720 2505 14 2 2619.50 2581.50 2550.45
1985 2715 2500 12 2 2615.50 2577.10 2546.80
1984 2710 2490 17 2 2621.25 2577.70 2540.00
1983 2710 2495 13 1 2614.75 2576.10 2540.80
1982 2720 2495 15 1 2617.25 2580.50 2542.10
1981 2690 2495 15 0 2617.75 2578.00 2544.10
1980 2725 2500 14 2 2621.50 2580.80 2548.10
1979 2705 2500 11 1 2615.25 2577.20 2546.30
1978 2725 2500 12 1 2613.75 2577.90 2547.35
1977 2690 2495 12 0 2613.00 2576.70 2545.35
1976 2695 2490 13 0 2613.75 2573.20 2540.40
1975 2780 2490 18 2 2629.25 2580.20 2544.05
1973 2785 none 14 1 2623.50 none
1971 2740 2480 16 1 2626.00 2583.40 2542.60

DoroPhil
15-01-2004, 10:09 PM
Hi Bill,

Does all this mean that every player currently on the rating list will receive 68 additional rating points on the ACF March rating list?

Bill Gletsos
15-01-2004, 11:20 PM
Possibly. ;)

Garvinator
15-01-2004, 11:25 PM
silly question now, but why would players get 68 more points? :oops:

15-01-2004, 11:30 PM
Because of the inflation by FIDE.

Garvinator
15-01-2004, 11:37 PM
oh no another inflation deflation debate ](*,) #-o

15-01-2004, 11:42 PM
Debate?

Bill Gletsos
15-01-2004, 11:43 PM
Because of the inflation by FIDE.
Well done. =D>

Bill Gletsos
15-01-2004, 11:45 PM
oh no another inflation deflation debate ](*,) #-o
Pay more attention and re-read the first post.
But more importantly :-k .

Bill Gletsos
15-01-2004, 11:46 PM
Debate?
He is probably referring to a debate on the old board regarding whether the ACF ratings were deflating.
However it would appear he has failed to undersrand the facts of the first post and the historical implication of the second post.

Therefore pay no attention to him, he doesnt understand. ;)

Garvinator
16-01-2004, 09:34 AM
Debate?
He is probably referring to a debate on the old board regarding whether the ACF ratings were deflating.
However it would appear he has failed to understand the facts of the first post and the historical implication of the second post.

Therefore pay no attention to him, he doesnt understand. ;)

bill, bill, bill. actually i was trying to anticipate some ppl saying that some or all players should get 68 points added to their acf rating. I had read the post the first time and then reread it after you claimed i should go back and re-read it.

I was also trying to anticipate that old debate again and trying to kill it off before it started 8-[

Bill Gletsos
16-01-2004, 10:08 AM
bill, bill, bill. actually i was trying to anticipate some ppl saying that some or all players should get 68 points added to their acf rating.
It would appear you seem to think they should not get 68 points added.

If this is the case then you obviously didn't pay as much attention to information posted on the BB or the ACF bulletins as I had believed. :(

The original post on this board is a carry over from the old ACF BB.

On the old BB I said the following.

the motion passed at the last ACF Council meeting:

That the ACF Rating Officers are authorised to adjust the ACF ratings as they see fit to try and bring the ACF ratings more in line with the FIDE ratings. This correction to take place where determined necessary prior to or at the same time as the publication of the first ACF rating list for each calendar year.
and

However keep in mind the Council also passed the following motion:

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
Graham and I are still currently working on this one.

Those motions were also reported in an ACF bulletin last year.

Now it is clear from the first motion that the aim is to try and keep the ACF and FIDE ratings in line as much as possible. THus if there is no deflation in the ACF system but the ratings have divereged by 68 points then there must be FIDE inflation. To remove this discrepancy then obviously points need to be added to the ACF ratings.

Garvinator
16-01-2004, 10:15 AM
It would appear you seem to think they should not get 68 points added.

I guess in some ways i dont think there should be 68 points added :oops: fide and acf use two different rating systems(as i understand it) so then it makes sense to me that there will be a disparity from time to time. For some reason, I feel that if the acf ratings are going to be increased, this time by 68 points, each time the fide ratings are higher, then we might as well use elo :o

Bill Gletsos
16-01-2004, 10:21 AM
I guess in some ways i dont think there should be 68 points added :oops: fide and acf use two different rating systems(as i understand it) so then it makes sense to me that there will be a disparity from time to time. For some reason, I feel that if the acf ratings are going to be increased, this time by 68 points, each time the fide ratings are higher, then we might as well use elo :o
Then you clearly do not understand. ](*,)
It isnt a glicko V elo issue.
Even when the ACF used the ELO system for ratings calculations there was a difference between ACF and FIDE ratings. This difference increased between 1980-1999 such that in 1999 there was a 150 point difference. Hence the 150 point uplift in April 2000.

Garvinator
16-01-2004, 10:23 AM
ok then, oh dear i get the feeling im going to have to ask bill to repeat himself :( , why do our ratings fall behind alot of the time?

ursogr8
16-01-2004, 10:24 AM
It would appear you seem to think they should not get 68 points added.

I guess in some ways i dont think there should be 68 points added :oops: fide and acf use two different rating systems(as i understand it) so then it makes sense to me that there will be a disparity from time to time. For some reason, I feel that if the acf ratings are going to be increased, this time by 68 points, each time the fide ratings are higher, then we might as well use elo :o

ggrayggray

I have been one encouraging Bill to keep ACF ratings in synch/scaling with FIDE ratings because I find many players in Victoria seduced into wanting as high a rating as possible. So our ACF product needs to be scaled up to match the FIDE product. Otherwise our revenue stream is in jeopardy from FIDE offering a 'higher' product.
Once we have parity established then the arguments on which rating system to use turns to the criteria of relative cost and accuracy. Bill's product is superior on both these criteria.

starter

How's that Bill. Good support? Can I have my 68 points now? :D

arosar
16-01-2004, 10:24 AM
Look Garvin mate....just stick to chess OK. You'll go nuts thinking and talking about bloody ratings. Did you win that game on ICC last night?

