PDA

View Full Version : ACF March 2004 Ratings



Pages : [1] 2

Bill Gletsos
27-02-2004, 11:05 PM
The ACF March 2004 rating lists have been sent to the webmaster and State Rating Officers.

Last year the ACF Council passed the following motion :"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."

In line with this all players December ratings were increased by 70 points prior to the calculation of the ratings for the March 2004 period.

Now someone is surely going to suggest that players lost most of the the 150 points that were added to virtually all players Dec 1999 ratings just prior to the April 2000 list. That however is not the case.

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 indicates 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.

Given the influx of many more Australians to the FIDE list over this 4 year period this has lead to an effective inflation trend in the FIDE ratings of Australian players especially those players below FM strength. It is from here that the 70 points is coming.

For the March 2004 rating period there were 80 Tournaments rated and 7818 games of which 5678 were in the ACF Classic rating system and 2140 were in the ACF Rapid rating system.


Top Players
2668!! 23 NSW Rogers, Ian [GM]
2509!! 11 VIC Johansen, Darryl K [GM]
2505! 0 NSW Wallace, John-Paul [IM]
2493!! 18 NSW Lane, Gary W [IM]
2458! 0 VIC Gluzman, Michael [IM]
2446!! 18 QLD Solomon, Stephen J [IM]
2439! 11 SA Tao, Trevor
2400!! 0 VIC Froehlich, Peter [IM]
2383! 0 VIC West, Guy [IM]
2381!! 11 SA Chapman, Mark [IM]
2380!! 11 VIC Smerdon, David C [IM]
2380!! 17 NSW Zhao, Zong-Yuan [IM]
2378!! 25 VIC Bjelobrk, Igor
2371!! 14 VIC Sandler, Leonid [IM]
2366! 0 VIC Teichmann, Erik [FM]
2359!! 25 VIC Depasquale, Chris J [FM]
2301! 0 NSW Tindall, Brett [FM]
2297!! 37 NSW Xie, George
2295!! 11 VIC Levi, Eddy L [FM]
2284!! 18 NSW Canfell, Gregory J [FM]
2280!! 26 VIC Baron, Michael [FM]
2276!! 7 VIC Rujevic, Mirko [FM]
2274!! 13 VIC Rashid, Abdulwahab [FM]
2272! 0 NSW Feldman, Vladimir [IM]
2271!! 11 NSW Reilly, Tim [FM]
2259! 0 QLD Stephson, David J
2252! 0 NSW Cook, Roger S
2252! 0 NSW Smirnov, Vladimir
2245!! 8 NSW Berezina - Feldman, Irina [IM]
2245! 7 NSW Flatow, A (Fred) [FM]
2244! 6 NSW Kabir, Ruhul
2243!! 26 WA Boyd, Tristan
2241!! 10 SA Zaric, Srboljub
2238!! 0 VIC Hamilton, Douglas G [FM]
2236!! 6 NSW Seberry, Ralph B
2234!! 30 NSW Tan, Justin
2204!! 0 QLD Sorokina, Anastasia [WIM]
2201! 0 NSW Bouchaaya, Tony
2196!! 0 NSW Goris, Robert
2186! 7 NSW Scott, Ronald
2184!! 17 NSW Rej, Tomek
2181! 0 NSW Hirschhorn, Jeremy K
2180!! 11 VIC Hacche, David J
2162! 0 NSW Ayvazyan, Armen
2162!! 11 VIC Chow, Samuel
2159!! 19 WA Barber, Haydn J [FM]
2152! 7 VIC Partsi, Dimitry
2151!! 17 NSW Yu, Ronald
2150!! 0 VIC Aghamalyan, Armen
2149! 11 VIC Booth, Stewart
2148!! 14 NSW Rose, Adrian P
2145! 0 WA McCamon, Boyd
2145! 0 NSW Zirdum, Ivan
2142! 0 NSW Dwyer, Danny [FM]
2141! 0 VIC Woodhams, Michael V
2139! 0 NSW Samar, Raul
2132! 0 SA Goldsmith, Alan D
2128!! 18 NSW Jones, Brian A [FM]
2126!! 0 NSW Ghenzer, Charles
2122! 0 WA Wilkins, Michael

Bill Gletsos
27-02-2004, 11:06 PM
Top Females
2245!! 8 NSW Berezina - Feldman, Irina [IM]
2204!! 0 QLD Sorokina, Anastasia [WIM]
2102!! 11 NSW Eriksson, Ingela
2098! 0 SA Nguyen, Giang
1992! 12 NSW Dekic, Biljana [WIM]
1950! 0 NSW Lip, Catherine [WFM]
1926! 0 WA Mills, Natalie
1923! 0 NSW Klimenko, Veronica [WFM]
1914! 6 NSW Moylan, Laura A [WIM]
1832!! 7 NSW Lane, Nancy L [WIM]
1774!! 11 VIC Szuveges, Narelle S [WFM]
1769!! 4 VIC Lee, Michelle
1704! 6 NSW Rogers, Cathy L
1704! 5 VIC Zivanovic, Andjelija
1700!! 15 NSW Huddleston, Heather
1663!! 17 ACT Oliver, Shannon
1655! 0 NSW Wettstein, Carla
1655!! 32 NSW Song, Angela
1652!! 14 WA Payne, Sophie
1640! 11 NSW Shields, Sylvia

Bill Gletsos
27-02-2004, 11:07 PM
Top Under 20
2380!! 19 VIC Smerdon, David C
2380!! 17 NSW Zhao, Zong-Yuan
2297!! 18 NSW Xie, George
2243!! 19 WA Boyd, Tristan
2234!! 18 NSW Tan, Justin
2184!! 17 NSW Rej, Tomek
2162!! 16 VIC Chow, Samuel
2151!! 15 NSW Yu, Ronald
2112!! 17 VIC Bourmistrov, Denis
2098! 18 SA Nguyen, Giang
2062!! 19 QLD Humphrey, Jonathan
2055!! 18 VIC Nemeth, Janos
2038!! 16 ACT Jovanovic, Peter
1982!! 15 NSW Lubarsky, Kostia
1950! 17 NSW Lip, Catherine
1941!! 16 NSW O'Chee, Kevin
1938!! 17 NSW Zvedeniouk, Ilia
1935!! 14 VIC Stojic, Dusan
1932!! 14 ACT Wei, Michael
1913!! 15 NSW Suttor, Vincent

Bill Gletsos
27-02-2004, 11:07 PM
Top Under 16
2151!! 15 NSW Yu, Ronald
1982!! 15 NSW Lubarsky, Kostia
1935!! 14 VIC Stojic, Dusan
1932!! 14 ACT Wei, Michael
1913!! 15 NSW Suttor, Vincent
1899!! 15 ACT Oliver, Gareth
1843!! 9 NSW Song, Raymond
1833!! 13 SA Obst, James
1821!! 15 VIC Jager, Jesse
1813!! 14 QLD Barnard, Casey T
1805!! 12 QLD Ly, Moulthun
1803!! 12 VIC Jia, Jing Qu
1790!! 15 VIC Lindberg, Douglas
1781!! 13 VIC Lugo, Ruperto
1773!! 12 ACT Ikeda, Junta
1769!! 13 VIC Lee, Michelle
1744!! 14 NSW Cronan, James
1723!! 15 WA Chapman, Mark J
1718!! 13 QLD Lazarus, Benjamin
1710!! 12 NSW Huang, Justin

Bill Gletsos
27-02-2004, 11:08 PM
Top Under 12
1843!! 9 NSW Song, Raymond
1655!! 11 NSW Song, Angela
1597!! 11 NSW Illingworth, Max
1559!! 10 VIC Lin, Zhigen Wilson
1503!! 11 QLD Anderson, Daniel C
1494!! 9 QLD Finke, Kelvin
1451!! 11 NSW Harris, Benjamin
1420! 9 VIC Morris, James
1416!! 11 NSW Wu, Edwin
1404! 11 QLD Xu, Mingda
1393!! 11 ACT Brown, Andrew
1387!! 11 WA Choong, Yita
1351!! 10 NSW Miranda, Adrian
1328!! 10 NSW Xu, William
1252!! 11 ACT Ung, Thomas
1226!! 11 VIC Ziffer, Rocheleh
1223!! 11 NSW Soltysik, Adelaide
1217!! 9 ACT Chow, Justin
1198! 11 VIC Carey, Jamie
1185!! 10 VIC Vijayakumar, Rengan

Bill Gletsos
27-02-2004, 11:09 PM
Top Females Under 20
2098! 18 SA Nguyen, Giang
1950! 17 NSW Lip, Catherine
1769!! 13 VIC Lee, Michelle
1704! 19 VIC Zivanovic, Andjelija
1700!! 14 NSW Huddleston, Heather
1663!! 17 ACT Oliver, Shannon
1655!! 11 NSW Song, Angela
1652!! 18 WA Payne, Sophie
1640! 19 NSW Shields, Sylvia
1516!! 14 NSW Reid, Vaness
1485! 19 QLD Mowles, Karina
1438!! 14 NSW Harris, Rebecca
1420! 16 VIC Richter, Sally-Anne
1375!! 14 QLD Jule, Alexandra
1334! 14 ACT Eldridge-Smith, Veronique
1333! 18 WA Cassidy, Kelly
1324!! 13 QLD Evans, Amy L
1319!! 13 VIC Hickman, Casey
1268! 17 VIC Lee, Melissa
1264! 17 NSW Chu, Winnie

Bill Gletsos
27-02-2004, 11:10 PM
Top Seniors
2245! 67 NSW Flatow, A (Fred)
2241!! 64 SA Zaric, Srboljub
2238!! 63 VIC Hamilton, Douglas G
2126!! 66 NSW Ghenzer, Charles
2050!! 76 WA Leonhardt, Wolfgang
2034!! 77 NSW Viner, Phillip J
2024!! 61 NSW Capilitan, Romeo
1983! 66 NSW Benson, Lachlan
1953! 66 SA Riches, L John
1921!! 62 NSW Gross, Mike C
1917! 68 NSW Hutchings, Frank P
1909!! 61 NSW Winter, George
1901! 63 NSW Markovic, Milosav
1898! 72 NSW Creech, Ken E
1861!! 65 NSW Tulevski, Vasil G
1855! 82 NSW Webb, Selwyn
1854! 76 ACT Korda, Augustin
1844! 71 NSW Lazaridis, Julius
1840!! 63 NSW Pilja, Djuro
1808! 64 NSW Macmillan, Malcolm

Bill Gletsos
27-02-2004, 11:11 PM
Top Improvers
1833 224 SA Obst, James
1227 217 NSW Adler, Nathan
1390 202 NSW Webster, David
1821 153 VIC Jager, Jesse
1217 131 ACT Chow, Justin
1242 123 VIC Sanderson, Paul
1309 114 NSW Chu, Louis
1716 111 VIC Emodi, Nick
732 111 ACT Flood, Christopher
971 110 VIC Mihalopoulos, Jason
1333 108 NSW Schuetz, Fred
1098 105 QLD Grigg-Tondeleir, Samuel
1054 105 QLD Fisher, Jayden
1583 102 QLD Quaresmini, Harvey A
1370 101 VIC Schlossberg, Jonathan
1137 101 NSW Lam, David
1653 97 SA Cooke, Tristrom
1498 96 WA Colic, Igor
1504 91 QLD Buciu, Avram
1856 90 VIC Fry, Peter
1766 90 VIC Hoholis, Harry
1755 90 NSW Field, Andrew
1805 89 QLD Ly, Moulthun
933 88 QLD Kinder, Jessica
1394 87 WA Humphries, Ryan
1795 85 NSW Bogdanoski, Petre
820 83 QLD Rynne, Tom
1701 83 ACT Neeman, Jeremy
1946 82 QLD Sonter, Matthew
1003 81 VIC Young, Darren

Bill Gletsos
27-02-2004, 11:11 PM
Busiest Players
1717 45 NSW Mendes da Costa, Alex
1597 43 NSW Illingworth, Max
1635 40 NSW Keuning, Anthony V
1559 40 VIC Lin, Zhigen Wilson
1821 39 VIC Jager, Jesse
1943 37 NSW Peters, Duncan
2297 37 NSW Xie, George
2009 37 VIC Dragicevic, Domagoj
1817 33 NSW Yu, Michael
2006 33 VIC Dizdarevic, Mehmedalija
1965 32 NSW Harp, Joel
1655 32 NSW Song, Angela
1843 31 NSW Song, Raymond
1833 31 VIC Frost, Peter
1738 31 WA Fedec, John
2063 30 NSW Lip, Michael
2106 30 NSW Charles, Gareth
2234 30 NSW Tan, Justin
1938 30 NSW Zvedeniouk, Ilia
1265 28 NSW Parker, Trent
1805 28 QLD Ly, Moulthun
1394 28 WA Humphries, Ryan
1431 27 VIC Reuben, Des
1654 26 NSW Escribano, Jose
1710 26 NSW Huang, Justin
1416 26 NSW Wu, Edwin
1328 26 NSW Xu, William
1432 26 NSW Tse, Jeffrey
1011 26 NSW Ng, Deborah
1375 26 NSW Boyce, Jamie
2280 26 VIC Baron, Michael
1319 26 VIC Hickman, Casey
1282 26 VIC Mustafa, Metin
2243 26 WA Boyd, Tristan
2012 25 NSW Wright, Neil S
1861 25 NSW Yap, Aaron
2359 25 VIC Depasquale, Chris J
1790 25 VIC Lindberg, Douglas
2378 25 VIC Bjelobrk, Igor
1424 25 VIC Kerksal, David

Rincewind
27-02-2004, 11:37 PM
I assume the top improvers delta score does not have the 70 point increase included. (considering the last the list is only just clearing 70 by 8 points)

Bill Gletsos
28-02-2004, 12:16 AM
You should think of it as starting out with your Dec 2003 rating plus 70 points. Then all the March ratings are calculated.

Cat
28-02-2004, 10:27 AM
Dear Bill

Having been sometimes somewhat critical of the ratings over the last 12 months, I'd like to congratulate you on the huge improvements you've initiated. Your defence of the Glicko system was admirable and you have indeed been vindicated in demonstrating that the system can be brought up to scratch :clap:

Congratulations to Ian Rogers too, whose outstanding victories have won him a truely amazing rating :clap:

Obviously I haven't seen the entire list for the Gold Coast, but the omens seem good. It would appear my own rating would have increased almost 200 points since November without playing a game - now that's what I call chess!

There is now a real prospect that our juniors can gain some dynamic movement in their ratings over the next 6-12 months and a chance that the higher rated players can feel confident of competing locally without being punished.

I guess the goal now will be to ensure these improvements are maintained beyond the next 6-12 months. There are some areas I feel may need further attention;

Regional variation You have said previously that you were sceptical that regional variation existed to any large extent. This may be so, but a simple check by cross-correlating regions using Sydney as 'ground zero' would be an insurance against what Matt termed as 'speciation'. This would ensure some stability for those players near the top end of the regional pool and allow regions like Box Hill and ACT (who have a large junior pool and may be in danger of regional 'compaction') some bouyancy. This would also assist those talented, enthusiastic isolated juniors, who perform well in major competitions in places like Brisbane and the Gold Coast, only to find their rating improvements are lost when they lose the odd game to an under-rated junior in their own town.

Special attention In regions such as Box Hill and the ACT where large numbers of rapidly improving juniors are entering the rating system there is a drag effect from the disproportionally large junior pool and small adult pool. As these juniors start to play frequently and they obtain very stable ratings there is a reverse migration effect from the adult rating level to the junior. As these regions are probably the most fertile spawning grounds for future champions, it is essential we maintain rating fluidity in the junior pool and stability in the adult pool. I would suggest these regions get special attention to ensure their ratings are reasonably accurately represented in real time. Its true that periodic corrections of 150 points every 3-4 years will ease the problem, but 3-4 years is a long time in the life of a 10 year old and a lot of disaffection can be generated in the interim.

Best wishes


David.

Bill Gletsos
28-02-2004, 01:41 PM
Dear Bill

Having been sometimes somewhat critical of the ratings over the last 12 months, I'd like to congratulate you on the huge improvements you've initiated. Your defence of the Glicko system was admirable and you have indeed been vindicated in demonstrating that the system can be brought up to scratch :clap:
Thanks.


Congratulations to Ian Rogers too, whose outstanding victories have won him a truely amazing rating :clap:
Ian may be embarressed by his rating but Graham Saint and I feel it is entirely justified. If no one had received the 70 point increase he would have still been nearly 160 points above his nearest rival. Back in April 2000 when we gave everyone 150 points and Ian zero we did this because evidence showed his ACF rating was inflated. This is definately not the case now. He deserves the 70 points. To not give them to him would distort the parity.


Obviously I haven't seen the entire list for the Gold Coast, but the omens seem good. It would appear my own rating would have increased almost 200 points since November without playing a game - now that's what I call chess!
I realise your being a little michevious here so I wont bite. ;)


There is now a real prospect that our juniors can gain some dynamic movement in their ratings over the next 6-12 months and a chance that the higher rated players can feel confident of competing locally without being punished.

I guess the goal now will be to ensure these improvements are maintained beyond the next 6-12 months. There are some areas I feel may need further attention;
We will just have to see what happens. I demand evidence rather than speculation.


Regional variation You have said previously that you were sceptical that regional variation existed to any large extent. This may be so, but a simple check by cross-correlating regions using Sydney as 'ground zero' would be an insurance against what Matt termed as 'speciation'. This would ensure some stability for those players near the top end of the regional pool and allow regions like Box Hill and ACT (who have a large junior pool and may be in danger of regional 'compaction') some bouyancy. This would also assist those talented, enthusiastic isolated juniors, who perform well in major competitions in places like Brisbane and the Gold Coast, only to find their rating improvements are lost when they lose the odd game to an under-rated junior in their own town.
We will have to wait and see what happens in the ACT now. We think the problem that the ACT had should not re-occur. As for Box Hill all the examples starter quoted were found to be unfounded when examined.


Special attention In regions such as Box Hill and the ACT where large numbers of rapidly improving juniors are entering the rating system there is a drag effect from the disproportionally large junior pool and small adult pool. As these juniors start to play frequently and they obtain very stable ratings there is a reverse migration effect from the adult rating level to the junior. As these regions are probably the most fertile spawning grounds for future champions, it is essential we maintain rating fluidity in the junior pool and stability in the adult pool. I would suggest these regions get special attention to ensure their ratings are reasonably accurately represented in real time.
The evidence does not necessarily support this. Large numbers of juniors maybe entering the pool but that does not equate to a large number of rapidly improving juniors entering the pool.

If juniors show a sudden change in underlying strength the Glicko2 will handle it.


Its true that periodic corrections of 150 points every 3-4 years will ease the problem, but 3-4 years is a long time in the life of a 10 year old and a lot of disaffection can be generated in the interim.
The 150 points issue has nothing to do with any of this. If every player receives the same increase it makes no difference to relative rating, hence it makes no difference to expected results.

Having said all of the above, I'll make the following observations.

In the FIDE pool regional variations based on country/zones is evident. Except for players at the top end of the FIDE list (I would estimate above 2300) I supect there is relatively little correlation between a 2000 FIDE rated aussie, a 2000 FIDE rated American or a 2000 FIDE rated Russian.

FIDE appear to do nothing to handle regional variation.

Like wise regional variation does not appear to be taken into consideration in any of the worlds ratings systems. It doesnt happen in the US or England.

Now that does not mean it shouldnt happen, it just means it currently doesnt happen. I suspect the main reason is lack of correlating data.

As for special treatment of juniors, I notice the USCF who has a massive number of as they call them "scholastic players" does absolutely nothign to effect their ratings. they all are subject to the standard ratings calculations.
Again this appears to be due to the fact that all juniors improve or dont improve at totally different rates and stages .

Rincewind
28-02-2004, 02:02 PM
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.

I rating came in within 2 points of the prediction (1822) so the amazing Glicko Estimator comes through again. :)

Unfortnuately last year's team now have 3 team members over 1800, so it won't be so easy this year.

Bill Gletsos
28-02-2004, 02:10 PM
Everyone please note that an ACT tournament was missing from the ratings. The ratings have been reprocessed and the changes reflected in the lists at the start of this thread.

Kevin Bonham
28-02-2004, 03:31 PM
Like wise regional variation does not appear to be taken into consideration in any of the worlds ratings systems. It doesnt happen in the US or England.

Now that does not mean it shouldnt happen, it just means it currently doesnt happen. I suspect the main reason is lack of correlating data.

It is also easy to claim but extremely difficult to prove.

To test it properly you 'd need to look at results of players from zone A against players from zone B both within zone A and zone B, to eliminate the possibility that players perform better/worse when travelling compared to when at home. (Or you could look at meetings on effectively "neutral" ground, if you could define that, but there would be few of these). Even then you would have to look at what kinds of players travelled from region to region, eg if those travelling from zone A to zone B were improving juniors and those travelling from B to A were declining seniors, then the picture would be distorted. Statistically, it's not simple, and you'd need enough evidence to show a statistically significant difference, which would be quite a lot.

Cat
28-02-2004, 03:33 PM
In the FIDE pool regional variations based on country/zones is evident. Except for players at the top end of the FIDE list (I would estimate above 2300) I supect there is relatively little correlation between a 2000 FIDE rated aussie, a 2000 FIDE rated American or a 2000 FIDE rated Russian.

FIDE appear to do nothing to handle regional variation.

Like wise regional variation does not appear to be taken into consideration in any of the worlds ratings systems. It doesnt happen in the US or England.

Now that does not mean it shouldnt happen, it just means it currently doesnt happen. I suspect the main reason is lack of correlating data.

As for special treatment of juniors, I notice the USCF who has a massive number of as they call them "scholastic players" does absolutely nothign to effect their ratings. they all are subject to the standard ratings calculations.
Again this appears to be due to the fact that all juniors improve or dont improve at totally different rates and stages .

I would just point out regionalism may be more acute in Australia. Certainly in the UK, a trip say from London to Leeds might take as little as 5hrs, so there's a much greater mixing of pools. Similarly in the US, there are greater numbers and more established clubs which can dilute the effect.

Bill Gletsos
28-02-2004, 04:37 PM
I would just point out regionalism may be more acute in Australia. Certainly in the UK, a trip say from London to Leeds might take as little as 5hrs, so there's a much greater mixing of pools. Similarly in the US, there are greater numbers and more established clubs which can dilute the effect.
I suspect thats all just supposition and not based on any actual data. :whistle:

PHAT
28-02-2004, 05:32 PM
It is also easy to claim but extremely difficult to prove.

To test it properly you 'd need to look at results of players from zone A against players from zone B both within zone A and zone B, to eliminate the possibility that players perform better/worse when travelling compared to when at home. (Or you could look at meetings on effectively "neutral" ground, if you could define that, but there would be few of these). Even then you would have to look at what kinds of players travelled from region to region, eg if those travelling from zone A to zone B were improving juniors and those travelling from B to A were declining seniors, then the picture would be distorted. Statistically, it's not simple, and you'd need enough evidence to show a statistically significant difference, which would be quite a lot.

Cannot agree that it is difficult. It may be time consuming to gather the data, but it is there and it can be done. eg. 2 zones Germans and Dutch, both in the zone Europe. Each player has DOB, Nationality, and FIDE rating , plus a history. correlations between ages, improvers, nationalities genders, etc cetera could be easly examined. The stats on such a data set would be large enough to run an ANOVA and get pretty strong evidence - if it exists - that pooling/speciation events are occur. Then it is up to others to show that pooling/speciation is not occuring in their corner of the world.

chesslover
28-02-2004, 06:23 PM
I suspect thats all just supposition and not based on any actual data. :whistle:

well it makes sense to me from a qualitiative point of view...

Cat
28-02-2004, 06:39 PM
I suspect thats all just supposition and not based on any actual data. :whistle:
No its based on elemental geography and a personal knowledge from the UK.

jenni
28-02-2004, 07:19 PM
I am far too cautious to say our ACT problem is totally fixed - I would want a few more rating periods to go by.

However I feel fairly happy at the moment with how our juniors' ratings are going. Certainly our rapidly improving juniors do seem to go up more quickly and reach their correct ratings. There will always be a bit of a problem in the ACT, because we have such a small pool of adults. However more of our juniors are starting to travel and play outside of Canberra and hopefully over time if we can retain a % of them, they will add to the adult pool.

Bill Gletsos
28-02-2004, 07:34 PM
well it makes sense to me
Well then I guess that proves my point then. :whistle:

Bill Gletsos
28-02-2004, 07:38 PM
No its based on elemental geography and a personal knowledge from the UK.
So your assumption about the US is based on no actual data and the UK based on your own opinion.

Unfortubnately David, based on some of your other opinions regarding ratings, I therefore cant put much faith in your suppositions.

chesslover
28-02-2004, 08:31 PM
So your assumption about the US is based on no actual data and the UK based on your own opinion.

Unfortubnately David, based on some of your other opinions regarding ratings, I therefore cant put much faith in your suppositions.

Surely, what david states makes sense?

UK is a far smaller place than us, and the distance between places in England is much less than the distances between Australian states.

Do you want to look at the geography of England in relation to Australia, before you post on this topic about regional variation, Supreme leader?

Kevin Bonham
28-02-2004, 08:49 PM
Cannot agree that it is difficult. It may be time consuming to gather the data, but it is there and it can be done. eg. 2 zones Germans and Dutch, both in the zone Europe. Each player has DOB, Nationality, and FIDE rating , plus a history. correlations between ages, improvers, nationalities genders, etc cetera could be easly examined. The stats on such a data set would be large enough to run an ANOVA and get pretty strong evidence - if it exists - that pooling/speciation events are occur. Then it is up to others to show that pooling/speciation is not occuring in their corner of the world.

I wouldn't consider one pool-drifting example enough to put the burden of proof on the no-drift case. Multiple examples demonstrating that pool-drifting was more common than not, and you've got my attention.

Even so, just having that result would not necessarily tell you which way drift would occur in a given case, unless it predicted drift as a function of demographic or other data that were adequately available in that given case.

Bill Gletsos
28-02-2004, 09:17 PM
Surely, what david states makes sense?
Possibly only if you are a goose. ;)


UK is a far smaller place than us, and the distance between places in England is much less than the distances between Australian states.
No s..t sherlock. I never would have guessed that. :whistle:


Do you want to look at the geography of England in relation to Australia, before you post on this topic about regional variation, Supreme leader?
No. :hand:

PHAT
28-02-2004, 09:31 PM
I wouldn't consider one pool-drifting example enough to put the burden of proof on the no-drift case. Multiple examples demonstrating that pool-drifting was more common than not, and you've got my attention.


When a study is very large and encompasses a representive population, its importance is greater - true? As such, it can be said that it's interpretation is the current theory. Therefore we assume that the theory applies else where - perhaps everywhere - until shown otherwise.

So, if it was shown that players from country A with a larger proportion of improving juniors preformed better against players from B with a smaller proportion, we must assume it to be a phenomenum in other overlapping pools of players - until contradicting evidence is found.

Kevin Bonham
28-02-2004, 09:58 PM
When a study is very large and encompasses a representive population, its importance is greater - true?

I am not sure there is such a thing as a representative population in this case. Even the choice of rating system will affect the behaviour in question.


As such, it can be said that it's interpretation is the current theory.

I do not know how you could ever convincingly exclude alternative interpretations of the result based on a single study, which effectively includes only two control samples (zone A vs zone A, zone B vs zone B) and four experiment samples. I would like to see a lot more samples. In one of my fields (forest ecology) one of the commonest causes of flawed studies is not having enough control sites to exclude alternative hypotheses. Someone in my department had to redo about a year of their thesis due to errors of this type.


So, if it was shown that players from country A with a larger proportion of improving juniors preformed better against players from B with a smaller proportion, we must assume it to be a phenomenum in other overlapping pools of players - until contradicting evidence is found.

I would not assume this as a stand-alone finding.

I would accept it as a provisional result if there was other evidence to support the proportion of improving juniors being a likely cause. For instance if the "improving" juniors tended to perform above their own rating in a subsequent ratings period when paired against their own zone as well as the opposing one. (I think it has to be subsequent to avoid circularity).

Cat
28-02-2004, 10:56 PM
So your assumption about the US is based on no actual data and the UK based on your own opinion.

Unfortubnately David, based on some of your other opinions regarding ratings, I therefore cant put much faith in your suppositions.

Bill, you've been doing ok on this thread so far, don't mess it up. I can produce data to demonstrate the bleeding obvious, but it would be a waste of my time and yours. Maybe we'll just have to put up with plain, simple common sense?

KB, any data on regional variation is better than none, which seems to be the situation at present. The question is do we really want to be going through this whole exercise again in another 3-4 years? Isn't it better to put in place some management and surveillance plan to ensure the ratings remain reasonably representative. Surely there are lessons to be learnt from this experience?

Kevin Bonham
28-02-2004, 11:09 PM
KB, any data on regional variation is better than none, which seems to be the situation at present.

I have no objection to people gathering data on this, indeed I encourage it. I am just pointing out that proving something is going on to the level where a change to the system is justified isn't a piece of cake.


Surely there are lessons to be learnt from this experience?

Which experience?

Bill Gletsos
28-02-2004, 11:11 PM
Bill, you've been doing ok on this thread so far, don't mess it up. I can produce data to demonstrate the bleeding obvious, but it would be a waste of my time and yours. Maybe we'll just have to put up with plain, simple common sense?
If anyone will mess it up its you, when you make generalisations.
As for common sense isnt all its cracked up to be. Often it can lead you to false conclusions and lead to total error.
I doubt players on the US east coast often play with those on the west coast. Or even that chess players in San Franciso play those in Los Angeles. As for the UK how often do those in the North play those in the south.


KB, any data on regional variation is better than none, which seems to be the situation at present. The question is do we really want to be going through this whole exercise again in another 3-4 years? Isn't it better to put in place some management and surveillance plan to ensure the ratings remain reasonably representative. Surely there are lessons to be learnt from this experience?
You seem to have this idea that the 70 points is related to something other than FIDE rating inflation. It doesnt. Therefore there is no 3-4 year period of repetition.

Bill Gletsos
28-02-2004, 11:14 PM
Which experience?
I suspect he is referring to the ACT adjustment. If he is then he is way off track.
If its the Gold Coast then he is also off track.
If its the 70 point uplift then hes also off track.

Guess we can all be thankful he's a doctor and not a train driver. :owned:

If its regional variability then there is no experience.

Alan Shore
29-02-2004, 12:00 AM
Interesting how the ratings were up and then taken away.. I got the sneak peek.. it's like have seen a film before the official release date, hehe. Now I can see precicely what fiddling goes on behind the scenes when they're released 'officially' :)

Incidently, thanks for your work on the ratings Bill.. the Glicko system is fantastic (albeit unforgiving) and really gives accurate ratings fast - good stuff.

Pillsbury
29-02-2004, 12:39 AM
ACF March 04 Ratings Master List: <deleted>

Thanks K! :owned:

Cat
29-02-2004, 08:27 AM
Which experience?


Obviously not! Rgihtly or wrongly, individuals use ratings not only as a reference for their performance relative to others but also as a measure of their own performance. They expect (possibly unreasonably) that a rating of 1500 is the same in 1999, 2002, 2004, 2006 in Sydney, Brisbane or New York even. If there is regional or national 'compaction' then they may become disaffected by their apparent regression and this can lead to significant bleeding from the clubs.

Now & Bill may say that this does not happen, that it's all to do with FIDE inflation and that there is do need for further action. What I see is adults on the Gold Coast loosing an endless stream of rating points to the larger junior pool, juniors static despite obvious improvement and periodic corrections.

I know you have confidence that Glicko can deal with this. What I'm saying is keep the situation monitored - some stability is required.

ursogr8
29-02-2004, 08:53 AM
Obviously not! Rightly or wrongly, individuals use ratings not only as a reference for their performance relative to others but also as a measure of their own performance. They expect (possibly unreasonably) that a rating of 1500 is the same in 1999, 2002, 2004, 2006 in Sydney, Brisbane or New York even. If there is regional or national 'compaction' then they may become disaffected by their apparent regression and this can lead to significant bleeding from the clubs.


Dave
As you point out there are some individuals who expect the rating system to give them the same number over a decade if their chess ability doesn't change over a decade.
But then there are also individuals who will pay big money to get a FIDE rating because this is a larger number than their current ACF rating. The big money goes out of Australia to FIDE.
So, there you have it. Two conflicting behaviour sets.

I encourage Bill along the path of keeping ACF ratings in parity with FIDE ratings.

starter

Bill Gletsos
29-02-2004, 09:48 AM
ACF March 04 Ratings Master List: <deleted>

Thanks K! :owned:
You goose, that list is unofficial.
Thats why you cannot navigate to it from the ACF ratings page. :wall:

Also its not the master list but the active players list. :rolleyes:

In future if you cant navigate to it then dont post links to it. :mad:

Garvinator
29-02-2004, 09:59 AM
You goose, that list is unofficial.
Thats why you cannot navigate to it from the ACF ratings page. :wall:

Also its not the master list but the active players list. :rolleyes:

In future if you cant navigate to it then dont post links to it. :mad:
woohoo another person has reached the elite status of goosedom :owned: :owned: :whistle: :whistle:

Rincewind
29-02-2004, 10:03 AM
You goose, that list is unofficial.
Thats why you cannot navigate to it from the ACF ratings page. :wall:

Also its not the master list but the active players list. :rolleyes:

In future if you cant navigate to it then dont post links to it. :mad:

It's worse than that. Until I moderated it, there was a posting of the entire (unofficial) active list IN THIS THREAD!

Garvinator
29-02-2004, 10:06 AM
It's worse than that. Until I moderated it, there was a posting of the entire (unofficial) active list IN THIS THREAD!
no disrespect to bill, but the list can still be reached from bills quote of pillsbury's post. So that might need to be moderated too.

Rincewind
29-02-2004, 10:13 AM
no disrespect to bill, but the list can still be reached from bills quote of pillsbury's post. So that might need to be moderated too.

A case of closing the game after the goose has bolted I think. But as I was passing the gate, I might as well close it before any more geese bolt through it. Even though a brief glance seems to indicate there are none left in the paddock. :D

Rincewind
29-02-2004, 11:08 AM
If I didn't navigate to it, how did I find it, oh supreme goose?

I didn't think I'd need to explain this to you, Jeo. You could have guessed it the same way I did, from the name standard that has been used in the past. :)

I do note that the ACF website has now published the March list. So I guess it is "official" now.

Take it easy, dude. ;)

skip to my lou
29-02-2004, 11:11 AM
I didn't think I'd need to explain this to you, Jeo. You could have guessed it the same way I did, from the name standard that has been used in the past. :)

I do note that the ACF website has now published the March list. So I guess it is "official" now.

Take it easy, dude. ;)

I did not guess it from the name it has been in the past.

I got to it by "clicking" on links through auschess.

Bill Gletsos
29-02-2004, 12:35 PM
Obviously not! Rgihtly or wrongly, individuals use ratings not only as a reference for their performance relative to others but also as a measure of their own performance. They expect (possibly unreasonably) that a rating of 1500 is the same in 1999, 2002, 2004, 2006 in Sydney, Brisbane or New York even.
Then they are being silly. After all no one with any brains really believes that a player FIDE rated 2600 now is the same as one back in the mid 70's. Ratings reflect the relativities of players against others in the pool at the current time not how they were 2, 5, 10 or 20 years ago.