AR

Bill Gletsos
16-01-2004, 10:34 AM
ok then, oh dear i get the feeling im going to have to ask bill to repeat himself :( , why do our ratings fall behind alot of the time?
Oh dear, oh dear, oh dear. ](*,) ](*,) ](*,)

Ours dont fall behind.
The FIDE ratings inflate and Australian players below 2100 ACF are generally overrated on the FIDE scale.

Bill Gletsos
16-01-2004, 10:35 AM
How's that Bill. Good support? Can I have my 68 points now? :D
:D :D

You may be lucky and get 70. ;)

Rincewind
16-01-2004, 10:43 AM
You may be lucky and get 70. ;)

Is this a part of a plot to stop Wollongong from winning the u-1800 grade matches this year? ;)

Garvinator
16-01-2004, 10:48 AM
Look Garvin mate....just stick to chess OK. You'll go nuts thinking and talking about bloody ratings. Did you win that game on ICC last night?

AR
yeah i won that game you saw. in the final i only drew which meant i came second in the bracketed elimination final system they use. i had 15 secs left and was about to queen a pawn to get two queens and a bishop v queen, but i queened the pawn and allowed perpetual check :(

ursogr8
16-01-2004, 11:04 AM
You may be lucky and get 70. ;)

Is this a part of a plot to stop Wollongong from winning the u-1800 grade matches this year? ;)

Can it be true? :shock:
Are you suggesting that NSW don't adjust their grades criteria after Bill adds 150 and 68 to the ACF ratings. :rolleyes:
Not only can we sell you the VIC 'States are an Association of Clubs' concept but we could also sell you our Grades rules. ;)

Bill Gletsos
16-01-2004, 11:17 AM
Are you suggesting that NSW don't adjust their grades criteria after Bill adds 150 and 68 to the ACF ratings. :rolleyes:
Given everybody gets the same increase no player is disadvantaged.

Not only can we sell you the VIC 'States are an Association of Clubs' concept
Been there done that and rejected it back in the 80's. ;)

but we could also sell you our Grades rules. ;)
Maybe you should check out NSW's before making that comment.

Rincewind
16-01-2004, 11:53 AM
Can it be true? :shock:
Are you suggesting that NSW don't adjust their grades criteria after Bill adds 150 and 68 to the ACF ratings. :rolleyes:

Why would they when the boundaries are arbitrary?

The only issue is having sufficient teams in each division to make a contest of it. Traditionally, the u1800 and u1600 are the most hotly contested as that's where the majority of players live.

Wollongong having a small player base may be greatly impacted by a raing shift. Were fortunate last year in have a full team + reserve in the upper end of the u1800 division. So while we were competitive on boards 1 and 2 (even if I do say so myself :oops: ) we usually had a fair rating advantage on the lower two boards.

But my comment was just in jest. If we have to play in the u2000, so be it.

My comment was more a response to something Bill said to me at the presentation night last year. The comment was made by someone that we would be able to play in the u1800 grade next year. I said "What are you talking about? My rating is going down!" To which Bill responded "Well, I can fix that". :D

Perhaps this 70 points adjustment was going through Bill's head even then. However, me personally it should affect me too much. If my rating hits 1750 in March as I estimate, an extra 70 points will only put me at 1820. So I will still be able to play board 1 provided we don't have any other over 1800 players in the team. But it would probably force our previous board 1 up to the next grade. :(

Bill Gletsos
16-01-2004, 12:04 PM
[My comment was more a response to something Bill said to me at the presentation night last year. The comment was made by someone that we would be able to play in the u1800 grade next year. I said "What are you talking about? My rating is going down!" To which Bill responded "Well, I can fix that". :D
Yes, I remember that conversation. :D
And you are correct, I was aware of a possible increase at that time, although all I was fairly certain of was that it would be at least 50 points.

Cat
16-01-2004, 12:19 PM
While you're in such a good mood Bill, can I suggested a couple of things for your esrtwhile consideration;

1. Regional parity - this would not be difficult to achieve by simply comparing interstate performance and standardising the results using NSW as the standard.

2. There is still the issue of regional stabilisation for the higher rated players. It remains the greatest single deterrant preventing our strongest players competing regularly in local competition. Its obviously a major problem in Box Hill and Canberra as well, and your suggestion of quaranteening the adults against junior contamination inhibits rather than encourages participation. With Starter talking about a rating product, why not make that product more player-friendly? Maybe in part this problem can be helped by considering the issue of regional parity.

Oepty
16-01-2004, 05:02 PM
I strongly disagree with the adding on of points to match the superior ACF rating system to the inferior FIDE rating system. Why should we corrupt the ACF rating system just to match with FIDE system? It just seems pointless. The FIDE ratings are becoming so much higher very quickly that the figures themselves are almost becoming a joke. I don't want to see ACF ratings of over 3o00 which will happen eventually. Bill you should be lobbying for FIDE (or lobbying the ACF to lobby FIDE) to change there rating system to one that doesn't inflate.
Scott

Bill Gletsos
16-01-2004, 05:38 PM
I strongly disagree with the adding on of points to match the superior ACF rating system to the inferior FIDE rating system. Why should we corrupt the ACF rating system just to match with FIDE system? It just seems pointless. The FIDE ratings are becoming so much higher very quickly that the figures themselves are almost becoming a joke. I don't want to see ACF ratings of over 3o00 which will happen eventually. Bill you should be lobbying for FIDE (or lobbying the ACF to lobby FIDE) to change there rating system to one that doesn't inflate.
Scott
Actually no corruption of the ACF system is involved.
Just think of it as parity pricing. :D

As for getting FIDE to change their system, apart from the fact I seriously believe they see it as an advantage, I think you would have buckleys getting it changed.