If there is regional or national 'compaction' then they may become disaffected by their apparent regression and this can lead to significant bleeding from the clubs.
If players strengths relative to others based on their results means their ratings are closer together then that is as it should be. That does not make "compaction" a problem.


Now & Bill may say that this does not happen, that it's all to do with FIDE inflation and that there is do need for further action.
Now you are really being a goose. The 70 point uplift has absolutely nothing to do with regionalisation, the previous ACT issue, compaction, God, the devil or even aliens from outer space.
It was simply to bring all ratings closer to FIDE ratings.


What I see is adults on the Gold Coast loosing an endless stream of rating points to the larger junior pool, juniors static despite obvious improvement and periodic corrections.
Did it ever actually occur to you that the adults might be getting worse.
And you just can't seem to get it through you head that not all juniors improve and even those that do improve dont do so at any sort of systematic rate.


I know you have confidence that Glicko can deal with this. What I'm saying is keep the situation monitored - some stability is required.
No I dont think I'll bother monitoring anything.

I can just use your method and use common sense. :rolleyes:

In which case all players will just have to pray.

Cat
29-02-2004, 01:06 PM
Dave
As you point out there are some individuals who expect the rating system to give them the same number over a decade if their chess ability doesn't change over a decade.
But then there are also individuals who will pay big money to get a FIDE rating because this is a larger number than their current ACF rating. The big money goes out of Australia to FIDE.
So, there you have it. Two conflicting behaviour sets.

I encourage Bill along the path of keeping ACF ratings in parity with FIDE ratings.

starter
Yes I agree completely, Starter. I'm suggesting we ensure a mechansim to achieve this real time in the future (as close as possible), rather than waiting for periodic corrections every few years.

Cat
29-02-2004, 01:18 PM
Bill Gletsos]Then they are stupid.

Thats as may be, but they are your constituency.




Now you are really being a goose. The 70 point uplift has absolutely nothing to do with regionalisation, the previous ACT issue, compaction, God, the devil or even aliens from outer space.
It was simply to bring all ratings closer to FIDE ratings.

Good work, but my rating has risen 200pts hasn't it? Why is that Bill? Look what you've done is great, a credit to you. It may never happen again. But surely we can attempt to put in place some simple mechanisms to avoid major corrections and maintain the ratings broadly in-line with FIDE.


Did it ever actually occur to you that the adults might be getting worse.
And you just can't seem to get it through you head that not all juniors improve and even those that do improve dont do so at any sort of systematic rate.

This is rehashing old ground, surely you don't what that all brought back up again?


No I dont think I'll bother monitoring anything.

I can just use your method and use common sense.


It would be a start!

Bill Gletsos
29-02-2004, 01:48 PM
Thats as may be, but they are your constituency.
Their inability to understand the obvious is not my concern.


Good work, but my rating has risen 200pts hasn't it? Why is that Bill?I thought I made it quite clear.
Your original 124 points was part of the GC correction. However I thought I made it clear to you that if corrections had been applied recalulations many players there would have lost points, to the extent that the net rating change for all GC players would have been zero. We took the decision not to adversekly effect any players in the recalculations.This therefore added points to the GC pool.

The 70 points have nothing to do with any local issues. In fact it could be just as easily argued that no points should be added on the grounds that problems with the FIDE ratings should not be a reason to adjust ACF ratings.
The 70 points you got is no different to thoase that starter, barry Cox or even I got. The 70 points is totally meaningless in the scheme of things when it comes to ratings calculations.


Look what you've done is great, a credit to you. It may never happen again. But surely we can attempt to put in place some simple mechanisms to avoid major corrections and maintain the ratings broadly in-line with FIDE. And its here you show no understanding of the FIDE issue. In fact there can be no automatic determination of what may happen to FIDe ratings because you have no knowledge of the FIDE ratings until they are published.
Irrespective of this you obviously only seem to wish to read/undertsand what you feel furthers your own agenda. In the very first post in this thread it says "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."
It is obvious from this that any FIDE alignment will in future occur every year in the march list.


This is rehashing old ground, surely you don't what that all brought back up again?
Only if you want to waste your time and mine.


It would be a start!
Yes but a pretty useless start.

paulb
29-02-2004, 03:28 PM
Apologies to everyone but I've used the moderator's sledgehammer to crack a walnut and gone berzerk in a blood-curdling frenzy of deletion. They don't call me "that rude Dutch count" and "Paul Pot" for nothing :)

It's an instance of Dutch diplomacy, as employed to effect in the International Court of Justice, and the signing of the Maastricht european union Treaty and The Hague Convention (predecessor to the Geneva Convention). The Dutch tactic is to invite warring parties to a meeting, and then insult both sides so much that they forget about their prior dispute and find common cause against their unspeakably rude Dutch hosts. Works a treat, every time.

Seriously, though, the choice was to spend the rest of this glorious summer afternoon on the net, instead of the beach, sorting through dozens of posts, or simply give a short, sharp slash with my moderator's machete. To those who've had worthy posts removed, my apologies. But think of my suntan!

In any case, the dispute is now historic, since the corrected ratings files are on the ACF website.

cheers - PaulB

Garvinator
29-02-2004, 03:31 PM
Apologies to everyone but I've used the moderator's sledgehammer to crack a walnut and gone berzerk in a blood-curdling frenzy of deletion. They don't call me "that rude Dutch count" and "Paul Pot" for nothing :)

It's an instance of Dutch diplomacy, as employed to effect in the International Court of Justice, and the signing of the Maastricht european union Treaty and The Hague Convention (predecessor to the Geneva Convention). The Dutch tactic is to invite warring parties to a meeting, and then insult both sides so much that they forget about their prior dispute and find common cause against their unspeakably rude Dutch hosts. Works a treat, every time.

Seriously, though, the choice was to spend the rest of this glorious summer afternoon on the net, instead of the beach, sorting through dozens of posts, or simply give a short, sharp slash with my moderator's machete. To those who've had worthy posts removed, my apologies. But think of my suntan!

In any case, the dispute is now historic, since the corrected ratings files are on the ACF website.

cheers - PaulB
yeah i noticed the thread went from 6 pages to 4 :D . I think the machete could have been wielded a little closer to the bottom of the stem :lol: :owned:

Alan Shore
29-02-2004, 07:13 PM
If it's not too much trouble, I'd like to be enlightened as to what changed were made to the final list from the provisional one - I didn't notice any changes to my ratings or to those of people I know (small sample I know, but still), so what was tweaked exactly? Thanks.

Bill Gletsos
29-02-2004, 07:29 PM
If it's not too much trouble, I'd like to be enlightened as to what changed were made to the final list from the provisional one - I didn't notice any changes to my ratings or to those of people I know (small sample I know, but still), so what was tweaked exactly? Thanks.
Two changes.
The Tuggeranong Vikings weekender from last December was missed by the ACT. Also a VIC tournament (Elwood Handicap) was submitted as a Rapid when it should have been a Normal.

Garvinator
29-02-2004, 08:30 PM
ok i have a question.

Now that all players(as i understand it) have had their ratings increased by 70 points to keep inline with fide ratings, will the australian championship entry rating of 2150 be increased to 2220?

Bill Gletsos
29-02-2004, 09:38 PM
ok i have a question.

Now that all players(as i understand it) have had their ratings increased by 70 points to keep inline with fide ratings, will the australian championship entry rating of 2150 be increased to 2220?
I would think highly unlikey, but it could go to 2200 or maybe back to 2250.

Garvinator
29-02-2004, 09:48 PM
I would think highly unlikey, but it could go to 2200 or maybe back to 2250.
so i take it then that the rating mark was not raised when the 150 points was added to ppl's ratings?

Kevin Bonham
29-02-2004, 10:16 PM
Obviously not! Rgihtly or wrongly, individuals use ratings not only as a reference for their performance relative to others but also as a measure of their own performance. They expect (possibly unreasonably) that a rating of 1500 is the same in 1999, 2002, 2004, 2006 in Sydney, Brisbane or New York even. If there is regional or national 'compaction' then they may become disaffected by their apparent regression and this can lead to significant bleeding from the clubs.

You were talking about learning from an experience. I simply asked which experience you were talking about because, as Bill points out, there are a range of different experiences or alleged experiences

As for compaction, points loss due to 'compaction' (assuming that is even still happening, which I doubt) would only affect those with high ratings. For those below the 1900s it was either a non-issue or a benefit.

If you don't think the 70 points can be explained purely by FIDE inflation then the onus is on you to explain why the ACF players with FIDE ratings have on average stayed remarkably static on the ACF scale. I'm not saying that the ACF stasis for these players proves the 70 points is FIDE inflation, but I'm assuming it until proven otherwise. For instance if the mean age of those players was far younger than the mean age of the pool as a whole, that might be of interest.

ursogr8
01-03-2004, 09:45 AM
Two changes.


Bill
I notice a nice note from bobby1972 (ratings people are tops ) that reminds me to say that on behalf of the Box Hill Chess Club we thank you and Graham for the huge volunteer effort you put in to produce these ratings updates. It is very much appreciated how you produce these figures which we use to produce fair and reasonable pairings.
regards, starter

Bill Gletsos
01-03-2004, 10:49 AM
Bill
I notice a nice note from bobby1972 (ratings people are tops ) that reminds me to say that on behalf of the Box Hill Chess Club we thank you and Graham for the huge volunteer effort you put in to produce these ratings updates. It is very much appreciated how you produce these figures which we use to produce fair and reasonable pairings.
regards, starter
Thanks starter.
Graham and I are aware that the majority of players appreciate our efforts.
In fact I think all the identifiable posters on the BB appreciate our efforts.

Oepty
01-03-2004, 03:37 PM
Bill. Why wasn't the Australian Schools Teams Championship rated? Also would the time control used have placed it on the rapid list?

Scott

Bill Gletsos
01-03-2004, 04:37 PM
Bill. Why wasn't the Australian Schools Teams Championship rated? Also would the time control used have placed it on the rapid list?
Simple. It appears the organiser did not submit it for rating.

chesslover
01-03-2004, 05:06 PM
Thanks starter.
Graham and I are aware that the majority of players appreciate our efforts.
In fact I think all the identifiable posters on the BB appreciate our efforts.

That is very true. Your efforts as well as Graham's and your hard work are very much appreciated and loved by all of us. Now if only you would change your stance on self nomination and appeals, you would be just perfect :p :p

chesslover
01-03-2004, 05:09 PM
in the top improvers list, you have not added 70 points to everyone in that list. Given that from the last rating list everyone got 70 points increase, that 70 points too should be added to the list

thus it should show James Obst increase by 294 points (not 224 as now), Nathan Adler by 287 (not 217 as now) etc etc

Bill Gletsos
01-03-2004, 06:42 PM
in the top improvers list, you have not added 70 points to everyone in that list. Given that from the last rating list everyone got 70 points increase, that 70 points too should be added to the list

thus it should show James Obst increase by 294 points (not 224 as now), Nathan Adler by 287 (not 217 as now) etc etc
No that is incorrect.
The James Obst "strength" only improved 224 points, not 294.
His effective pre March rating was 1609.

Kevin Bonham
01-03-2004, 08:05 PM
Ian may be embarressed by his rating but Graham Saint and I feel it is entirely justified.

Bill, could you post an Ian Rogers rating history summary (like the ones you've posted for several other players) here sometime? Just interested in looking at how Glicko handles players who are regularly scoring very high percentages within their ratings pool/fishbowl. On the same note, could you post one for some patzer called Bonham? Not urgent, just curious.

Bill Gletsos
01-03-2004, 09:46 PM
Bill, could you post an Ian Rogers rating history summary (like the ones you've posted for several other players) here sometime? Just interested in looking at how Glicko handles players who are regularly scoring very high percentages within their ratings pool/fishbowl. On the same note, could you post one for some patzer called Bonham? Not urgent, just curious.
No problem.
Here are the figures.


Results For Player Ian Rogers
Period Rating Perf Score Games
Aug 2000 2587 0 22.5 30
Dec 2000 2598 2686 22.0 24
Apr 2001 2596 2582 15.0 18
Aug 2001 2582 2432 5.5 7
Dec 2001 2583 2589 7.0 8
Apr 2002 2563 2478 10.5 13
Aug 2002 2561 2558 24.0 28
Dec 2002 2544 2479 20.0 23
Mar 2003 2552 2193 6.0 6
Jun 2003 2581 2815 14.5 15
Sep 2003 2591 2625 33.5 36
Nov 2003 2593 0 0.0 0
Dec 2003 2610 2803 11.5 12
Mar 2004 2668 2625 20.0 23

Results For Player Kevin Bonham
Period Rating Perf Score Games
Aug 2000 1893 0 10.0 14
Dec 2000 1898 1915 21.5 27
Apr 2001 1903 1915 13.5 21
Aug 2001 1892 1844 9.0 12
Dec 2001 1879 1854 21.0 26
Apr 2002 1849 1796 11.5 23
Aug 2002 1892 2091 6.0 6
Dec 2002 1931 2043 23.0 25
Mar 2003 1950 1792 9.0 9
Jun 2003 1956 2000 5.0 6
Sep 2003 1908 1638 4.0 6
Dec 2003 1951 2056 24.0 27
Mar 2004 2039 2187 10.5 11

Kevin Bonham
02-03-2004, 01:25 AM
Thanks Bill.

I agree that Rogers' rating is justified having seen those because if one looks at his form over the last year, it turns out that he is actually performing somewhere in the 2730s (in post-70 pts terms) from 63 games from a start rating equivalent to 2622. Awesome effort really. So in giving him 46 points for those results, the system is measuring 63 games over 1 year at about 40% of the evidence of his current playing strength. That seems absolutely fair to me.

I was interested in seeing whether my increase (even without the 70 pts) over the last two years reflected actual improvement or just some strange artifact of the system. It seems I've actually got a bit better relative to my opponents. I work out rough PRs for my tournaments as I'm going along but they don't tell the full story because of the errors caused by batching - in the last rating period the estimate was 200 points out for this reason.

Bill Gletsos
02-03-2004, 08:58 AM
Thanks Bill.

I agree that Rogers' rating is justified having seen those because if one looks at his form over the last year, it turns out that he is actually performing somewhere in the 2730s (in post-70 pts terms) from 63 games from a start rating equivalent to 2622. Awesome effort really. So in giving him 46 points for those results, the system is measuring 63 games over 1 year at about 40% of the evidence of his current playing strength. That seems absolutely fair to me.
Also remember that Ian like all players over 2300 is only getting 75% of their true Glicko2 increases and decreases.

Kerry Stead
02-03-2004, 10:07 AM
Also remember that Ian like all players over 2300 is only getting 75% of their true Glicko2 increases and decreases.
Why is that Bill? I thought that sort of thing only applied to decreases, as there was not the opportunity for higher rated players to play people significantly above their rating to recover points lost on 'off days', as there would be for someone around say 1600.

Bill Gletsos
02-03-2004, 10:10 AM
Why is that Bill? I thought that sort of thing only applied to decreases, as there was not the opportunity for higher rated players to play people significantly above their rating to recover points lost on 'off days', as there would be for someone around say 1600.
Because prior to implementing it testing showed it wasnt statistically valid to reduce decreases but not increases.

Oepty
02-03-2004, 11:05 AM
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 indicates 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.



Bill
I would like to see some similar figures to the figures you have shown above. You appear to have included players who were active on the ACF rating list, but not active on the FIDE list. Is this correct?

If it is I would like to look at the comparison of the figures of players that were active on both FIDE and ACF rating lists at both times if that is possible.
I also would like to look at what has happened to players with higher ratings, players with FIDE ratings over 2300 and over 2200 to see if has been much change there.

It is a pity that the ATSC was not submitted, although I think the time limit, 55 plus 10sec from start is actually rapid. Is that correct?
Scott

Bill Gletsos
02-03-2004, 11:36 AM
Bill
I would like to see some similar figures to the figures you have shown above. You appear to have included players who were active on the ACF rating list, but not active on the FIDE list. Is this correct?
Yes and no.

There are two seperate issues.
1) When looking at how the ACF ratings changed for players with FIDE ratings we looked at all players who had played between April 2000 and Dec 2003. players. This was 212 players.

2) When it came to determining the uplift we only looked at players who we considered active on the Dec 2003 ACF list and the Jan 2004 FIDE list. We also ignored all players whose ACF ratings were below 2050. The true uplift was 68 points but we decided to give everyone 70. :owned:

Therefore someone like Alex Wohl was included in 1) but not in 2) because he is not active on the ACF list.

For those rated over 2300 on the ACF list the difference between their ACF and FIDE ratings is on average 29 points. For all those ACF rated over 2200 the average difference is 41 points and for all those ACF rated over 2100 its 58 points.



It is a pity that the ATSC was not submitted, although I think the time limit, 55 plus 10sec from start is actually rapid. Is that correct?
Scott
If that was the time limit then it is a rapid.
However I do know that previous ATSC comps have been rated as rapid.

Oepty
02-03-2004, 12:28 PM
2) When it came to determining the uplift we only looked at players who we considered active on the Dec 2003 ACF list and the Jan 2004 FIDE list. We also ignored all players whose ACF ratings were below 2050.


Bill. Thanks, but why ignore people with ACF ratings below 2050?
Scott

Bill Gletsos
02-03-2004, 12:47 PM
Bill. Thanks, but why ignore people with ACF ratings below 2050?
Scott
Because although not necessarily true for everyone, for the vast majority their FIDE ratings are statistically "crap". :whistle:

Rincewind
02-03-2004, 01:24 PM
If that was the time limit then it is a rapid.
However I do know that previous ATSC comps have been rated as rapid.

Out of interest, how do you determine whether a incremental time control is blitz, rapid or classical?

I assume you make take a reasonable number of moves and if the total time based on this measure was > than the thresholds of 15min and 60min then it would count as the higher time control.

A reasonable number of moves to consider would be in the range 30-40.

Now 55min+10s/move is 60 minutes if there are 30 moves. I would have thought this was serious enough to be counted as classical.

Can someone please enlighten me? Thanks in advance.

Kerry Stead
02-03-2004, 01:30 PM
Because although not necessarily true for everyone, for the vast majority their FIDE ratings are statistically "crap". :whistle:

:eek: :cry: so I'm just another crap statistic :cry:

Bill Gletsos
02-03-2004, 01:32 PM
Out of interest, how do you determine whether a incremental time control is blitz, rapid or classical?

I assume you make take a reasonable number of moves and if the total time based on this measure was > than the thresholds of 15min and 60min then it would count as the higher time control.

A reasonable number of moves to consider would be in the range 30-40.

Now 55min+10s/move is 60 minutes if there are 30 moves. I would have thought this was serious enough to be counted as classical.

Can someone please enlighten me? Thanks in advance.
You obviously have been too busy doing maths puzzles or discussing olympiad selections and missed the post http://www.chesschat.org/showpost.php?p=7863&postcount=3 ;)

Also check out http://www.chesschat.org/showpost.php?p=7949&postcount=13

Bill Gletsos
02-03-2004, 01:40 PM
:eek: :cry: so I'm just another crap statistic :cry:
Actually you are the reason I used the words "although not necessarily true for everyone". ;)

Rincewind
02-03-2004, 01:42 PM
You obviously have been too busy doing maths puzzles or discussing olympiad selections and missed the post http://www.chesschat.org/showpost.php?p=7863&postcount=3 ;)

Actually doing real math revision :eek: - and (I think) the Olympiad Selection furore had mostly settled down by then.

Follow up questions then, if I may.

What is the guidelines for distinguishing Rapid from Blitz?

Are those guidelines (Rapid/Blitz and Classical/Rapid) ACF guidelines or FIDE guidelines?

If the former, what is to be done regarding the application of the Blitz and Rapid appendices of the Laws or Chess?

I suppose the point has been made that 59min + 29sec/move gives you much more time than 30min + 30sec/move but the first is Rapid and the second is Classical. Where is the sense there?

No offense intended, just trying to get the issues and an uderstanding of the guidelines. Thanks in advance, again.

Kerry Stead
02-03-2004, 01:49 PM
Actually you are the reason I used the words "although not necessarily true for everyone". ;)

So where's my extra 150 points then? If I'm 2100-odd FIDE, shouldn't I be 2100-odd ACF too?? ;) :hmm:

Bill Gletsos
02-03-2004, 01:59 PM
So where's my extra 150 points then? If I'm 2100-odd FIDE, shouldn't I be 2100-odd ACF too?? ;) :hmm:
Nice try. :rolleyes:

Maybe since your 1969 ACF you should be 1969 FIDE. :doh: :whistle:

Bill Gletsos
02-03-2004, 02:03 PM
Actually doing real math revision :eek: - and (I think) the Olympiad Selection furore had mostly settled down by then.

Follow up questions then, if I may.

What is the guidelines for distinguishing Rapid from Blitz?

Are those guidelines (Rapid/Blitz and Classical/Rapid) ACF guidelines or FIDE guidelines?

If the former, what is to be done regarding the application of the Blitz and Rapid appendices of the Laws or Chess?

I suppose the point has been made that 59min + 29sec/move gives you much more time than 30min + 30sec/move but the first is Rapid and the second is Classical. Where is the sense there?

No offense intended, just trying to get the issues and an uderstanding of the guidelines. Thanks in advance, again.

As you would have noticed in those two posts I quoted, prior to FIDE making a ruling at Bled all games played with an increment were neither Blitz nor Rapid no matter how small the original time allocation was because the rules stated that Blitz and Rapid required all moves to be made in a fixed time control.

When we came up with the ACF definition this was well before FIDE had made this decision.

Now that FIDE have, I see no reason why the ACF Council should not review our definitions of what constitutes a normal time control.

Garvinator
02-03-2004, 02:59 PM
can i ask what was fide's decision regarding this matter? I think i already know but i just want to be sure :D

Bill Gletsos
02-03-2004, 03:21 PM
can i ask what was fide's decision regarding this matter? I think i already know but i just want to be sure :D
You obviously did not read the second message I provided the url for above.
Careful of you might get labelled with the g word. ;)

Garvinator
02-03-2004, 03:23 PM
You obviously did not read the second message I provided the url for above.
Careful of you might get labelled with the g word. ;)
im sorry no i didnt look at the url posted previously because i remembered the posts that had already taken place on this topic ;) .

And btw all i did was ask a question, i did not make a silly statement then carry it on :whistle:

Bill Gletsos
02-03-2004, 03:26 PM
im sorry no i didnt look at the url posted previously because i remembered the posts that had already taken place on this topic ;) .

And btw all i did was ask a question, i did not make a silly statement then carry it on :whistle:
Ok then maybe I'll just call you a slacker instead. ;)

Garvinator
02-03-2004, 03:30 PM
Ok then maybe I'll just call you a slacker instead. ;)
oh no ive started having images of you being mr strickland from back to the future trilogy :p

Bill Gletsos
02-03-2004, 03:34 PM
oh no ive started having images of you being mr strickland from back to the future trilogy :p
:whistle:

Garvinator
02-03-2004, 03:35 PM
:whistle: :lol: :lol: :owned: :owned:

Bill Gletsos
02-03-2004, 03:53 PM
:lol: :lol: :owned: :owned:
Careful or starter is going to accuse you of making useless posts to up your post count. :doh: :whistle:

Garvinator
02-03-2004, 03:57 PM
Careful or starter is going to accuse you of making useless posts to up your post count. :doh: :whistle:
and then by accusing me he gets up his post count, then i have to reply and so we both get a higher post count :eek: so its a win win situation :whistle:

Rincewind
02-03-2004, 04:02 PM
As you would have noticed in those two posts I quoted, prior to FIDE making a ruling at Bled all games played with an increment were neither Blitz nor Rapid no matter how small the original time allocation was because the rules stated that Blitz and Rapid required all moves to be made in a fixed time control.

When we came up with the ACF definition this was well before FIDE had made this decision.

Now that FIDE have, I see no reason why the ACF Council should not review our definitions of what constitutes a normal time control.

OK. So it looks like the FIDE standard is to be applied for the appendices and that uses a formula something like the one I assumed except with a number of 60 for the reasonable number of moves for a game (instead of 30-40). No problem with that.

I take it from your reply that this standard might also be applied to the ACF rating process which I think makes sense.

However, is there currently a standard defintion of what is too fast to be declared as rapid play? Obviously G5 is not rated in the rapid list. What about something like G10+5s/move? It's rapid by the FIDE definition, is it rapid by the ACF definition?

Bill Gletsos
02-03-2004, 04:07 PM
OK. So it looks like the FIDE standard is to be applied for the appendices and that uses a formula something like the one I assumed except with a number of 60 for the reasonable number of moves for a game (instead of 30-40). No problem with that.

I take it from your reply that this standard might also be applied to the ACF rating process which I think makes sense.

However, is there currently a standard defintion of what is too fast to be declared as rapid play? Obviously G5 is not rated in the rapid list. What about something like G10+5s/move? It's rapid by the FIDE definition, is it rapid by the ACF definition?
Actually as far as I recall we only rate rapids whose initial is 15 or greater.

Garvinator
02-03-2004, 04:09 PM
Actually as far as I recall we only rate rapids whose initial is 15 or greater.
not meaning to sound like a rating officer ;) but i remember something similar myself, that games with an initial time of 15 mins are considered to short for the rapid system.

Bill Gletsos
02-03-2004, 04:14 PM
not meaning to sound like a rating officer ;) but i remember something similar myself, that games with an initial time of 15 mins are considered to short for the rapid system.
No originally some people wanted to rate lightning in the rapid system.
We settled on a minimum time limit of 15 minutes for an ACF rated rapid event.

Rincewind
02-03-2004, 04:16 PM
Actually as far as I recall we only rate rapids whose initial is 15 or greater.

This leads to the whimsical example of 10min + 1min/move

Classical by the FIDE definition, too fast to be rated as even rapid by the ACF. ;) I assume if the ACF maintained a Blitz rating list it would find its way onto that.

Ian Rout
02-03-2004, 04:44 PM
I can see a case for rating guillotine finishes on the rapid list regardless of the initial time allocation, but the opposite, treating all non-guillotine finishes as non-rapid, is obviously nonsense.

Incidentally how rigorous are people in other States about rating rapid games? As far as I can see only about two rapid events per year are submitted from ACT.

Bill Gletsos
02-03-2004, 04:56 PM
This leads to the whimsical example of 10min + 1min/move

Classical by the FIDE definition, too fast to be rated as even rapid by the ACF. ;) I assume if the ACF maintained a Blitz rating list it would find its way onto that.
No, based on the decision at Bled it is considered by FIDE as a Blitz game.

Rincewind
02-03-2004, 05:07 PM
No, based on the decision at Bled it is considered by FIDE as a Blitz game.

Hang on 60 * 1 min + 10 min = 70 min

What am I missing?

Bill Gletsos
02-03-2004, 05:37 PM
I can see a case for rating guillotine finishes on the rapid list regardless of the initial time allocation, but the opposite, treating all non-guillotine finishes as non-rapid, is obviously nonsense.
I agree some non guillotines could be quite stupid.
Currently anything with an initial less than 15 irrespective of whether it has an increment or not is not rated in rapid or normal.
Currently anything between 15-59 mins with no increment is rated on the rapid.
Currently only tournaments with an initial of 30-59 and an increment greater than or equal to 30 from move one are rated as normal.


Incidentally how rigorous are people in other States about rating rapid games? As far as I can see only about two rapid events per year are submitted from ACT.
The main user of the Rapid list is QLD and they are mostly junior events. Victoria I think is second behind QLD.

Bill Gletsos
02-03-2004, 05:38 PM
Hang on 60 * 1 min + 10 min = 70 min

What am I missing?
Nothing I am. :doh:
I read it as 10min + 1sec/move.

Cat
02-03-2004, 10:02 PM
You were talking about learning from an experience. I simply asked which experience you were talking about because, as Bill points out, there are a range of different experiences or alleged experiences

As for compaction, points loss due to 'compaction' (assuming that is even still happening, which I doubt) would only affect those with high ratings. For those below the 1900s it was either a non-issue or a benefit.

If you don't think the 70 points can be explained purely by FIDE inflation then the onus is on you to explain why the ACF players with FIDE ratings have on average stayed remarkably static on the ACF scale. I'm not saying that the ACF stasis for these players proves the 70 points is FIDE inflation, but I'm assuming it until proven otherwise. For instance if the mean age of those players was far younger than the mean age of the pool as a whole, that might be of interest.

I'm not going to get into semantic arguments over cause and effect, but you know my feelings are that historically the rating system hasn't been dynamic enough to reflect junior development. Bill is right when he says that we (on the Gold Coast) have not been submitting enough of our games for ACF rating and although this was not done consciously, up until now it really wouldn't have been too fruitful because the Gold Coast Rating average was so low that in many of our tournaments players were essentially losing points simply by participating.

The situation has now changed and I feel genuinely optimistic. I will certainly be emphasing to the club how important it is that we take advantage of the situation to submit more of our games for rating. Since the establishment of the Gardiner Chess Centre there has been an even greater influx of new juniors into chess. My concern is that if we find that their playing strength moves rapidly ahead of their rating, we may find ourselves back in the same situation.

Anyway, for now things look good. My rating has gone up so much with so little effort that if I start playing chess as well I might be challenging for an olympaid spot in a few years! I think a few of our juniors will be looking at this old patzer and fancying their chances I dare say - that's got to be a good thing. So thanks Bill and you know you can rely on me to keep you posted on events. :lol:

Kevin Bonham
03-03-2004, 12:01 AM
I think a few of our juniors will be looking at this old patzer and fancying their chances I dare say - that's got to be a good thing.

Out of curiosity how did you find it playing against the juniors personally before the correction was applied? Were you getting thrashed by juniors hundreds of points below you, or were you improving fast enough yourself that they weren't that much of an issue?

Cat
03-03-2004, 11:12 AM
Out of curiosity how did you find it playing against the juniors personally before the correction was applied? Were you getting thrashed by juniors hundreds of points below you, or were you improving fast enough yourself that they weren't that much of an issue?

If you look at my rapid rating you'll find the answer! Its been steadily falling despite corrections. A year ago I was playing a young fellow rated about 1100. After a few moves he lost a pawn, my position looked comfortable and I thought I'd exchange into a winning endgame. The tenacious little guy got me into time trouble, I made a few weak moves and eventually blundered. He managed a few more upsets before the end of the comp - he was of course Moulthun Ly. In the last 12 months I've also lost to Chris Hardy and Kelvin Finke in rapid games, both excellent prospects and significantly underrated a year ago.

Following the release of March 2004 ratings, I analysed the rating distributions on the Coast and I note a paucity of players in the 1100-1300 rating bracket. In essence, we now have 2 population distributions, one with a mean around 900 and another around 1500. I suspect that over the years our adults in the 1000-1400 rating range have been feeling the heat from the juniors and have lost motivation.

Bill Gletsos
03-03-2004, 11:36 AM
Following the release of March 2004 ratings, I analysed the rating distributions on the Coast and I note a paucity of players in the 1100-1300 rating bracket. In essence, we now have 2 population distributions, one with a mean around 900 and another around 1500. I suspect that over the years our adults in the 1000-1400 rating range have been feeling the heat from the juniors and have lost motivation.
Of course my mate starter would say you should market/cater for the adults who have lost motivation.
You could therefore just run a few tournaments restricted to adults, especially for those adults who have lost motivation.

Rincewind
03-03-2004, 11:48 AM
Of course my mate starter would say you should market/cater for the adults who have lost motivation.
You could therefore just run a few tournaments restricted to adults, especially for those adults who have lost motivation.

Perhaps just include u1400 and u1200 rating prizes for seniors only in your existing tournaments. Watch out for unscrupulous 1500 seniors sandbagging their rating to get on the gravy train. :eek:

ursogr8
03-03-2004, 11:53 AM
Of course my mate starter would say you should market/cater for the adults who have lost motivation.
You could therefore just run a few tournaments restricted to adults, especially for those adults who have lost motivation.

Bill
I am sure I am not the only poster who is interested in tournament design. In fact it is probably the most popular topic on the board if you include

time controls
fees
discount
entry quals
location
round/day
prizes
rating divisions
etc.

And all designers would be interested in their 'customer's' view.

Perhaps Dave needs to focus on other than adults-only tourneys. :hand:

I use as guiding principles

> competitive games
>> recognise and reward

when thinking about tournament design.

But then I end up in favour of Fischer_Random, so, many people think my methodology has a screw loose. ;)

starter

Cat
03-03-2004, 11:56 AM
Yes, all these things are part of the solution, but of course we also want to encourage junior v's adult competition as part of the chess culture. I hope you're not referring to me BJC? Moi! Sandbagging??

Bill Gletsos
03-03-2004, 12:10 PM
Yes, all these things are part of the solution, but of course we also want to encourage junior v's adult competition as part of the chess culture.
I wasnt suggesting that you have no tournaments where juniors play adults.
I was just suggesting you may want to have some.

Bill Gletsos
03-03-2004, 12:12 PM
But then I end up in favour of Fischer_Random, so, many people think my methodology has a screw loose. ;)
You probably lucky the "l" and the "g" key are not close to each other otherwise a typing error could be embarrassing. ;) :whistle:

Rincewind
03-03-2004, 12:28 PM
I hope you're not referring to me BJC? Moi! Sandbagging??

Honest Avatar,

At the time, I wasn't referring to anyone in particular. As they say in the movies, any resemblance to any actual chess players, living or dead, is purely coincidental. But if the shoe fits... :lol:

ursogr8
03-03-2004, 12:41 PM
Perhaps just include u1400 and u1200 rating prizes for seniors only in your existing tournaments. Watch out for unscrupulous 1500 seniors sandbagging their rating to get on the gravy train. :eek:

Baz :eek:

I thought we had all agreed that 'sand-bagging' was prevented by leaving the choice of thresh-holds on rating prizes until after the first round. And then if there are to be 5 divisions (say) just do the maths to split the field equally.
In other words....don't advertise at the start, UNDER 1400 prize available.

starter

Rincewind
03-03-2004, 04:46 PM
I thought we had all agreed that 'sand-bagging' was prevented by leaving the choice of thresh-holds on rating prizes until after the first round. And then if there are to be 5 divisions (say) just do the maths to split the field equally.
In other words....don't advertise at the start, UNDER 1400 prize available.

There are two problems with that approach. One general and one specific to the problem David is trying to address.

The general problem is it leaves a question mark over the divisional prize threshold settings and whoever determines the thresholds could be accused of division fixing. Therefore issues of conflicts of interest need to be clearly addressed.

The specific problem is that in this case the object is specifically the attract a certain market sector back to tournament play. So it would be expected that the u1200 and u1400/seniors divisions would be under represented. The idea is to provide additional incentive to those players.

Alan Shore
03-03-2004, 06:09 PM
Here's something to think about though.. there is no law against sandbagging... it's impossible to prove too. But I have to agree with Barry; I think rating prizes should be advertised - you'd feel pretty hard done by if you're last in your rating category and have little chance of attaining anything from the tournament, which one may have only played in to win something. A lot of chess players are only after the money, even the top ones (I won't name names) but let's face it, one can play chess on the internet (FICS, ICC etc.) without having to fork out $50 entry fees with no hope of winning a cent.

ursogr8
03-03-2004, 06:52 PM
There are two problems with that approach. One general and one specific to the problem David is trying to address.

The general problem is it leaves a question mark over the divisional prize threshold settings and whoever determines the thresholds could be accused of division fixing. Therefore issues of conflicts of interest need to be clearly addressed.