Bill Gletsos
16-01-2004, 05:42 PM
While you're in such a good mood Bill, can I suggested a couple of things for your esrtwhile consideration;
I'm always in a good mood. It is just sometimes I'm in a better mood. ;)

1. Regional parity - this would not be difficult to achieve by simply comparing interstate performance and standardising the results using NSW as the standard.
I'm not prepared to concede that their are regional differences. :-''
Take Gold Coast CC for example.
Our research showed that if all games amongst all GC members were ignored then if you lloked at all GC members as a group the average rating change of a GC member was around 1.0 rating point. Some would go up but others would have gone down. When we made the adjustment to GC ratings in November we decided to not disadavantage anyone by reducing their rating. The impact of this is that we added approx 3800 points to the GC pool. Eventually those points will be correctly distributed across the GC, QLD and ACF pools due to player interactions.

2. There is still the issue of regional stabilisation for the higher rated players. It remains the greatest single deterrant preventing our strongest players competing regularly in local competition. Its obviously a major problem in Box Hill and Canberra as well, and your suggestion of quaranteening the adults against junior contamination inhibits rather than encourages participation.
There you go making unfounded statements again. :rolleyes:
It is not a major problem in Canberra.
It does not appear to ever having been a problem in Canberra.
The junior problem there showed no indication that it was preventing adult participation.

With Starter talking about a rating product, why not make that product more player-friendly? Maybe in part this problem can be helped by considering the issue of regional parity.
It has nothing to do with being player friendly or unfriendly. It has to do with predictive accuracy.

Cat
16-01-2004, 06:26 PM
I'm not prepared to concede that their are regional differences.

Whether you concede it or not, doesn't it make sense to attempt to ensure 2000 in QLD is the same as 2000 in NSW &amp; Victoria, etc.?

There you go making unfounded statements again.
It is not a major problem in Canberra.
It does not appear to ever having been a problem in Canberra.
The junior problem there showed no indication that it was preventing adult participation.

OK so its only a major problem on the Gold Coast &amp; Box Hill. Isn't that enough?

Bill Gletsos
16-01-2004, 06:36 PM
I'm not prepared to concede that their are regional differences.

Whether you concede it or not, doesn't it make sense to attempt to ensure 2000 in QLD is the same as 2000 in NSW &amp; Victoria, etc.?
I would argue that they are.

There you go making unfounded statements again.
It is not a major problem in Canberra.
It does not appear to ever having been a problem in Canberra.
The junior problem there showed no indication that it was preventing adult participation.

OK so its only a major problem on the Gold Coast &amp; Box Hill. Isn't that enough?
I would suggest that failure of players to play in tournaments with many juniors is not a rating issue but a marketing one. Perhaps you need to run tournaments that will attract those players.

Cat
16-01-2004, 06:51 PM
I would argue that they are.

And if they weren't, either now or sometime in the future, would you attempt to standardise them? Hypothetically speaking of course.

I would suggest that failure of players to play in tournaments with many juniors is not a rating issue but a marketing one. Perhaps you need to run tournaments that will attract those players.

One way to market to them is to say, 'come and play our juniors. We know they're good - very good, better than their rating suggests. We know that playing them might put a dent in your rating, and we can see that that might be a bit rough. So I'll tell you what, we'll put some stability back into your rating.' How does that sound?

Bill Gletsos
16-01-2004, 08:06 PM
I would argue that they are.

And if they weren't, either now or sometime in the future, would you attempt to standardise them? Hypothetically speaking of course.
I'm not prepared to second guess myself as to what I may or may not do in the future. 8)

I would suggest that failure of players to play in tournaments with many juniors is not a rating issue but a marketing one. Perhaps you need to run tournaments that will attract those players.

One way to market to them is to say, 'come and play our juniors. We know they're good - very good, better than their rating suggests. We know that playing them might put a dent in your rating, and we can see that that might be a bit rough. So I'll tell you what, we'll put some stability back into your rating.' How does that sound?
Do what you feel you must. :rolleyes:

16-01-2004, 08:46 PM
I'm not prepared to second guess myself as to what I may or may not do in the future. 8)

Thanks Bill, I just love this statement, :D ( I know I have taken it in isolation [-X )

If you can't guess what you are going to do what hope do we have? :-k

If Matthew was here I am sure he would have had some fun with this. \:D/

I think I will go and ban myself for a couple of hours as I have had a few glasses of medicinal Cabernet Sauvignon, don't want the Grand Poobah to get in first. :oops:

Kevin Bonham
16-01-2004, 09:01 PM
You're not too drunk if you can still spell what you've been drinking. :mrgreen:

Bill Gletsos
16-01-2004, 10:41 PM
I'm not prepared to second guess myself as to what I may or may not do in the future. 8)

Thanks Bill, I just love this statement, :D ( I know I have taken it in isolation [-X )
Personally I think it sums things up nicely. ;)

If you can't guess what you are going to do what hope do we have? :-k
Situations change. What you may do one day you may not the next.
Also as Barry showed in his marble puzzle just giving an intuitive answer can lead you astray. #-o
I try to avoid being wrong, especially when it can be avoided. 8)

If Matthew was here I am sure he would have had some fun with this. \:D/
He might, but I couldn't care less what his opinion would be. =;
He is prone to put his mouth in gear before engaging his brain. :-''

I think I will go and ban myself for a couple of hours as I have had a few glasses of medicinal Cabernet Sauvignon, don't want the Grand Poobah to get in first. :oops:
Enjoy. :)

antichrist
17-01-2004, 04:09 PM
I take it that no resolution was adopted that barred minor comps clashing with major comps?

Bill Gletsos
17-01-2004, 04:19 PM
I take it that no resolution was adopted that barred minor comps clashing with major comps?
What has that question got to do with the topic.

Oepty
19-01-2004, 02:50 PM
Bill. I am not sure what parity pricing is so your analogy is lost on me.

By changing the ACF rating system, by adding on points to match an inflating system is just making the stable ACF system into a system that inflates. If that is not corrupting the system I do not know what is.
Also I think that players like to keep track of whether there rating is going up and down. This is made more difficult by the adding on of points that have nothing to do with performance.
I understand that there has been some teething problems with the system which has meant there has been some recalculations of the ratings, but I think that there has been enough fiddling with system. I believe there should be a line drawn under the ratings from December 2003 and if there has to be any more changes to the system it should be done based on December 2003 ratings not ratings from 2000. The effect, which I don't believe was deliberate, of all of the recalculations that have been done is I and anyone else who wishes to our own analysis on the rating system cannot because we don't have the ratings. Although I think the Glicko 2 system is the best for Australia I think some effort needs to be made to make the system more open and accessable. I don't mean you should be teaching people what the Glicko 2 system is, that is up to an individual but the means should be made available for people to look at the system.

Scott

Bill Gletsos
19-01-2004, 03:20 PM
Bill. I am not sure what parity pricing is so your analogy is lost on me.
I'll leave that as an exercise for you to research. ;)

By changing the ACF rating system, by adding on points to match an inflating system is just making the stable ACF system into a system that inflates. If that is not corrupting the system I do not know what is.
It would seem you do not understand what is stable and what isnt or what is corrupting and what isnt.

If the system is stable then if the same number of points is added to every players rating that does not make it suddenly unstable. Likewise if the same number of points is deducted from every one.

Also I think that players like to keep track of whether there rating is going up and down. This is made more difficult by the adding on of points that have nothing to do with performance.
Its a straight addition of points. Its just basic maths.
It isnt that difficult to comprehend.

I understand that there has been some teething problems with the system which has meant there has been some recalculations of the ratings, but I think that there has been enough fiddling with system. I believe there should be a line drawn under the ratings from December 2003 and if there has to be any more changes to the system it should be done based on December 2003 ratings not ratings from 2000.
In some respects I agree but in some I disagree.

For example switching from Glicko to Glicko2 wouldnt have been such a major improvement if we had only implemented it from Dec 2002 instead of doing what we did and rerunning the syetm under glicko2 from Dec 2000. Likewise with the introduction from June 2003 of the changes to to rating calculations for players over 2200.

The effect, which I don't believe was deliberate, of all of the recalculations that have been done is I and anyone else who wishes to our own analysis on the rating system cannot because we don't have the ratings.
That may well be true, but those people who wish to do their own analysis on the rating system are in a very significant minority. I would suggest that there would be less than a dozen people in Australiai who would do any sort of analysis on the rating system as a whole over multiple rating periods.
I would contend that accuracy is more important than satisfying the needs of that minority.

Although I think the Glicko 2 system is the best for Australia I think some effort needs to be made to make the system more open and accessable. I don't mean you should be teaching people what the Glicko 2 system is, that is up to an individual but the means should be made available for people to look at the system.
This is a seperate issue to the point above.
Barry Cox's Glicko calculator is very good at estimating a players new rating.

antichrist
20-01-2004, 03:45 PM
I take it that no resolution was adopted that barred minor comps clashing with major comps?
What has that question got to do with the topic.

It may not fit in with the topic, but I thought at one stage the ACF was considering banning such comps. As this was the next meeting of the ACF as far as I knew so I thought it may have come up.

Bill Gletsos
20-01-2004, 03:48 PM
Well you moron why didnt you just start another topic.

antichrist
20-01-2004, 04:09 PM
Well you moron why didnt you just start another topic.

Without all the fillerbusting you could have answered the question and I would be out of here. Don't forget those Yugoslav clocks.

Oepty
22-01-2004, 01:08 PM
Bill. My main point is that by adding points to a stable rating system, you make it effectively the same as having a inflationary system of the same amount. Therefore I contend that by adding points you are making the ACF rating system inflationary and therefore corrupting it.
You are of course right that it is simple maths to take off the amount that has been added on, but I still would prefer not to do it.
Can you please tell me what my ratings where on the December 2000, 2001, 2002 lists?
Scott

Bill Gletsos
22-01-2004, 01:32 PM
Bill. My main point is that by adding points to a stable rating system, you make it effectively the same as having a inflationary system of the same amount. Therefore I contend that by adding points you are making the ACF rating system inflationary and therefore corrupting it.
That is not correct.
If a system is inflating the only ratings that can inflate are those of players who are actually playing games. Players who have been inactive cannot inflate and are in effect deflated against the active players.
However we are adding the points to all the players whether active or inactive. Therefore there is no inflation.

Can you please tell me what my ratings where on the December 2000, 2001, 2002 lists?
Unr, Unr, 1201

Rincewind
22-01-2004, 01:33 PM
Bill. My main point is that by adding points to a stable rating system, you make it effectively the same as having a inflationary system of the same amount. Therefore I contend that by adding points you are making the ACF rating system inflationary and therefore corrupting it.
You are of course right that it is simple maths to take off the amount that has been added on, but I still would prefer not to do it.
Can you please tell me what my ratings where on the December 2000, 2001, 2002 lists?

Scott, no system is an absolute measure of anything. Ratings are all about differences. For example there is no logical reason for rating not to be negative.

Bill Gletsos
22-01-2004, 01:50 PM
The point Scott was making previously was that due to changes in the rating system and rerunning from the dec 2000 period, previously puiblished lists may be differnent to the actual ratings for those periods now. Hence he asked me directly for his rating in those periods.

Oepty
27-01-2004, 04:57 PM
Hello Bill. Okay I accept that there is that difference. I just don't want to gain rating points without actually gaining them by winning chess games. The fact I gain points means absolutely nothing. There seems to be no reason for doing this. Matching the clearly inferior FIDE rating system just seems stupid to me, but it also seems I am alone in my thoughts.
Hello Barry. Yes I agree that the ratings themselves have no absolute meaning, rather they are relatative to each other, but I don't think that fact gives any reason just to add points whenever it is felt there is a need to.
Scott

Rincewind
27-01-2004, 05:16 PM
Hello Bill. Okay I accept that there is that difference. I just don't want to gain rating points without actually gaining them by winning chess games. The fact I gain points means absolutely nothing. There seems to be no reason for doing this. Matching the clearly inferior FIDE rating system just seems stupid to me, but it also seems I am alone in my thoughts.