The specific problem is that in this case the object is specifically the attract a certain market sector back to tournament play. So it would be expected that the u1200 and u1400/seniors divisions would be under represented. The idea is to provide additional incentive to those players.

Well Baz, you see it is like this south of the border

1 Our administrators are as clean as driven snow, and highly respected. So, when they say the "the divisionl breaks are here, here and here", then the players say "but of course". So problem #1 is not a VIC problem.
2 I concede #2 when the tournament has the objective you imply. Yes, truth in advertising must be backed by actuality.

starter

ursogr8
03-03-2004, 07:01 PM
You probably lucky the "l" and the "g" key are not close to each other otherwise a typing error could be embarrassing. ;) :whistle:

Bill, :eek:

The thread with the new ratings seems to be petering out after only 115 posts. What has happened? :confused: It used to go on and on.
Are you doing something right?
You, personally seemed to be reduced to imagining dyslexia.

starter

Bill Gletsos
03-03-2004, 10:01 PM
Bill, :eek:

The thread with the new ratings seems to be petering out after only 115 posts. What has happened? :confused: It used to go on and on.
Are you doing something right?
We believe so.

You, personally seemed to be reduced to imagining dyslexia.
Nothing dyslexic about the words loose and goose. :whistle:

Cat
04-03-2004, 06:21 PM
Bill, :eek:

The thread with the new ratings seems to be petering out after only 115 posts. What has happened? :confused: It used to go on and on.
Are you doing something right?
You, personally seemed to be reduced to imagining dyslexia.

starter

Yes, I think the problems have largely been addressed for the moment and generally everyone is pretty pleased with the outcome. Bill feels confident the Glicko system is dynamic enough to cope with junior ratings into the future, I remain a little sceptical. I know many of our local juniors remain significantly underrated, but with the injection of rating points into the Gold Coast region the situation should dramatically improve over the next 6-12 months as the points redistribute.

chesslover
04-03-2004, 09:20 PM
Bill, :eek:

The thread with the new ratings seems to be petering out after only 115 posts. What has happened? :confused: It used to go on and on.
Are you doing something right?
starter

Bill, our beloved Supreme Leader is THE MAN!!!! or rather he is the ACFrating officer :)

I think that over time Bill has convinced most of the detractors of the Glicko2's correctness. I also think that most of the problems came about when there were some titled players who argued about Glicko2 when Bill introduced it, but I guess practical experience of the system has proven it's merits

Well done Supreme Leader!! Long may you reign!!

Rincewind
04-03-2004, 10:48 PM
Bill, our beloved Supreme Leader is THE MAN!!!! or rather he is the ACFrating officer :)

I think that over time Bill has convinced most of the detractors of the Glicko2's correctness. I also think that most of the problems came about when there were some titled players who argued about Glicko2 when Bill introduced it, but I guess practical experience of the system has proven it's merits

Well done Supreme Leader!! Long may you reign!!

Is that the first for the year? And not till March. Quite a record! ;)

Alan Shore
04-03-2004, 10:48 PM
Bill, our beloved Supreme Leader is THE MAN!!!! or rather he is the ACFrating officer :)

I think that over time Bill has convinced most of the detractors of the Glicko2's correctness. I also think that most of the problems came about when there were some titled players who argued about Glicko2 when Bill introduced it, but I guess practical experience of the system has proven it's merits

Well done Supreme Leader!! Long may you reign!!

I think Mr chesslover may be starting the 'Bill Gletsos Fan Club' ...! ;)

Garvinator
04-03-2004, 10:54 PM
I think Mr chesslover may be starting the 'Bill Gletsos Fan Club' ...! ;)
what do you mean start, cl is the inaugural member and has life membership :lol: :lol: :lol:

Alan Shore
04-03-2004, 10:58 PM
what do you mean start, cl is the inaugural member and has life membership :lol: :lol: :lol:

Hehe, I see.. this must go back further than I had first imagined :)

Bill Gletsos
04-03-2004, 11:02 PM
what do you mean start, cl is the inaugural member and has life membership :lol: :lol: :lol:
Unfortunately for him, I want no involvement with the club. :hand:

Bill Gletsos
04-03-2004, 11:04 PM
Is that the first for the year? And not till March. Quite a record! ;)
Yes, mostly this year he has been busy "hoovering" up to the likes of Matt, David, Paulb and Kevin. :whistle:

Cat
04-03-2004, 11:24 PM
Bill, our beloved Supreme Leader is THE MAN!!!! or rather he is the ACFrating officer :)

I think that over time Bill has convinced most of the detractors of the Glicko2's correctness. I also think that most of the problems came about when there were some titled players who argued about Glicko2 when Bill introduced it, but I guess practical experience of the system has proven it's merits

Well done Supreme Leader!! Long may you reign!!

Look, Bill has done a good job and deserves credit. Is this post deliberately designed to visit fire & brimstone upon him as retribution for his tawdriness towards you, or are you just being naive? It seems there are just a few too many pokers in the fire.

Alan Shore
04-03-2004, 11:27 PM
It sounded sincere enough to me David.. just.. a little excessive :)

Cat
04-03-2004, 11:59 PM
It sounded sincere enough to me David.. just.. a little excessive :)

It's not a matter of essential correctness, but recognition of a rating system that has gone through a number of corrections to arrive at a point that is mutually agreeable.

Bill Gletsos
05-03-2004, 04:26 PM
It's not a matter of essential correctness, but recognition of a rating system that has gone through a number of corrections to arrive at a point that is mutually agreeable.
If you believe that then you are kidding yourself.
Any changes to the system have nothing to do with arriving at a point that is mutally agreeable.

1) We introduced the Glicko system, because it was superior to ELO. The Glicko2 wasn't publically available at that time. We enchanced how unrated players were first given a rating. We also didnt allow players to increase or decrease their ratings by more than their performance ratings.
2) We introduced the Glicko2 system which was better thanh the Glicko.
3) We identified a potential problem with players over 2200 and made allowances for it.
4) We corrected the ACT junior problem. A problem that had absolutely nothing to do with Glicko. It stemmed from the ELO system from the 90's.
5) We made a correction at the same time for the Gold Coast even though the cause of 4 and 5 are essentially different. The Gold coast problem also has its roots in the ELO system. Now our correction here was essentially a little too generous, but it only really effected a limited number of people.

Now virtually all of the changes were planned and were done because Graham and I believed they were in the interest of the ratings system, not because it might actually please anybody.

Now a few months prior to introduction of the Glicko2 we suspected 3 but hadnt run any actual simulations to verify it and were looking at how to handle 4.
However after introducing the Glicko2 we wanted to see what effect it might have on 3 and 4. We therefore delayed doing anything about 3 and continued looking at 4.
The rating debate then started in the ACF Bulletin in March 2003. All this did was delay us even more in handling 4.
We then introduced 3 whilst still working on 4.
We then did 4 and at the same time now also did 5 because it was convenient.

Cat
05-03-2004, 06:46 PM
If you believe that then you are kidding yourself.
Any changes to the system have nothing to do with arriving at a point that is mutally agreeable.

1) We introduced the Glicko system, because it was superior to ELO. The Glicko2 wasn't publically available at that time. We enchanced how unrated players were first given a rating. We also didnt allow players to increase or decrease their ratings by more than their performance ratings.
2) We introduced the Glicko2 system which was better thanh the Glicko.
3) We identified a potential problem with players over 2200 and made allowances for it.
4) We corrected the ACT junior problem. A problem that had absolutely nothing to do with Glicko. It stemmed from the ELO system from the 90's.
5) We made a correction at the same time for the Gold Coast even though the cause of 4 and 5 are essentially different. The Gold coast problem also has its roots in the ELO system. Now our correction here was essentially a little too generous, but it only really effected a limited number of people.

Now virtually all of the changes were planned and were done because Graham and I believed they were in the interest of the ratings system, not because it might actually please anybody.

Now a few months prior to introduction of the Glicko2 we suspected 3 but hadnt run any actual simulations to verify it and were looking at how to handle 4.
However after introducing the Glicko2 we wanted to see what effect it might have on 3 and 4. We therefore delayed doing anything about 3 and continued looking at 4.
The rating debate then started in the ACF Bulletin in March 2003. All this did was delay us even more in handling 4.
We then introduced 3 whilst still working on 4.
We then did 4 and at the same time now also did 5 because it was convenient.

I know you better than to think you would go out of your way to please me Bill. And I have no arguments with anything you've said. It was a poor turn of phrase of mine delivered in a hurry late at night. 'Mutually agreeable' was not meant to be a viz a viz statement - I meant there was general satisfaction with the outcome of the rating changes, nothing more, nothing less. It was given in answer to CL & Starter as to why there was no criticism on this rating thread - I think we're all happy aren't we?

ursogr8
07-03-2004, 12:45 PM
You probably lucky the "l" and the "g" key are not close to each other otherwise a typing error could be embarrassing. ;) :whistle:

Bill

When the new ratings were placed on our notice-board last Friday night, together with the rationale of maintaining parity with FIDE ratings, there was tremendous
> interest in the numbers
>> positive acceptance of the rationale
>>> goodwill generated to the ratings officers.

A succesful PR exercise all-round and a definite step forward in the understanding by our juniors of how ratings are maintained. Well done.

starter

ps I could have summarised that to 'every one likes 70 points added". ;)

Alan Shore
07-03-2004, 05:23 PM
I'll guarantee you that people would have looked less favourably on a 70 point decrease if it had of been so required :)

arosar
08-03-2004, 08:57 AM
I'll just dump this here: http://www.bworld.com.ph/current/A&L/ang.html

AR

Rhubarb
08-03-2004, 02:32 PM
I'll just dump this here: http://www.bworld.com.ph/current/A&L/ang.html

AR

Taking your lead, Amiel, I'll just dump this here:

ACF Master File Top 50, March 2004


1 2668 !! Rogers, Ian NSW
2 2567 ?? Levin, Naum L NSW
3 2561 Wohl, Aleksander H QLD
4 2548 ? Jamieson, Robert M VIC
5 2538 ? Gedevani, Dimitri NSW
6 2509 !! Johansen, Darryl K VIC
7 2505 ! Wallace, John-Paul NSW
8 2493 !! Lane, Gary W NSW
9 2491 ?? Travers, Roy B NSW
10 2488 ?? Kaltenbacher, Sorin NSW
11 2468 ? Davidovic, Aleksanda SA
12 2468 ? Prods, Arvids VIC
13 2464 ? Kanikevich, Alexande NSW
14 2458 ! Gluzman, Michael VIC
15 2456 ?? Brodie, Alaistar I NSW
16 2454 ? Kontorovich, Michael VIC
17 2448 ? Goldenberg, Igor VIC
18 2446 !! Solomon, Stephen J QLD
19 2439 ! Tao, Trevor SA
20 2436 ? Belin, Igor NSW
21 2427 ? Smith, Murray D NSW
22 2422 Speck, Nicholas S VIC
23 2418 ? Selim, Nabil NSW
24 2408 ? Pickles, Stephen M NSW
25 2406 ?? Brown, Philip J NSW
26 2400 !! Froehlich, Peter VIC
27 2393 ?? Kewley, Douglas SA
28 2393 Jordan, Bill VIC
29 2390 ?? Kerr, Stephen G R NSW
30 2389 ?? Zeidler, Sven VIC
31 2389 ?? Chavich, Tom ACT
32 2383 ! West, Guy VIC
33 2381 !! Chapman, Mark SA
34 2380 !! Smerdon, David C VIC
35 2380 !! Zhao, Zong-Yuan NSW
36 2378 !! Bjelobrk, Igor VIC
37 2378 ? Vassilaropoulos, Nic SA
38 2371 !! Sandler, Leonid VIC
39 2368 ?? O'Conner, M WA
40 2367 ?? Horton, Les SA
41 2366 Rutherford, Simon A VIC
42 2366 ?? Sulik, Frank X SA
43 2366 ! Teichmann, Erik VIC
44 2365 ?? Fardell, Dan NSW
45 2361 ?? Bombek, Peter NSW
46 2359 !! Depasquale, Chris J VIC
47 2357 ?? Maclaurin, Norman D NSW
48 2355 ?? Jenkinson, Simon W VIC
49 2354 ?? Bell, Graham M WA
50 2352 ? Nutu-Gajic, Daniela SA

Well I hate to be a spoilsport, bulletin board members, but not everyone is satisfied with the rating system and Bill Gletsos's increasingly vociferous defence of it.

I thought that the latest correction was supposed to realign ACF ratings with FIDE ratings, but clearly the ratings of many people on this list (those without !! next to their name) are so absurd as to be beyond a joke. Discounting the ones who are deceased, many of them were probably over-rated before they were gifted 220 points.

Now of course, Glicko2 claims to be able to correct this as soon as the non-active players start playing again, but meanwhile they're just sitting there with ridiculously inflated ratings.

So why were completely inactive players given 220 points?

Why does Bill argue so vehemently for his beloved Glicko when the top end becomes ever more farcical?

If the answer is that inactive players' ratings are essentially meaningless, why do they have the right to be on the Master File at all?

Either remove inactive players entirely or use a system that can realistically assess them, I say.

P.S. On a personal note, it occurs to me that if I had taken a hiatus from chess a few years back, when my "pre-Glicko" ACF rating was 2280-something, I would be over 2500 by now. Ridiculous to be sure, but as not as ridiculous as numbers 9, 10, 12 and 15 on the above list, to take the most egregious examples.

P.P.S. I couldn't seem to find the reference in this thread where Bill explains to the NSWCA council members that one of the reasons the FIDE system is inherently inflationary is because "a tournament winner can't lose rating points". I wonder why I couldn't find it?

Bill Gletsos
08-03-2004, 03:36 PM
Taking your lead, Amiel, I'll just dump this here:

ACF Master File Top 50, March 2004


1 2668 !! Rogers, Ian NSW
2 2567 ?? Levin, Naum L NSW
3 2561 Wohl, Aleksander H QLD
4 2548 ? Jamieson, Robert M VIC
5 2538 ? Gedevani, Dimitri NSW
6 2509 !! Johansen, Darryl K VIC
7 2505 ! Wallace, John-Paul NSW
8 2493 !! Lane, Gary W NSW
9 2491 ?? Travers, Roy B NSW
10 2488 ?? Kaltenbacher, Sorin NSW
11 2468 ? Davidovic, Aleksanda SA
12 2468 ? Prods, Arvids VIC
13 2464 ? Kanikevich, Alexande NSW
14 2458 ! Gluzman, Michael VIC
15 2456 ?? Brodie, Alaistar I NSW
16 2454 ? Kontorovich, Michael VIC
17 2448 ? Goldenberg, Igor VIC
18 2446 !! Solomon, Stephen J QLD
19 2439 ! Tao, Trevor SA
20 2436 ? Belin, Igor NSW
21 2427 ? Smith, Murray D NSW
22 2422 Speck, Nicholas S VIC
23 2418 ? Selim, Nabil NSW
24 2408 ? Pickles, Stephen M NSW
25 2406 ?? Brown, Philip J NSW
26 2400 !! Froehlich, Peter VIC
27 2393 ?? Kewley, Douglas SA
28 2393 Jordan, Bill VIC
29 2390 ?? Kerr, Stephen G R NSW
30 2389 ?? Zeidler, Sven VIC
31 2389 ?? Chavich, Tom ACT
32 2383 ! West, Guy VIC
33 2381 !! Chapman, Mark SA
34 2380 !! Smerdon, David C VIC
35 2380 !! Zhao, Zong-Yuan NSW
36 2378 !! Bjelobrk, Igor VIC
37 2378 ? Vassilaropoulos, Nic SA
38 2371 !! Sandler, Leonid VIC
39 2368 ?? O'Conner, M WA
40 2367 ?? Horton, Les SA
41 2366 Rutherford, Simon A VIC
42 2366 ?? Sulik, Frank X SA
43 2366 ! Teichmann, Erik VIC
44 2365 ?? Fardell, Dan NSW
45 2361 ?? Bombek, Peter NSW
46 2359 !! Depasquale, Chris J VIC
47 2357 ?? Maclaurin, Norman D NSW
48 2355 ?? Jenkinson, Simon W VIC
49 2354 ?? Bell, Graham M WA
50 2352 ? Nutu-Gajic, Daniela SA

Well I hate to be a spoilsport, bulletin board members, but not everyone is satisfied with the rating system and Bill Gletsos's increasingly vociferous defence of it.

I thought that the latest correction was supposed to realign ACF ratings with FIDE ratings, but clearly the ratings of many people on this list (those without !! next to their name) are so absurd as to be beyond a joke. Discounting the ones who are deceased, many of them were probably over-rated before they were gifted 220 points.

Now of course, Glicko2 claims to be able to correct this as soon as the non-active players start playing again, but meanwhile they're just sitting there with ridiculously inflated ratings.

So why were completely inactive players given 220 points?
Greg,

Firstly there is nothing secret about this list so to suggest its a list people are not supposed to see is simply wrong. After all any man and his dog can create it using Excel and the master file list off the ACF web site.

Secondly this isnt a Glicko v ELo issue.

After all the original 150 points was added well before the Glicko was introduced.

Without the original 150 point increase that virtually all players received then many of the top active players would now be rated below many of the inactive players anyway. This situation would not have been due to any deflation in the ACF ratings.

Now in the current situation if you take the 70 points off all the players the top 10 are:

1 2598 !! Rogers, Ian NSW
2 2497 ?? Levin, Naum L NSW
3 2491 Wohl, Aleksander H QLD
4 2478 ? Jamieson, Robert M VIC
5 2468 ? Gedevani, Dimitri NSW
6 2439 !! Johansen, Darryl K VIC
7 2435 ! Wallace, John-Paul NSW
8 2423 !! Lane, Gary W NSW
9 2421 ?? Travers, Roy B NSW
10 2418 ?? Kaltenbacher, Sorin NSW

Is Levin or Jamieson really that strong.
I doubt it.
But it doesnt matter. They arent active.
They are not going to get selected for Olympiad teams etc.

Finally the point to remember is all ratings are relative.
If you add points to player A then you should add them to players B, C and D.


Why does Bill argue so vehemently for his beloved Glicko when the top end becomes ever more farcical?
Try not to replace fact with fiction.

As noted above the first 150 points all occured under ELO.
This isnt a Glicko issue or an ELO issue.
The same situation would occur if we still used ELO, except ELO wouldnt be able to correct itself when/if the players every return to active chess.


If the answer is that inactive players' ratings are essentially meaningless, why do they have the right to be on the Master File at all?
Once rated always rated.
Its no different than Bobby Fischer still being rated on the FIDE list at 2780.


Either remove inactive players entirely or use a system that can realistically assess them, I say.
If I used ELO the same problem would still exist, just like it did back in Dec 1999/April 2000 under ELo when the 150 point uplift occurred.


P.S. On a personal note, it occurs to me that if I had taken a hiatus from chess a few years back, when my "pre-Glicko" ACF rating was 2280-something, I would be over 2500 by now. Ridiculous to be sure, but as not as ridiculous as numbers 9, 10, 12 and 15 on the above list, to take the most egregious examples.
Firstly you only reached 2286 after the 150 point increase in the April 2000 list. Without it you would have been 2136 on the April 2000 list.
So you would have only reached a maximum of 2286+70 of 2356.
Not really that close to 2500 after all.


P.P.S. I couldn't seem to find the reference in this thread where Bill explains to the NSWCA council members that one of the reasons the FIDE system is inherently inflationary is because "a tournament winner can't lose rating points". I wonder why I couldn't find it?
Actually I believe I originally made that comment back on the old BB.
Also if someone is going to quote what i said to the NSWCA Council members then it would be good if they actually got it right.
I believe my comment was that over many years the FIDE system has been infalting for various reasons.
I suspect your informant just assumed I meant the last 3-4 years.
If so he was mistaken as I was referring to the situation from the 1980's onwards. After all it was from the 1980's to 1999 that was used to look at the FIDE inflation and led to the 150 point increase.

The main reason was the rating floor. however this really only became a problem once it dropped from 2200 to 2000 and so many more national tournaments started to be FIDE rated. Another cause was Schevingan tournaments.
I did also mention that the fact that the winner of a tournament did not lose rating points had led to inflation and that this was why the Super GM's ratings originally inflated.
The operative word there is had. I did not say it was a current problem. The reason I did not say that is because FIDE removed it a number of years ago.

chesslover
08-03-2004, 05:31 PM
I'll guarantee you that people would have looked less favourably on a 70 point decrease if it had of been so required :)
since most of teh rating points went up, most people were happy

if everyone had gone down by 70 points instead, then people would be unhappy

since rating is relative, as Bill keeps on saying, why not have regular increases every rating period to keep people happy? that way the rating points go up but the relative performance stays the same

people will be happy every time the rating results are released

Rhubarb
08-03-2004, 05:44 PM
Greg,

Bill,

Thank you for your measured response.

I agree, let us "try not to replace fact with fiction".

[QUOTE]"Firstly there is nothing secret about this list so to suggest its a list people are not supposed to see is simply wrong. After all any man and his dog can create it using Excel and the master file list off the ACF web site."
I did not claim that the Master File list was "secret"; simply that it is never published anywhere such that the everyday viewer can see the extreme upwards distortion of the ratings of inactive players.



"Secondly this isnt a Glicko v ELo issue. After all the original 150 points was added well before the Glicko was introduced."
I did not claim that the original 150 points were added under the Glicko system; simply that they were added indiscriminately to players - active and inactive.



"Without the original 150 point increase that virtually all players received then many of the top active players would now be rated below many of the inactive players anyway."
Yes Bill, but the inactive players got 150 points as well. Your argument is just obfuscation.



"This situation would not have been due to any deflation in the ACF ratings."
Regardless, I am not arguing about any inflation/deflation in ACF ratings. Once again, your sentence is pointless.



"Now in the current situation if you take the 70 points off all the players the top 10 are:

1 2598 !! Rogers, Ian NSW
2 2497 ?? Levin, Naum L NSW
3 2491 Wohl, Aleksander H QLD
4 2478 ? Jamieson, Robert M VIC
5 2468 ? Gedevani, Dimitri NSW
6 2439 !! Johansen, Darryl K VIC
7 2435 ! Wallace, John-Paul NSW
8 2423 !! Lane, Gary W NSW
9 2421 ?? Travers, Roy B NSW
10 2418 ?? Kaltenbacher, Sorin NSW

Is Levin or Jamieson really that strong.
I doubt it.
But it doesnt matter. They arent active.
They are not going to get selected for Olympiad teams etc."
What the...? Now you're arguing my case. That is exactly my point. If you agree that inactive players aren't that strong, why gift them points? And while we're on the top 10 case (even though I started with the top 50), why is Alex Wohl so high, relative to FIDE? The answer is because he hasn't recently played in ACF-rated events. Once again, Bill, why does he get 70 points for free? (Let's not even start on numbers 9 and 10 in your list, the ratings of whom even you surely cannot defend.)



"Finally the point to remember is all ratings are relative.
If you add points to player A then you should add them to players B, C and D."
Yes Bill, I do actually comprehend that ratings are relative. But once again, you have to distinguish between active and inactive players. If active players in a particular rating segment are losing points under a system - no matter how statistically valid that rating system may be internally - then it is folly to give rating points to inactive players when you want to make "corrections", i.e. realign it with the FIDE rating system.




"Try not to replace fact with fiction."
I'm honestly trying Bill. I think you should start trying a lot harder.



"As noted above the first 150 points all occured under ELO.
This isnt a Glicko issue or an ELO issue.
The same situation would occur if we still used ELO, except ELO wouldnt be able to correct itself when/if the players every return to active chess."
With my limited understanding, I believe the Glicko system may well be statistically superior to ELO - but only under perfect conditions, which may or may not be applicable to a small and widespread chess community such as Australia's. But I'm not arguing the relative merits of ELO and Glicko, just the indiscriminate "corrections" to inactive players.



"Once rated always rated.
Its no different than Bobby Fischer still being rated on the FIDE list at 2780."
Well yes it is different, Bill, because the FIDE list is standalone and - inflation or not - is unaffected by other systems, unlike the ACF system which is adjusted to realign with FIDE to the benefit of inactive players. (As an aside, was the Fischer-Spassky contraband match of 1991 actually FIDE-rated? If so, that would surely have contravened UN sanctions? Oh no, hang on, we're talking about FIDE.)



"If I used ELO the same problem would still exist, just like it did back in Dec 1999/April 2000 under ELo when the 150 point uplift occurred."
I don't doubt there are massive problems, the possible solutions to which are extremely complicated. I'm simply pointing out an obvious flaw.



"Firstly you only reached 2286 after the 150 point increase in the April 2000 list. Without it you would have been 2136 on the April 2000 list.
So you would have only reached a maximum of 2286+70 of 2356.
Not really that close to 2500 after all."
When I said a "few years back", I admit I was stretching the definition a bit. I was referring to my December 1990 rating of 2284. Nevertheless, it's disingenous of you to suggest that I would have hypothetically added 150 points to my rating twice. My original point is that if I had stopped playing in early 1991, my ACF rating would now be 2504. When you consider that many of the players on the ACF top 50 list haven't played a game since the seventies, it's not so ridiculous.



"Actually I believe I originally made that comment back on the old BB.
Also if someone is going to quote what i said to the NSWCA Council members then it would be good if they actually got it right.
I believe my comment was that over many years the FIDE system has been infalting for various reasons.
I suspect your informant just assumed I meant the last 3-4 years.
If so he was mistaken as I was referring to the situation from the 1980's onwards. After all it was from the 1980's to 1999 that was used to look at the FIDE inflation and led to the 150 point increase.
The main reason was the rating floor. however this really only became a problem once it dropped from 2200 to 2000 and so many more national tournaments started to be FIDE rated. Another cause was Schevingan tournaments.
I did also mention that the fact that the winner of a tournament did not lose rating points had led to inflation and that this was why the Super GM's ratings originally inflated.
The operative word there is had. I did not say it was a current problem. The reason I did not say that is because FIDE removed it a number of years ago."
I don't want to get into this argument about your discussions with council members as to why the FIDE system is inflationary (I agree it is anyway). Suffice it to say, you yourself replied to "your informant", as you call him, in this very bulletin board not two weeks ago - but the evidence has mysteriously disappeared...

Regards,
Greg

Rhubarb
08-03-2004, 05:49 PM
I'm sorry everyone, but I'm new to this BB. Is there a way I can reply to particular sections of a post to make it clear as to who's saying what? In the above, all of Bill's previous comments are in inverted commas.

chesslover
08-03-2004, 05:59 PM
Greg,

Without the original 150 point increase that virtually all players received then many of the top active players would now be rated below many of the inactive players anyway. This situation would not have been due to any deflation in the ACF ratings.


Is Levin or Jamieson really that strong.
I doubt it.
But it doesnt matter. They arent active.
They are not going to get selected for Olympiad teams etc.

Finally the point to remember is all ratings are relative.
If you add points to player A then you should add them to players B, C and D.




But Greg said that inactive players should not have points added on. I think that makes perfect sense. If they have not played for so long, then why add points to these players?

How can you say that Jammo and Levin are not going to get selected. I thought in the Olympic thread, it was decided that rating was not the only source of selection - which means that based on other factors Jammo and levin could be selected

skip to my lou
08-03-2004, 05:59 PM
use the quote tags

chesslover
08-03-2004, 06:08 PM
use the quote tags

how do you use quote tags? :confused:

i know how to reply back to the whole post of the person I am replying to, but do not know how to do replay back with my comments after each paragraph or sentence

Kerry Stead
08-03-2004, 06:09 PM
I'm sorry everyone, but I'm new to this BB. Is there a way I can reply to particular sections of a post to make it clear as to who's saying what? In the above, all of Bill's previous comments are in inverted commas.

Why yes Greg, there are easy ways to do things here.

To do the above, I clicked on the Reply bottun in your post, and it quoted your post in a separate colour (far more easy to distinguish from using quotation marks).

Also in the reply box (not in the 'Quick Reply' box at the foot of each page, rather the box you get when you click 'Post Reply'), there is a series of buttons on the toolbar, and the one on the far right is the quote button.

Alternatively, you can use the tag [QUOTE] to start a quote, and [ /QUOTE] (without the space) to close the quote.

Either way, its all pretty straightforward to use one you get used to it! :hmm:

At least now Bill will be sure to work on his grammatical errors before he posts ... I'm sure you'll be keeping a sharp eye on him in that regard! :doh:

Bill Gletsos
08-03-2004, 07:00 PM
But Greg said that inactive players should not have points added on. I think that makes perfect sense. If they have not played for so long, then why add points to these players?
Because you goose, ratings are relative.
If player A gets X points player B should also get X points irrespective of whether they played or not.


How can you say that Jammo and Levin are not going to get selected. I thought in the Olympic thread, it was decided that rating was not the only source of selection - which means that based on other factors Jammo and levin could be selected
Because you goose they have no recent results at all for any selector to determine how strong they are.

Rhubarb
08-03-2004, 07:04 PM
use the quote tags

Thanks Jeo - big of you.


Why yes Greg, there are easy ways to do things here.

To do the above, I clicked on the Reply bottun in your post, and it quoted your post in a separate colour (far more easy to distinguish from using quotation marks).

Also in the reply box (not in the 'Quick Reply' box at the foot of each page, rather the box you get when you click 'Post Reply'), there is a series of buttons on the toolbar, and the one on the far right is the quote button.

Alternatively, you can use the tag to start a quote, and (without the space) to close the quote.

Either way, its all pretty straightforward to use one you get used to it! :hmm:

At least now Bill will be sure to work on his grammatical errors before he posts ... I'm sure you'll be keeping a sharp eye on him in that regard! :doh:

thanks, got it now and will try not to get used to it :p

grammatical errors? i believe you have your diplomatic hat on now Kezza :eh:

Alan Shore
08-03-2004, 07:20 PM
Regarding the issue brought up by Greg Canfell, yes, those players are most likely overrated, yet inactive. The implication is, when/if they do return to chess, the glicko system's power can be shown - for example, if Mr Levin returned to chess and performed at only around 2200 strength, his new rating would immediately reflect the discrepancy between rating and performance, moreso because of the '??' rating differential. The new rating I would estimate for him would be around 2300, a very big drop, but very much in line with a 2200 performance with such an uncertain 2500+ rating.

Now if we were using the old ELO system.. yikes.. we'd have a nasty situation!

Kevin Bonham
08-03-2004, 07:25 PM
On the issue of inactive players being overrated (Bill's heard a lot of this before, but Greg may not have) :

(i) Previously I pointed out that since Glicko had started, top players (those above 2000-ish) had generally been shedding points without the Australian pool as a whole deflating. No-one has yet proved exactly why this is happening but my view was that the ratings pool had been overstretched by the under-responsiveness and approximations of past ratings systems and that this was being rectified. I haven't looked at the impact of the top-player dynamism reduction on this, or whether it's still happening.

(ii) Keeping a player's rating the same when they've been inactive for decades is methodically dodgy, even if it is common practice. Just because FIDE do it means nothing; after all FIDE still use ELO, which we regard as pretty much discredited. Fischer still being rated 2780 by FIDE is a joke. Super-GMs of Fischer's era who have remained active over the same period have lost points as age has caused their abilities to decline. Why should an inactive player be assumed to have kept their skills intact over a period in which strength loss is so likely even in an active player?

(iii) I agree with Bill that over-rated rusties are not a big issue, no one is going to take their ratings seriously for any normal purpose and if they come out and play any games they won't appear in the top players list and their ratings will be quickly sorted out by the way Glicko works. However, there is one valid argument for making the ratings of inactive players more realistic - accurate seeding for Swiss draws should those players enter a Swiss.

I know Bill is investigating this behind the scenes and I'll wait to see what comes out of it - it should be done through sound analysis rather than just gut feels that player X is overrated. My hunch is that a decay factor for inactivity will prove to be at least arguably mathematically sound. Not more than 10 points per year at a guess, and not kicking in for several years. However Bill and Graeme may well come to a different conclusion.

Bill Gletsos
08-03-2004, 07:44 PM
I did not claim that the Master File list was "secret"; simply that it is never published anywhere such that the everyday viewer can see the extreme upwards distortion of the ratings of inactive players.
Because its irrelevant.


I did not claim that the original 150 points were added under the Glicko system; simply that they were added indiscriminately to players - active and inactive.
Sorry but I have to disagree.
You said:

Why does Bill argue so vehemently for his beloved Glicko when the top end becomes ever more farcical?
That implies this is a Glicko issue.



Yes Bill, but the inactive players got 150 points as well. Your argument is just obfuscation.
I disagree.
Its a simple fact that ratings are relative.
If you add X points to player A you should add X points to player B, irrespective of whether they are active or not.
now there maybe be reasons why you would not do this but the most relevant ones are not applicable in the current case.



Regardless, I am not arguing about any inflation/deflation in ACF ratings. Once again, your sentence is pointless.
I disagree that my statement was pointless. remember this is not private correspondence. others are reading it.
I'm therefore just preempting any possible comments regarding inflation/deflation by either yourself or any other BB poster.



What the...? Now you're arguing my case. That is exactly my point. If you agree that inactive players aren't that strong, why gift them points? And while we're on the top 10 case (even though I started with the top 50), why is Alex Wohl so high, relative to FIDE? The answer is because he hasn't recently played in ACF-rated events. Once again, Bill, why does he get 70 points for free? (Let's not even start on numbers 9 and 10 in your list, the ratings of whom even you surely cannot defend.)
I'm not arguing your case at all.
I'm just pointing out that since rating are relative its immaterial whether players are active or not if the rating system is neither inflating or deflating. All players deserve the same increase.




Yes Bill, I do actually comprehend that ratings are relative. But once again, you have to distinguish between active and inactive players. If active players in a particular rating segment are losing points under a system - no matter how statistically valid that rating system may be internally - then it is folly to give rating points to inactive players when you want to make "corrections", i.e. realign it with the FIDE rating system.
No, you dont have to distinguish between active and inactive.
Well firstly their is no indication that players in a particular rating segment are losing points.
Secondly I showed that of the 212 active FIDE rated players the average rating movement of those players over 4 years was less than 1 rating point. Thats not some theoretical situation is based on actual numbers.




I'm honestly trying Bill. I think you should start trying a lot harder.
Sorry Greg but you making statements not based on any facts.



With my limited understanding, I believe the Glicko system may well be statistically superior to ELO - but only under perfect conditions, which may or may not be applicable to a small and widespread chess community such as Australia's. But I'm not arguing the relative merits of ELO and Glicko, just the indiscriminate "corrections" to inactive players.
I'm glad you are not going to argue the relative merist of ELO and Glicko because there is nothing to indicate that the Glicko is only superior to ELO under perfect conditions.
Their is no indiscriminate correction to inactive players.
Under the current conditions all players whether active or inactive deserve the same ratings increase.




Well yes it is different, Bill, because the FIDE list is standalone and - inflation or not - is unaffected by other systems, unlike the ACF system which is adjusted to realign with FIDE to the benefit of inactive players. (As an aside, was the Fischer-Spassky contraband match of 1991 actually FIDE-rated? If so, that would surely have contravened UN sanctions? Oh no, hang on, we're talking about FIDE.)
Its not to the benefit of inactive players.
Prior to the 70 point increase Wohl out rated Johansen, Wallace and Lane.
He deserves to still out rate them. The actual rating number is unimportant.



I don't doubt there are massive problems, the possible solutions to which are extremely complicated. I'm simply pointing out an obvious flaw.
I disagree its not a flaw. In fact to not give them the 70 points would be a flaw.