Scott, it is possible that the ACF rating system is deflating, meaning that people are losing rating points when they are not actually any worse a chess player. To determine whether this is a case you your first have to define how you would measure deflation.

Hello Barry. Yes I agree that the ratings themselves have no absolute meaning, rather they are relatative to each other, but I don't think that fact gives any reason just to add points whenever it is felt there is a need to.

of course there is also not ny reason NOT to. The reason to do is to maintain a parity and comparability between ACF and FIDE ratings. It can't be 100% of course but the best that can be done under the circumstances.

Oepty
28-01-2004, 12:39 PM
Barry. I think that trying to maintain parity with the FIDE rating system as a reason just stinks, but I seem to be alone in this view. I also doubt the ACF rating system is deflating, but if it is it should be fixed, not just having points added to it.
Scott

Cat
28-01-2004, 03:08 PM
OK Scott, seeing as you're so committed to purity, perhaps you could help me with this problem? Anne is a bright young prospect, plays regularly at the club, many games every week - shes rated 1000. Alex is an nice old boy, we see him from month to month - he's in decline, but also rated 1000. One day they both play Fred rated 1500 and win. Which of the following should be correct;

a) Anne should get more rating points than Alex for her win.

b) Alex should get more rating points that Anne for his win.

c) They should both get the same points for the same performance.

Garvinator
28-01-2004, 03:59 PM
c) They should both get the same points for the same performance.

I know the question wasnt directed at me, but my answer is above

Bill Gletsos
28-01-2004, 04:31 PM
Firstly this has nothing to do with the thread topic.

Under the Elo system they would both gain the same amount of points.

However David is actually trying to be clever and catch people out.

He knows that under Glicko the movement is based on the players RD. The higher the RD the higher the movement.

Now based on his comments that Anne plays many games a week on could assume her RD is low. Alex on the other hand is likely to have a higher RD than Anne. Therefore under Glicko Anne will go up less than Alex if this was the only game rated in the period.

Now under Glicko2 there is a volatility. If the RD assumtions arbove are true and neither Anne or Alex are volatile then Anne will go up less than Alex.
Even if Anne has a high volatility and Alex a low Anne is still likely to go up less than Alex as the RD is still more improtant when only considering one game.

Now David will start with the argument that Anne is a young active junior who is obviously improving. Of course their is no evidence in the scenario to support this claim. Therefore she should have gone up more than the old declining Alex. Likewise there is no evidence in the scenario that old Alex is declining.

Of course be that as it may, since this is all hypothesis with no evidence then one could hypothesise the following:
Old Alex had at one stage been closer to 1500, maybe even over 1500. He doesnt play at that strength on a regular basis now hence his 1000 rating but he has been know to play at that strength. Anne on the other hand is just a junior who has never palyed at 1500. She is improving. But just because she has improved over the past 2-3 rating periods does not mean she will this time. Therefore it is not unreasonable that Alex goes up more than Anne.

Maybe David does like this scenario. However as I said on the old BB, I'm prepared to follow Professor Glickmam than David when it comes to ratings theory.

Cat
29-01-2004, 09:20 AM
For what its worth, I'd agree with you Garvin.

Garvinator
29-01-2004, 09:46 AM
For what its worth, I'd agree with you Garvin.
:clap: in your example David, you didnt mention how many games each person had played, so therefore without that data, a person has to assume that it is the same amount for each person. Therefore the answer must be their rating is equally affected.

Had you mentioned their previous RD's, then of course their rating change would be different.

Oepty
29-01-2004, 10:17 AM
Bill has explained what woould happen under Glicko2 very well. In pure chess senses you might find that Anne won here game because Fred dropped his queen on move 7 while Alex played like a 2000 strength player in there game or vice versa. A rating system cannot, and I am not suggesting it should take this into account. I believe that the way Glicko2 handles it is probably the best. You have no idea what Alex has done in the month since he last played. Maybe he has spent a lot of time on chess because he has recognised that his level of play had been declining and wished to change that. The simple fact is we have little information about Alex's level of play compared to Anne who plays very regularly. I believe, the same as Professor Glickman, it should be treated as being less reliable, hence the higher RD.
Scott

Bill Gletsos
29-01-2004, 10:56 AM
:clap: in your example David, you didnt mention how many games each person had played, so therefore without that data, a person has to assume that it is the same amount for each person. Therefore the answer must be their rating is equally affected.
You made an assumption based on no evidence to support your assumption.
In fact it appeared David made a point of noting that Anne was a very active player and Alex was significantly less active. To ignore this piece of information seems dubious to me.

To correctly answer the question you should have allowed for a number of the most likely possibilites and answered based on those scenarios. :hmm:

I'd give you a fail on the question. :sad:

Garvinator
29-01-2004, 11:54 AM
You made an assumption based on no evidence to support your assumption.
[QUOTE]In fact it appeared David made a point of noting that Anne was a very active player and Alex was significantly less active. To ignore this piece of information seems dubious to me.

but in my opinion david did not say they were rated games for the acf, they could have just been social games at the club on a friday night.

To correctly answer the question you should have allowed for a number of the most likely possibilites and answered based on those scenarios. :hmm: i prefer just to answer the question as it is asked and then berate the questioner for a bad question :doh:

I'd give you a fail on the question. :sad:no sweeney marking systems thanks :p

Bill Gletsos
29-01-2004, 12:21 PM
but in my opinion david did not say they were rated games for the acf, they could have just been social games at the club on a friday night.

i prefer just to answer the question as it is asked and then berate the questioner for a bad question :doh:
In that case you should have berated him in the same post that you gave your answer.

no sweeney marking systems thanks :p
This is the Gletsos marking system. Its superior. :owned:

Cat
29-01-2004, 10:30 PM
You made an assumption based on no evidence to support your assumption.
In fact it appeared David made a point of noting that Anne was a very active player and Alex was significantly less active. To ignore this piece of information seems dubious to me.