When I said a "few years back", I admit I was stretching the definition a bit. I was referring to my December 1990 rating of 2284. Nevertheless, it's disingenous of you to suggest that I would have hypothetically added 150 points to my rating twice.
Sorry but I did not expect your "few years back" to be 13.
As for being disingenous I wasnt.
Based on my belief that a few years back was in fact only a "few" the April 2000 period when your rating was 2286 seemed the one you were referring to. It therefore just appeared that you had forgotten that it was at that time that the 150 points was added.


My original point is that if I had stopped playing in early 1991, my ACF rating would now be 2504. When you consider that many of the players on the ACF top 50 list haven't played a game since the seventies, it's not so ridiculous.
And if that was the case and it was based on your 1991 rating then since it is 2504?? people would realise it would be highly unlikely that you now played at that strength.



I don't want to get into this argument about your discussions with council members as to why the FIDE system is inflationary (I agree it is anyway). Suffice it to say, you yourself replied to "your informant", as you call him, in this very bulletin board not two weeks ago - but the evidence has mysteriously disappeared...
I would have thought your informant would have to be a NSWCA council member otherwise getting information second hand is prone to be inaccurate.
Now as far as I know the only NSWCA council members who post on this BB is Matt and Kerry. Now Matt wasnt at the last Council meeting and I dont recall mentioning any of this in a reply to Kerry or anyone else on this BB in the past few weeks, so I dont know who you could be refering to.
Even if I had, there is no reason why I would have deleted the post.
Perhaps a Moderator deleted it for some reason or other, but I cannot imagine why.

Garvinator
08-03-2004, 07:51 PM
can i offer you a suggestion bill, post the url's for your comments regarding Ian Rogers comments regarding the rating system. This might help in reducing the amount of repeated information you will have to give :doh: unless you want to keep up your post count :hmm: :uhoh:

Bill Gletsos
08-03-2004, 07:53 PM
At least now Bill will be sure to work on his grammatical errors before he posts ... I'm sure you'll be keeping a sharp eye on him in that regard! :doh:
I was in a hurry to post and go home. :whistle:

Bill Gletsos
08-03-2004, 08:01 PM
grammatical errors? i believe you have your diplomatic hat on now Kezza :eh:
Of course Kerry has now set himself up to be reminded every time he makes a grammatical error. ;)

chesslover
08-03-2004, 08:55 PM
Because you goose, ratings are relative.
If player A gets X points player B should also get X points irrespective of whether they played or not.


Because you goose they have no recent results at all for any selector to determine how strong they are.

then why not go with my suggestion to the post before this?

give everyone a rating increase every time you release the ratings. It should not matter, as it is all relative but it will keep the chess players happy. My partner thinks that if I go down I am not playing well, but if it went up each time she will be happy and think I am good and let me play more. The fact that it is all relative should mean that there should be no problems with having regular 20 point increases each rating period right?

Garvinator
08-03-2004, 08:58 PM
i would just like to point out that i thought greg canfell's title to one of his posts was extremely funny :lol: :lol: :lol:

the title, edited due to bb incompetence

Rhubarb
08-03-2004, 09:29 PM
can i offer you a suggestion bill, post the url's for your comments regarding Ian Rogers comments regarding the rating system. This might help in reducing the amount of repeated information you will have to give :doh: unless you want to keep up your post count :hmm: :uhoh:

My sincerest apologies to you ggrayggray, whoever you may be. I have not had the great good fortune to anonymously and omnisciently oversee all the discussions of the rating system to this point, unlike your good self.

Please bear with me while I attempt to:

a) Recover those hallowed URLs of which you speak, such that I can aspire to your great enlightenment.

b) Engage the Great Man, Bill Gletsos, in a discussion about why the ratings of inactive players are completely whacko.

My humblest deferences,
greg

Garvinator
08-03-2004, 09:36 PM
My sincerest apologies to you ggrayggray, whoever you may be. I have not had the great good fortune to anonymously and omnisciently oversee all the discussions of the rating system to this point, unlike your good self.

Please bear with me while I attempt to:

a) Recover those hallowed URLs of which you speak, such that I can aspire to your great enlightenment.

b) Engage the Great Man, Bill Gletsos, in a discussion about why the ratings of inactive players are completely whacko.

My humblest deferences,
greg

As this is the bb and i dont know if you were being serious or just taking the piss out of me, ill take this post as a bit of a laugh, the reason being is that I was just attempting to give bill a rest so he doesnt run up more on his post count :whistle: If you took offense to my previous post, then my apologies.

Btw, yes greg there has been a large debate previously regarding the glicko2 system and most issues there of. My real name is Garvin and I am from queensland (dont hold that against me :p )

chesslover
08-03-2004, 09:38 PM
b) Engage the Great Man, Bill Gletsos, in a discussion about why the ratings of inactive players are completely whacko.

My humblest deferences,
greg

greg, forget it mate

you can never convince the GreatMan, our Supreme Leader that he is wrong in relation to rating - or for any other matter to boot.

I agree with you that the rating of inactive players are whacko, and that Bill should not add any rating adjustments to them. That just rewards a person for doing nothing, as they will get points for merely playing whilst FMs like you have to fighthard in a very competitive field to get some increase.

If they are inactive, then tough, they should get no point increases. I know thta, you know that and so do many people, but Bill will NEVER admit he is
W-R-O-N-G.

if you are lucky he might admit he is "mistaken", but that has never happened in this BB. Sometimes when I have caught him out he just says that he was just "testing" us

So that means that Bill is infalliable and is a God on earth, or he just refuses to admit when he is wrong - as he is in this instance with inactive players

skip to my lou
08-03-2004, 09:46 PM
how do you use quote tags? :confused:

i know how to reply back to the whole post of the person I am replying to, but do not know how to do replay back with my comments after each paragraph or sentence

[.quote="blahblah".]blahblah[./quote.]

without the dots

skip to my lou
08-03-2004, 09:49 PM
Thanks Jeo - big of you.

Im sure it is.

skip to my lou
08-03-2004, 09:50 PM
Rtfm

Rhubarb
08-03-2004, 09:50 PM
As this is the bb and i dont know if you were being serious or just taking the piss out of me, ill take this post as a bit of a laugh, the reason being is that I was just attempting to give bill a rest so he doesnt run up more on his post count :whistle: If you took offense to my previous post, then my apologies.

Btw, yes greg there has been a large debate previously regarding the glicko2 system and most issues there of. My real name is Garvin and I am from queensland (dont hold that against me :p )

Garvin, no offence taken and yes i was taking the piss.

You're right though, you guys shouldn't have to go through the whole thing again.

Kevin Bonham
08-03-2004, 09:50 PM
I agree with you that the rating of inactive players are whacko, and that Bill should not add any rating adjustments to them. That just rewards a person for doing nothing, as they will get points for merely playing whilst FMs like you have to fighthard in a very competitive field to get some increase.

I don't think you've made much effort to understand the issue.

Everyone got the rating adjustements, active or inactive. The rating adjustments made no difference to the rating points differences between active and inactive players - Gedevani, Levin, et al were just as overrated before the adjustments as after. No one was rewarded, in relative terms, by the 70 point increase - it's just that suddenly people have started noticing the difference between these guys' ratings and those of the active players, when they could have noticed the same difference before. (Actually some people did notice the same differences before, like Alex Wohl writing in to say he hoped if he stayed out of Australia long enough he could top the Australian rating list).

The increases were for parity with FIDE. FIDE is known to have an inflating ratings pool. There is no justification for giving the inactive players no correction, because if there is some reduction that should be applied to inactive players, it is very unlikely to be exactly 70 points. Furthermore, if you were tweaking ratings for inactivity, you would probably take more off someone who was inactive for 20 years than someone inactive for 10.


If they are inactive, then tough, they should get no point increases. I know thta, you know that and so do many people, but Bill will NEVER admit he is W-R-O-N-G.

You are in the W-R-O-N-G on this one. :hand:

Bill Gletsos
08-03-2004, 09:51 PM
a) Recover those hallowed URLs of which you speak, such that I can aspire to your great enlightenment.
I wouldn't bother Greg.
Those threads were mainly Glicko V ELO threads and are irrelevant to our discussion.



b) Engage the Great Man, Bill Gletsos, in a discussion about why the ratings of inactive players are completely whacko.
Your entitled to your opinion.
However there is no good reason why the 70 points should not be added to all players.
The excuse that they make the inactives look whacko unfortunately doesn't cut it.

Bill Gletsos
08-03-2004, 09:53 PM
greg, forget it mate

you can never convince the GreatMan, our Supreme Leader that he is wrong in relation to rating - or for any other matter to boot.

I agree with you that the rating of inactive players are whacko, and that Bill should not add any rating adjustments to them. That just rewards a person for doing nothing, as they will get points for merely playing whilst FMs like you have to fighthard in a very competitive field to get some increase.

If they are inactive, then tough, they should get no point increases. I know thta, you know that and so do many people, but Bill will NEVER admit he is
W-R-O-N-G.

if you are lucky he might admit he is "mistaken", but that has never happened in this BB. Sometimes when I have caught him out he just says that he was just "testing" us

So that means that Bill is infalliable and is a God on earth, or he just refuses to admit when he is wrong - as he is in this instance with inactive players
Do us all a favour you goose and stop posting about a subject you know nothing about.

Garvinator
08-03-2004, 09:54 PM
Do us all a favour you goose and stop posting about a subject you know nothing about.
:lol: :lol: that would leave cl with very little to post about :p

PHAT
08-03-2004, 09:55 PM
My 2 cents =

1. Rating numbers are unitless
2. Rating numbers are relative (to all others in the pool)
3. Therefore we can add or subtract to/from them mutlipy or divide them, and it does not matter one iota.
4. BG is only keeping the Glicko Australian pool comparable to the FIDE pool.

Therefore I side with BG on this.

5. The "rusty effect" has not been researched enough to assess its magnitude and acceleration.
6. BG is addressing the "speciation" that occurs in pools with large numbers of juniors.

chesslover
08-03-2004, 10:03 PM
Everyone got the rating adjustements, active or inactive. The rating adjustments made no difference to the rating points differences between active and inactive players

(Actually some people did notice the same differences before, like Alex Wohl writing in to say he hoped if he stayed out of Australia long enough he could top the Australian rating list).

The increases were for parity with FIDE. FIDE is known to have an inflating ratings pool. There is no justification for giving the inactive players no correction, because if there is some reduction that should be applied to inactive players, it is very unlikely to be exactly 70 points. Furthermore, if you were tweaking ratings for inactivity, you would probably take more off someone who was inactive for 20 years than someone inactive for 10.

You are in the W-R-O-N-G on this one. :hand:

Ok, fair point

I admit that I am not a maths whiz, and get easily confused by rating talk.

So the main problem is that whilst everyone should get the increase, some people who have been inactive far longer should get a lesser increase than those who are relatively more active?

Also if active top players are losing points by playing, and inactive players are gaining points by doing nothing (i know that all players get the same increase inactive or active) then one day you may get the inactive player top of the master list rating. Thus Smerdon for exampole could have lost 50 points in a bad tournament (net gain 20 points), whilst someone like Jammo whilst inactive could have got 70 points (net gain 70 points).

This means that if an active player plays badly, the relative difference in rating between them and the inactive player (who by definition does not play and thus will not lose points) will tilt in the favour of the inactive player

PHAT
08-03-2004, 10:03 PM
However there is no good reason why the 70 points should not be added to all players.
The excuse that they make the inactives look whacko unfortunately doesn't cut it.

But it has to cut it .... for now. Remember, it is all about relativity. EG

Three players:

2300 !!
2400 ??
2500 !

The relativity between them is 100 points
and
the 2400 is ?? unreliable

Add 70 points

2370 !!
2470 ??
2570 !

The relativity remains the same = 100 points
and
the 2470 is still ?? unreliable.

Therefore, Greg, I think your issue is not the 70 points. Your issue is the current lack of a "rust factor".

Rhubarb
08-03-2004, 10:07 PM
greg, forget it mate

you can never convince the GreatMan, our Supreme Leader that he is wrong in relation to rating - or for any other matter to boot.

I agree with you that the rating of inactive players are whacko, and that Bill should not add any rating adjustments to them. That just rewards a person for doing nothing, as they will get points for merely playing whilst FMs like you have to fighthard in a very competitive field to get some increase.

If they are inactive, then tough, they should get no point increases. I know thta, you know that and so do many people, but Bill will NEVER admit he is
W-R-O-N-G.

if you are lucky he might admit he is "mistaken", but that has never happened in this BB. Sometimes when I have caught him out he just says that he was just "testing" us

So that means that Bill is infalliable and is a God on earth, or he just refuses to admit when he is wrong - as he is in this instance with inactive players

I caught everything you said, chesslover, but I'm going to try anyway - if only because I recently read a quote from Ralph Nader (a truly Great Man):
"Turn on to politics, otherwise it'll turn on you."

Garvinator
08-03-2004, 10:09 PM
I caught everything you said, chesslover, but I'm going to try anyway - if only because I recently read a quote from Ralph Nader (a truly Great Man):
"Turn on to politics, otherwise it'll turn on you."
oh no greg big mistake, now cl will give you a lesson american politics :eek:

Rhubarb
08-03-2004, 10:16 PM
oh no greg big mistake, now cl will give you a lesson american politics :eek:
you're dead right garvin. :eek:

ok, I am only interested in the issue of over-rated players on the ACF list, not the validilty of the Elo, oo the Glicko, or (God forbid) American politics :hand:

Garvinator
08-03-2004, 10:18 PM
you're dead right garvin. :eek:

ok, I am only interested in the issue of over-rated players on the ACF list, not the validilty of the Elo, oo the Glicko, or (God forbid) American politics :hand:
actually you might be lucky cause he is trying to get himself out of a jam relating to corruption issues :lol:

chesslover
08-03-2004, 10:21 PM
I caught everything you said, chesslover, but I'm going to try anyway - if only because I recently read a quote from Ralph Nader (a truly Great Man):
"Turn on to politics, otherwise it'll turn on you."

Mate, I am very much into US presidential politics. There is a thread there on the non chess section of this BB, that has posts on this topic which you may wish to contribute to if you so wish to.

I love Nader, as I am a president Bush fan, and if Nader's votes had gone to Gore in Florida and New hampshire Bush would have lost to Gore. Gore lost 270-271 in the closest election.

This year it seems that even the Greenies in the US recongnise this, and have not endoresed nader. One of his fans, Michael Moore, also said that Nader supporters should vote strategically and plumbed for General Clark.

nader is now running as an independent and will thus poll farless than he did in 2000. All Bush has to do is to hold the states he won in 2000, and due to demographic changes he will get 278 electoral votes. Thus experts say kerry has to win Florida, New H or Ohio to wrest the presidency

I guess this is not the place to talk non topical issues such as US presidential electtions, but if you want to we can discuss further at the US Presidential elections thread

Garvinator
08-03-2004, 10:25 PM
thread police and prosecutor have been my names, now i can claim another title, pre emption king :eek:

Kevin Bonham
08-03-2004, 10:28 PM
So the main problem is that whilst everyone should get the increase, some people who have been inactive far longer should get a lesser increase than those who are relatively more active?

I prefer to think of it this way.

Everyone deserved 70 pts for parity with FIDE.

A completely seperate issue is whether rusty players should lose points.

If it is decided that they should, then it could be that some rusty players would lose somewhere between 0-70 points (which would cancel out to "getting a lesser increase"). It could be that some very rusty players would deserve to lose more than 70 points (and actually lose points overall).


Also if active top players are losing points by playing, and inactive players are gaining points by doing nothing (i know that all players get the same increase inactive or active) then one day you may get the inactive player top of the master list rating. Thus Smerdon for exampole could have lost 50 points in a bad tournament (net gain 20 points), whilst someone like Jammo whilst inactive could have got 70 points (net gain 70 points).

This is only true while the top players lose more points than they gain. At one stage early in Glicko this appeared to be the case (for whatever reason) but I think it has stopped.


This means that if an active player plays badly, the relative difference in rating between them and the inactive player (who by definition does not play and thus will not lose points) will tilt in the favour of the inactive player

That is as it should be, by and large. A player playing badly might lose 80 points in a tournament (that's about what I expect to lose for this weekend's disaster). A player not playing at all for several years would still only be declining in playing strength at a fraction of that 80 points per year.

Rhubarb
08-03-2004, 10:47 PM
I wouldn't bother Greg.
Those threads were mainly Glicko V ELO threads and are irrelevant to our discussion.


Your entitled to your opinion.
However there is no good reason why the 70 points should not be added to all players.
The excuse that they make the inactives look whacko unfortunately doesn't cut it.

Thank you for confining our discussion to inactive players. While I have (unintentionally) touched on many other important rating issues (system validity, Elo v. Glicko, compaction, isolated pools, inflation, etc.), those issues, as you say, are irrelevant to our discussion.

But if, by looking at the Top 200 players on the ACF Master File for March 2004 (I can email it to you if you like - but then again, every man and his dog can produce such an Excel file), you do not see something seriously wrong with the way inactive players are rated, then you are truly blind and have no philosophical right to be running a national rating system.


My Best Regards,
Good Night,
Greg

Bill Gletsos
08-03-2004, 10:56 PM
But if, by looking at the Top 200 players on the ACF Master File for March 2004 (I can email it to you if you like - but then again, every man and his dog can produce such an Excel file), you do not see something seriously wrong with the way inactive players are rated, then you are truly blind and have no philosophical right to be running a national rating system.
So far I have treated you with respect.
I therefore think that last comment is uncalled.