To correctly answer the question you should have allowed for a number of the most likely possibilites and answered based on those scenarios. :hmm:

I'd give you a fail on the question. :sad:

I'd have to agree with Bill on this. The fact I didn't mention the RD's is irrelevant because the situation was clearly hypothetical.

Cat
31-01-2004, 12:01 AM
Bill has explained what woould happen under Glicko2 very well. In pure chess senses you might find that Anne won here game because Fred dropped his queen on move 7 while Alex played like a 2000 strength player in there game or vice versa. A rating system cannot, and I am not suggesting it should take this into account. I believe that the way Glicko2 handles it is probably the best. You have no idea what Alex has done in the month since he last played. Maybe he has spent a lot of time on chess because he has recognised that his level of play had been declining and wished to change that. The simple fact is we have little information about Alex's level of play compared to Anne who plays very regularly. I believe, the same as Professor Glickman, it should be treated as being less reliable, hence the higher RD.
Scott

OK, so this was in a local competition, Alex did well, but Anne did brilliantly. All their opponents played frequently and had very stable ratings. Amazingly they both played the same players. The cross-table looks like this;

Opponent(Rating)----------Anne-------------------Alex

Fred(1500)-----------------W----------------------W
Joe(1400)------------------W----------------------W
Ivan(1300)-----------------W----------------------L
Peter(1200)----------------W----------------------W
Catherine(1100)------------W----------------------L
Bruce(1000)----------------W----------------------W
Dave(900)------------------W----------------------W

Q1. Who will be higher rated after the competition, Alex or Anne?

Q2. What will their ratings be?

Again, the question is hypothetical, there are no absolute right answers, some interpretation is necessary.

Bill Gletsos
31-01-2004, 03:14 PM
OK, so this was in a local competition, Alex did well, but Anne did brilliantly. All their opponents played frequently and had very stable ratings. Amazingly they both played the same players. The cross-table looks like this;

Opponent(Rating)----------Anne-------------------Alex

Fred(1500)-----------------W----------------------W
Joe(1400)------------------W----------------------W
Ivan(1300)-----------------W----------------------L
Peter(1200)----------------W----------------------W
Catherine(1100)------------W----------------------L
Bruce(1000)----------------W----------------------W
Dave(900)------------------W----------------------W

Chess is good because it makes one consider all the possibilites and not take at face value a seemingly innocent move. :cool:

Others may just think you are just asking a simple question however experience shows that people often have other agendas.
I'm of the opinion that you have a predetermined answer and are just trying to justify it by a seemingly random hypothetical question.

For the moment I'll play your little game. :wink:

Q1. Who will be higher rated after the competition, Alex or Anne?
Based on the data in this and the original question it is reasonable to assume the following.
Anne possibly plays around 40-50 normal rated games in a rating period(this can be determined by looking at the figures for ACT and NSW juniors who are by far the most active juniors in the normal system). She would likely have an RD around 70. Here volatility is most likely high if she has been improving from rating period to rating period. If however she has not been improving at all for the past few rating periods he volatility will be normal. Note however that 7 games is not very significant as a measure of possible significant improvement.

Alex on the other hand probably plays 7-10 games in a rating period. He would likely therefore have an RD of around 90. His volatility is most likely normal.

Under Elo Anne will be higher rated.
Under Glicko Anne would be higher rated.
Under Glicko2 Anne would be higher rated and higher rated than under Glicko.

Q2. What will their ratings be?
Under Elo Anne would be 1074 and Alex 1044.
Under Glicko Anne would be 1139 and Alex 1116.
Under Glicko2 if Anne has a high volatility Anne would be 1149, if her volatility was normal around 1143. Alex would be 1116.

If Alex only played every second rating period then his RD would still be most likely no more than 100. If his RD had been 100 then his rating would be 1133 in both Glicko and Glicko2.

Now given 7 games isnt a good enough sample for the volatility to have a significant effect lets assume Alex and Anne score the 7 results above in either 3 or 4 tournaments all within the one rating period.

Now for 3 tournaments:
Under ELO Anne's rating would be 1223, Alex 1133.
Anne's Glicko would be 1308, Alex 1228.
With a normal volatility Anne's Glicko2 1410, with a high volatility 1446 and Alex 1232.

Now for 4 tournaments:
Under ELO Anne's rating would be 1298, Alex 1179
Anne's Glicko would be 1359, Alex 1261
With a normal volatility Anne's Glicko2 1528, with a high volatility 1545 and Alex 1265.

Again, the question is hypothetical, there are no absolute right answers, some interpretation is necessary.
Of course some answers are more correct than others. :hmm:

It is entirely reasonable that the player with the less reliable rating(higher RD) flutuates more than the player with a reliable rating. Even so irrespective of whether a player is reliable or un-reliable they may also be volatile or non volatile. The volatility factor will carry over from rating period to rating period and if the results in a rating period are significantly different then the rating increase will accelerate.

Cat
31-01-2004, 04:28 PM
Anne possibly plays around 40-50 normal rated games in a rating period(this can be determined by looking at the figures for ACT and NSW juniors who are by far the most active juniors in the normal system). She would likely have an RD around 70. Here volatility is most likely high if she has been improving from rating period to rating period. If however she has not been improving at all for the past few rating periods he volatility will be normal. Note however that 7 games is not very significant as a measure of possible significant improvement.

Alex on the other hand probably plays 7-10 games in a rating period. He would likely therefore have an RD of around 90. His volatility is most likely normal.

What a beautiful finesse! In fact, Anne loves her chess and has played so much her rating is very stable, with a RD of around 30 or !!. Alex, well his wifes been sick and he's struggling with late hours at work and you know, he hasn't played a rated game in 2yrs, he has a ?? rating. Now I wonder if this makes any difference?

Kevin Bonham
31-01-2004, 05:43 PM
What a beautiful finesse! In fact, Anne loves her chess and has played so much her rating is very stable, with a RD of around 30 or !!. Alex, well his wifes been sick and he's struggling with late hours at work and you know, he hasn't played a rated game in 2yrs, he has a ?? rating. Now I wonder if this makes any difference?

Maybe I get where this is heading, you want a scenario where some old rusty who has a few good results goes up faster than some junior who's doing spectacularly.

But if Anne has been playing so heavily her rating is a very stable 1000-odd (would RD really get down to 30 even for a very active junior?), then for her to suddenly, out of nowhere, put on that kind of a burst and get 7/7 vs an expected 2/7 - you'd have to be sceptical about how much of that was improvement vs fluke.

I had one of these kinds of performances in my mid-teens; having been rated around 1190 and only improving slowly I suddenly scored 8/9 at PR c.1730 in an interclub (the loss being on time in a drawn ending in the last round to a c.1700 after our team had won the season). There was a bonus points system in place in those days and mostly due to that performance I went up to 1411 on the next list. My subsequent play (mean PR for the next year's tournaments was about 1450) showed that the 8/9 was slightly more fluke than improvement.

If Anne is really playing 40 rated games per period or so then something in the region Bill's suggested sounds about right for those 7 wins. If her performance was either a total fluke or a lot of real improvement then at her activity level the system will not take long to fix it.

Bill Gletsos
31-01-2004, 05:53 PM
What a beautiful finesse! In fact, Anne loves her chess and has played so much her rating is very stable, with a RD of around 30 or !!. Alex, well his wifes been sick and he's struggling with late hours at work and you know, he hasn't played a rated game in 2yrs, he has a ?? rating. Now I wonder if this makes any difference?
No one on the master file has an RD of 30. Even those playing 50-70 games per rating period, so an RD of 70 is more likely. The lowest is possibly 50 something. The chance of 30 is 0%.

As for Alex, well this post totally contradicts what you said in your first question. :whistle:

It would take a lot longer than 2 years for a player with even a space after their rating to get a ? let alone a ??. They would be unlikely to have an RD over 150.

However since you seem to now be pulling numbers out your ass :rolleyes: , lets go to extremes and assume poor old Alex hasnt played for 25 years and has an RD of 350.

Also lets make Anne's RD 50. Lets also assume that she has a normal volatility.

Under those scenario's:

Under Glicko Anne's rating would be 1093, Alex with an RD of 150 is then 1210 and if his RD was 350 then 1365.
Under Glicko2 Anne's rating is 1099, Alex with an RD of 150 is then 1210 and if his RD was 350 then 1368.

Under the 3 tournament scenario then:
Anne's Glicko would be 1224, Alex for RD 150 would be 1323 and for RD 350 it is then 1417.
Anne's Glicko2 is 1455 and Alex with an RD of 150 is then 1325 and if his RD was 350 then 1418.

For 4 tournaments:
Anne's Glicko would be 1276, Alex for RD 150 would be 1350 and for RD 350 it is then 1420.
Anne's Glicko2 is 1549 and Alex with an RD of 150 is then 1349 and if his RD was 350 then 1425.

Of course all of this is consistent. Remember RD's are not linear. Alex with an RD of 150 or 350 is significantly more unreliable than Anne.
Now under Glicko for a small sample of just 7 games Alex will out rate Anne.
Even with more games Alex still outrates Anne.
Glicko2 however has no problem. If based on 7 games Alex still should outrate Anne but based on many games Anne will easily out rate Alex. :cool:

All of this is as it should be. :hand:

Bill Gletsos
31-01-2004, 06:00 PM
Maybe I get where this is heading, you want a scenario where some old rusty who has a few good results goes up faster than some junior who's doing spectacularly.
I was way ahead of you on this one Kevin. It seemed fairly obvious to me what his intention was since his original question. :whistle:

But if Anne has been playing so heavily her rating is a very stable 1000-odd (would RD really get down to 30 even for a very active junior?), then for her to suddenly, out of nowhere, put on that kind of a burst and get 7/7 vs an expected 2/7 - you'd have to be sceptical about how much of that was improvement vs fluke.
Good points.

If Anne is really playing 40 rated games per period or so then something in the region Bill's suggested sounds about right for those 7 wins. If her performance was either a total fluke or a lot of real improvement then at her activity level the system will not take long to fix it.
Yes, 7 results isnt really enough to determine whether its true performance improvemnt or a fluke.
The 3 tournaments/4 tournaments scenario demonstrates that the system will handle true improvement.

ChessGuru
31-01-2004, 07:12 PM
I'm interested what if Old Rusty has his 1000?? rating and then plays 7 players rated 500 and wins 7/7.

ChessGuru
31-01-2004, 07:26 PM
Actually let me make my question clearer:

Lets say IM Rusty has been in retirement from tournament chess for a long time. He is rated about 2350 (and going up all the time!) but hasn't played in 5-7 years. He wants to get back into chess but is scared that his RD will be 200+ and one or two losses will put him back to 2000. So his plan is cunning...build back his RD by playing against really weak players, never play anyone over (say) 1500 and win 100% of his games. Then he can play real chess again without his rating plummeting. OR perhaps winning 30/30 games (or 50/50) against nobody's with low ratings will increase his rating? Can he get to 2500 and secure a low RD?

How does this change with a player 2400, 2200, 2000, 1800, 1600 etc. ie. Is there something which protects top players ratings?

My point being that IM Rusty probably is really only 2100 strength now after a long break, but wants still the prestige of his big rating! He hopes with a bit of practice in 6 months or so he'll be back at 2400 strenght but would like his rating to stay there. So would this plan work?

Bill Gletsos
31-01-2004, 07:41 PM
I'm interested what if Old Rusty has his 1000?? rating and then plays 7 players rated 500 and wins 7/7.
He will go up around 110 points to 1110.
However let me quickly point out that playing twice as many games wont make him go up 220 points.
If he played 14 he only goes up 141 points to 1141, therefore those additional 7 are only worth 31 points.
If he played 21 games he would only go up 1154 so those last 7 would only gain him 13 points.