My Best Regards,
Good Night,
Greg
Your regards seem at odds with your previous sentence.

chesslover
08-03-2004, 11:01 PM
So far I have treated you with respect.
I therefore think that last comment is uncalled.


poor petal

Bill Gletsos
08-03-2004, 11:03 PM
Greg,

Generally the following situation is possible.
Two rating pool exists A and B
Over time one pool becomes out of sync with the other rating pool.
Three possible scenarios from amongst a number are:
1) Pool A neither deflates nor inflates. Pool B inflates. Lets say the difference between Pool A and be is found to be 70 points. Then all players in A gets a 70 point increase irrespective of activity because the ratings are all relative.

2) Pool A deflates over time and pool B inflates. Now lets say the difference between pool A and B is 150 points. Now this 150 points represents the current difference, however the make up of that difference is that for every year pool A deflated 10 points (this is overly simplistic and it is unlikely to be a straight line as being used here). Therefore if a player in pool A last played 2 years ago he should only get 130 points, if he last played 5 years ago he only gets 100 points and if he last played 10 years ago he only gets 50 points.

3) It could be that pool A is inflating and Pool B is inflating. Now lets say the difference between pool A and B is 150 points. In this case you might find that pool A inflates at 10 points a year therefore inactive players in Pool A are at a disadvantage compared to active players. There ois thus 2 ways to handle this.
a) Ass an extra 10 points for each year they were inactive. Thus all players get a minimum of 150 but if a player was inactive for by years he gets 200 and if inactive for 10 years he gets 2050.
b) deduct 10 points from active players for every year they were active to keep them in line with the inactive players then add 150 to each player. Hence this equates to inactives getting 150 points but a player who has been active for the past 5 years only gets 100 and if he has been active for the past 10 years he only gets 50 points.


Now even those situations in 2) and 3) are not that simple to resolve as it might seem from the above. This is because ideally the correction of the inflation in pool A should have occurred on a rating period by rating period basis at the time. This means then that as a player becomes inactive his rating was in line with others in the pool at the time of his inactivity and that those active in the pool would have had their ratings adjusted for the inflation. This adjustment of the active players ratings at the time would havve effected their rating changes for the subsequent period and so on and so forth. Therefore in 2) above a player who has been active may with the calculation done at the end of the sequence end up having 140 points added to their current rating, but if it had been done on a rating period by rating period basis then his increase compared to his current rating might only be 127 points.

Now on top of all that there is the other issue of should inactive players lose rating points for simply being inactive.


It was with all these issues in mind that the ACF Rating Officers got the ACF Council at the September 2003 Council meeting to pass not only the motion that allowed us to try and keep the ACF and FIDE ratings inline but also get passed the 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

That last motion is fairly all encompassing in allowing for corrections.

The problem is Greg, it isnt exactly easy to determine the correct path to follow to cater for the above described situations.

However not knowing how to implement said corrections at this time is no reason not to give everyone the increase, because after all no one should put much faith in ? or ?? ratings for inactive players.

Bill Gletsos
08-03-2004, 11:10 PM
poor petal
Hey dipshit,
I just said it was uncalled for.
I did not say it worried me or traumatised me, etc. :whistle:
I'm not some emotionally fragile little petal like you CL. :rolleyes:

So go back to bleating about being unfairly done by with regards the prize for best BB post you moron. :hand:

Garvinator
08-03-2004, 11:14 PM
Hey dipshit,
im still trying to work out how these ruuuddeee words are getting passed the censor :whistle:


So go back to bleating about being unfairly done by with regards the prize for best BB post you moron. :hand:
so go back to bleating about being ALLEGEDLY unfairly done by with regards the prize for best BB post you moron :hand: . It should have an allegedly in it cause cl has not proven yet that he has been jipped enough to get himself in a tie for first with starter.

Rhubarb
08-03-2004, 11:54 PM
Greg,

Generally the following situation is possible.
Two rating pool exists A and B
Over time one pool becomes out of sync with the other rating pool.
Three possible scenarios from amongst a number are:
1) Pool A neither deflates nor inflates. Pool B inflates. Lets say the difference between Pool A and be is found to be 70 points. Then all players in A gets a 70 point increase irrespective of activity because the ratings are all relative.

2) Pool A deflates over time and pool B inflates. Now lets say the difference between pool A and B is 150 points. Now this 150 points represents the current difference, however the make up of that difference is that for every year pool A deflated 10 points (this is overly simplistic and it is unlikely to be a straight line as being used here). Therefore if a player in pool A last played 2 years ago he should only get 130 points, if he last played 5 years ago he only gets 100 points and if he last played 10 years ago he only gets 50 points.

3) It could be that pool A is inflating and Pool B is inflating. Now lets say the difference between pool A and B is 150 points. In this case you might find that pool A inflates at 10 points a year therefore inactive players in Pool A are at a disadvantage compared to active players. There ois thus 2 ways to handle this.
a) Ass an extra 10 points for each year they were inactive. Thus all players get a minimum of 150 but if a player was inactive for by years he gets 200 and if inactive for 10 years he gets 2050.
b) deduct 10 points from active players for every year they were active to keep them in line with the inactive players then add 150 to each player. Hence this equates to inactives getting 150 points but a player who has been active for the past 5 years only gets 100 and if he has been active for the past 10 years he only gets 50 points.


Now even those situations in 2) and 3) are not that simple to resolve as it might seem from the above. This is because ideally the correction of the inflation in pool A should have occurred on a rating period by rating period basis at the time. This means then that as a player becomes inactive his rating was in line with others in the pool at the time of his inactivity and that those active in the pool would have had their ratings adjusted for the inflation. This adjustment of the active players ratings at the time would havve effected their rating changes for the subsequent period and so on and so forth. Therefore in 2) above a player who has been active may with the calculation done at the end of the sequence end up having 140 points added to their current rating, but if it had been done on a rating period by rating period basis then his increase compared to his current rating might only be 127 points.

Now on top of all that there is the other issue of should inactive players lose rating points for simply being inactive.


It was with all these issues in mind that the ACF Rating Officers got the ACF Council at the September 2003 Council meeting to pass not only the motion that allowed us to try and keep the ACF and FIDE ratings inline but also get passed the motion:


That last motion is fairly all encompassing in allowing for corrections.

The problem is Greg, it isnt exactly easy to determine the correct path to follow to cater for the above described situations.

However not knowing how to implement said corrections at this time is no reason not to give everyone the increase, because after all no one should put much faith in ? or ?? ratings for inactive players.


Bill,

Tomorrow is another day and I will attempt to understand the above arguments then.

You are right in that you treated me with nothing but respect in the previous exchanges, and that I inexcusably failed to do so in my previous post.

Please accept my apologies,

Greg

Bill Gletsos
08-03-2004, 11:57 PM
Bill,

Tomorrow is another day and I will attempt to understand the above arguments then.

You are right in that you treated me with nothing but respect in the previous exchanges, and that I inexcusably failed to do so in my previous email.

Please accept my apologies,

Greg
Greg,
Apology accepted.
I look forward to continuing the debate with you.

Regards,
Bill

Ian_Rogers
09-03-2004, 02:10 AM
There is a precedent for adjusting some players ratings and not others, just to keep the list sensible. When everyone was given 150 points, I wasn't, a decision with which I fully agreed. The decision was arbitrary but sensible.
There is no harm in making the top 50 list more sensible by giving inactive players less than 70 points bonus (and totally inactive players much less than 220 points) - if the inactive players feel hard done by not getting 70 points they can always start playing and take the points from players who have received the bonuses.

P.S. I don't know of anyone who suggests that the FIDE list has inflated by as much as 70 points in just the last few years. Therefore it would seem highly likely that the Australian list is also deflating in comparison with the FIDE list- Bill's scenario B.

arosar
09-03-2004, 08:33 AM
I only have rudimentary maths/stats skills but even I can appreciate that condition 'X' for N1 player has to be true for N2 player right through to NN player. I understand that this fandangled new system, Glicko2, is supposed to be very sensitive - so I'd expect it to just do its job once this 'inflated' players start playing. FMD, I've ever even heard of a coupla those players with funny names.

AR

Bill Gletsos
09-03-2004, 10:35 AM
There is a precedent for adjusting some players ratings and not others, just to keep the list sensible. When everyone was given 150 points, I wasn't, a decision with which I fully agreed. The decision was arbitrary but sensible.
Hi Ian,
All of thats true, however Graham and I also felt at the time that your ACF rating was unduly inflated because of the 336 rule that had been in place.



There is no harm in making the top 50 list more sensible by giving inactive players less than 70 points bonus (and totally inactive players much less than 220 points) - if the inactive players feel hard done by not getting 70 points they can always start playing and take the points from players who have received the bonuses.
As I said in my response to Greg above, Graham and I are looking at it.



P.S. I don't know of anyone who suggests that the FIDE list has inflated by as much as 70 points in just the last few years. Therefore it would seem highly likely that the Australian list is also deflating in comparison with the FIDE list- Bill's scenario B.
I dont think I said the FIDE list had inflated by 70 points
What I said was:

Given the influx of many more Australians to the FIDE list over this 4 year period this has lead to an effective inflation trend in the FIDE ratings of Australian players especially those players below FM strength. It is from here that the 70 points is coming.
Therefore in relation to Australian players their FIDE ratings appear to have inflated by 70 points.
For the whole of the FIDE list I would expect it to be much less.
I know that for the top 20, top 50 and top 100 of the whole of the FIDE list between Jan 2000 and Jan 2004, the average FIDE ratings incresed by 12.30, 11.82 and 11.76 respectively.
Graham and I can find no evidence that the ACF list is deflating or inflating to any significant degree.

Garvinator
09-03-2004, 10:38 AM
how could you tell that the list is inflating. i would take inflation in this context to me that the general chess populations ratings are increasing, but their playing standard is remaining the same. Is this correct?

Rincewind
09-03-2004, 12:40 PM
I only have rudimentary maths/stats skills but even I can appreciate that condition 'X' for N1 player has to be true for N2 player right through to NN player. I understand that this fandangled new system, Glicko2, is supposed to be very sensitive - so I'd expect it to just do its job once this 'inflated' players start playing. FMD, I've ever even heard of a coupla those players with funny names.

That ain't necessarily so. The mathematics as applied to ratings make assumptions as to what are reasonable for the case in point. Stats isn't my strong suit but I think the whole logistical curve thing is based on the assumption that you are dealing with a behaviour which can be expressed in the form x' = ax^2 + bx, which might make a good approximation but I don't think anyone is suggesting it is exact.

So if there are reasons to treat particular sets of players differently to others then there is nothing intrinsically invalid with that method, even if one set contains only 1 player. Although I think the more selective the set the more closely the reason to treat that set differently should be examined.

peter_parr
09-03-2004, 02:12 PM
From the ACF Master file top 20 March 2004

ACF FIDE Diff
1 2668 !! Rogers, Ian NSW 2582 86+
2 2567 ?? Levin, Naum L NSW 2420 147+
3 2561 Wohl, Aleksander H QLD 2398 163+
4 2548 ? Jamieson, Robert M VIC 2444 104+
5 2538 ? Gedevani, Dimitri NSW 2380 158+
6 2509 !! Johansen, Darryl K VIC 2519 10-
7 2505 ! Wallace, John-Paul NSW 2410 95+
8 2493 !! Lane, Gary W NSW 2423 70+
9 2491 ?? Travers, Roy B NSW 2205 286+
10 2488 ?? Kaltenbacher, Sorin NSW - -
11 2468 ? Davidovic, Aleksanda SA 2392 76+
12 2468 ? Prods, Arvids VIC 2205 263+
13 2464 ? Kanikevich, Alexande NSW 2410 54+
14 2458 ! Gluzman, Michael VIC 2403 55+
15 2456 ?? Brodie, Alaistar I NSW decd -
16 2454 ? Kontorovich, Michael VIC 2385 69+
17 2448 ? Goldenberg, Igor VIC 2335 113+
18 2446 !! Solomon, Stephen J QLD 2397 49+
19 2439 ! Tao, Trevor SA 2390 49+
20 2436 ? Belin, Igor NSW 2319 117+

Top 20, 1 decd, 1 not rated FIDE, 12 - 70 or more, 5 - 49 or more, 1 - lower

70 ACF rating points were added to each player to bring the ACF ratings in line with FIDE but the top 20 are now over 100 points (average) above FIDE.

Note from the top 20 ACF - 12 are over 70 rating points above their FIDE rating, another 5 more than 48 ratings points above their FIDE rating. From the other 3 one has no FIDE rating, one is deceased and only one of the top 20 Johansen is 10 rating points behind his FIDE rating (He was "Glickoed" from the Australian Champs).

So from the top 18 ACF players with FIDE ratings they are a total of 1954 rating points (average 108 rating points higher per player) higher on the ACF list. Rogers (currently in Iceland) asked if he could deduct the extra 70 points from his ACF rating but this was NOT PERMITTED.

Rogers standard of play is 2582 not 2668
Wohl standard of play is 2398 not 2561 etc.

Regards
Peter Parr

Bill Gletsos
09-03-2004, 03:43 PM
[FONT=Courier New]From the ACF Master file top 20 March 2004

ACF FIDE Diff
1 2668 !! Rogers, Ian NSW 2582 86+
2 2567 ?? Levin, Naum L NSW 2420 147+
3 2561 Wohl, Aleksander H QLD 2398 163+
4 2548 ? Jamieson, Robert M VIC 2444 104+
5 2538 ? Gedevani, Dimitri NSW 2380 158+
6 2509 !! Johansen, Darryl K VIC 2519 10-
7 2505 ! Wallace, John-Paul NSW 2410 95+
8 2493 !! Lane, Gary W NSW 2423 70+
9 2491 ?? Travers, Roy B NSW 2205 286+
10 2488 ?? Kaltenbacher, Sorin NSW - -
11 2468 ? Davidovic, Aleksanda SA 2392 76+
12 2468 ? Prods, Arvids VIC 2205 263+
13 2464 ? Kanikevich, Alexande NSW 2410 54+
14 2458 ! Gluzman, Michael VIC 2403 55+
15 2456 ?? Brodie, Alaistar I NSW decd -
16 2454 ? Kontorovich, Michael VIC 2385 69+
17 2448 ? Goldenberg, Igor VIC 2335 113+
18 2446 !! Solomon, Stephen J QLD 2397 49+
19 2439 ! Tao, Trevor SA 2390 49+
20 2436 ? Belin, Igor NSW 2319 117+

Top 20, 1 decd, 1 not rated FIDE, 12 - 70 or more, 5 - 49 or more, 1 - lower

70 ACF rating points were added to each player to bring the ACF ratings in line with FIDE but the top 20 are now over 100 points (average) above FIDE.

Note from the top 20 ACF - 12 are over 70 rating points above their FIDE rating, another 5 more than 48 ratings points above their FIDE rating. From the other 3 one has no FIDE rating, one is deceased and only one of the top 20 Johansen is 10 rating points behind his FIDE rating (He was "Glickoed" from the Australian Champs).
Firstly let me say, I find it interesting you did not raise or debate this issue with me when I saw you last week. After all it was at that time I saw and commented the list on your wall that Greg posted here yesterday.

Now, it turns out that unfortunately for you its pretty useless when comparing what the uplift if any should be to include inactive players on either the FIDE list or the ACF list.
After all the aim of the whole thing is to determine how out of alignment the ACF ratings are with regards to FIDE for those who are actually active.

Now as I noted in an earlier post:

2) When it came to determining the uplift we only looked at players who we considered active on the Dec 2003 ACF list and the Jan 2004 FIDE list. We also ignored all players whose ACF ratings were below 2050. The true uplift was 68 points but we decided to give everyone 70.

For those rated over 2300 on the ACF list the difference between their ACF and FIDE ratings is on average 29 points. For all those ACF rated over 2200 the average difference is 41 points and for all those ACF rated over 2100 its 58 points.


As for johansen being "Glickoed" its just a silly statement.
If the old ACF ELO system was still in place Darryl would have been rated 2510. :hand:
Yep, it sure likes like he was "Glickoed". :doh:



[So from the top 18 ACF players with FIDE ratings they are a total of 1954 rating points (average 108 rating points higher per player) higher on the ACF list. Rogers (currently in Iceland) asked if he could deduct the extra 70 points from his ACF rating but this was NOT PERMITTED.
Lets get something straight here, we dont change players ratings just because they request it.
As I noted in a previous post in this thread:

Ian may be embarressed by his rating but Graham Saint and I feel it is entirely justified. If no one had received the 70 point increase he would have still been nearly 160 points above his nearest rival.
Ian must therefore gain whatever points everyone else gains to maintain parity.


[Rogers standard of play is 2582 not 2668
Wohl standard of play is 2398 not 2561 etc.
Actually Rogers standard of play is 160 points above Johansen, thats all.

Kevin Bonham
09-03-2004, 04:17 PM
I am going to do a little resorting of Peter's list because there is something he is overlooking here. I'm sorting it into those with ! and !! ratings, and the rest. So:

GROUP 1

1 2668 !! Rogers, Ian NSW 2582 86+
6 2509 !! Johansen, Darryl K VIC 2519 10-
7 2505 ! Wallace, John-Paul NSW 2410 95+
8 2493 !! Lane, Gary W NSW 2423 70+
14 2458 ! Gluzman, Michael VIC 2403 55+
18 2446 !! Solomon, Stephen J QLD 2397 49+
19 2439 ! Tao, Trevor SA 2390 49+

GROUP 2 (FIDE unrateds or deceased removed)

2 2567 ?? Levin, Naum L NSW 2420 147+
3 2561 Wohl, Aleksander H QLD 2398 163+
4 2548 ? Jamieson, Robert M VIC 2444 104+
5 2538 ? Gedevani, Dimitri NSW 2380 158+
9 2491 ?? Travers, Roy B NSW 2205 286+
11 2468 ? Davidovic, Aleksanda SA 2392 76+
12 2468 ? Prods, Arvids VIC 2205 263+
13 2464 ? Kanikevich, Alexande NSW 2410 54+
16 