Bill Gletsos
31-01-2004, 07:57 PM
Actually let me make my question clearer:

Lets say IM Rusty has been in retirement from tournament chess for a long time. He is rated about 2350 (and going up all the time!) but hasn't played in 5-7 years. He wants to get back into chess but is scared that his RD will be 200+ and one or two losses will put him back to 2000. So his plan is cunning...build back his RD by playing against really weak players, never play anyone over (say) 1500 and win 100% of his games. Then he can play real chess again without his rating plummeting. OR perhaps winning 30/30 games (or 50/50) against nobody's with low ratings will increase his rating? Can he get to 2500 and secure a low RD?

How does this change with a player 2400, 2200, 2000, 1800, 1600 etc. ie. Is there something which protects top players ratings?

My point being that IM Rusty probably is really only 2100 strength now after a long break, but wants still the prestige of his big rating! He hopes with a bit of practice in 6 months or so he'll be back at 2400 strenght but would like his rating to stay there. So would this plan work?

It wont work.
The RD is not a measure of activity.
A player rated 2400(really only 2100) beating a host of 1500's wont really help his RD as beating them really proves very little about how accurate the 2400 rating is.

Secondly, how is IM 2400 going to get to play all these weak players. Matches are not allowed. An IM playing in a club tournament where he out rated everyone by 500-100 points would look suspicious.

In reality if IM 2400 is only 2100 strength then his rating should reflect that by his results. Anything else just means he is incorrectly over rated.

The effect for players rated 2200 and under is identical.
Those rated over 2300 have their increasees and decreases reduced by 25% making their rating movements more stable. Those between 2200 and 2300 have their rating change incresed or reduced on a sliding scale of 0%-25%.

Cat
01-02-2004, 07:49 AM
Thanks BG & KB

So what was the point of my question? It was really directed at Scott, who was trying to defend the purity of our rating system, suggest bringing the ratings more into line with FIDE is corrupting the system.

In our rating system (rightly or wrongly) there are biases and the weighting of those biases is to some extent, arbitary. A rating system is a functional thing, not an immutable law. As Bill & Kevin have previously said, the ratings were compacting (moving towards the mean), and in a (essentially) closed system (like The Gold Coast) if this were to continue the potential for dynamic interchange would reduce. Given 70% of our players are juniors, -striving for improvement this situation can lead to frustration and disappointment, even disaffection.

Bringing the ratings into line with FIDE is a functional improvment, at least for the Gold Coast because;
-it produces greater variation for dynamic interplay
-it counters the large influx of low rated juniors entering from the bottom of the system dragging the ratings lower
-there is less fear of rating loss which may encourage greater participation at the top end of town ( a coach rated say 1800 cannot afford to loose rating points, it may affect his livelihood. If he's 1900 and those players rated 1800 are that little less able, then he can afford to take a risk).

This is yet another functional improvement to the rating system. Bill has listened to criticism and responded. Well done Bill, keep it up!

Bill Gletsos
01-02-2004, 12:47 PM
As Bill & Kevin have previously said, the ratings were compacting (moving towards the mean),
I don't think I ever said that.
It was one of Kevin's suggestions. :)

and in a (essentially) closed system (like The Gold Coast) if this were to continue the potential for dynamic interchange would reduce.
It would be under Elo because the rating change based on a game between two players is zero sum.
However under Glicko and more so under Glicko2 that is not the case.

Bringing the ratings into line with FIDE is a functional improvment, at least for the Gold Coast because;
-it produces greater variation for dynamic interplay
-it counters the large influx of low rated juniors entering from the bottom of the system dragging the ratings lower
-there is less fear of rating loss which may encourage greater participation at the top end of town ( a coach rated say 1800 cannot afford to loose rating points, it may affect his livelihood. If he's 1900 and those players rated 1800 are that little less able, then he can afford to take a risk).
How is this true?
Every player gains the same number of points in the FIDE related uplift. A coach rating relative to his students ratings will still have the same differential after the uplift.

This is yet another functional improvement to the rating system. Bill has listened to criticism and responded. Well done Bill, keep it up!
Actually thats not true. Neither criticism nor praise by people have anything to do with it. We dont make changes on the whim of opinion. :hand:
It is quite simple if the maths/stats make sense then their is a change. :hmm:

Ever since the 150 point uplift back in April 2000 Graham and I have been monitoring FIDE/ACF drift.
However whilst the ratings system was going thru changes of Glicko to Glikco2 it would have been impractical to have made a further uplift.
Now that the Glicko2 system has been in place for 15-18 mths it seems reasonable to again look at uplifts.

Kevin Bonham
01-02-2004, 02:16 PM
As Bill & Kevin have previously said, the ratings were compacting (moving towards the mean), and in a (essentially) closed system (like The Gold Coast) if this were to continue the potential for dynamic interchange would reduce.

Probably not actually, the kind of compaction I was talking about (assuming it was that) was of the 2000+ players and particularly the 2200+ players compacting down and the rest compacting up with respect to them.

I have not looked at fresh figures for a while but I doubt it's happening much anymore, the system has had a fair while to shake up the overstretched ELO pool now, and the top-player stability tweak would also have negated it to some degree.

Also, while any such compaction would produce convergence, this should not be mistaken with convergence towards a mean. If there is a significant skill difference between two players, the rating difference may have reduced but would not be eliminated, so two players who might have been an overstretched 600 apart (say a 2300 who was really 2220 strength and a 1700 who was really 1780 strength) would not have both ended up at 2000, they would have ended up bobbing around their true playing strengths.

Cat
01-02-2004, 10:04 PM
How is this true?
Every player gains the same number of points in the FIDE related uplift. A coach rating relative to his students ratings will still have the same differential after the uplift.

Because most people aren't aware that the rating will have gone up, it comes down to perception. Mum's looking for coaches, the rating affects marketability. There are other factors of course.

Bill Gletsos
01-02-2004, 10:54 PM
Because most people aren't aware that the rating will have gone up
They will when it is mentioned in the ACF bulletin at the time the March ratings are published. :whistle:

Garvinator
01-02-2004, 10:57 PM
They will when it is mentioned in the ACF bulletin at the time the March ratings are published. :whistle: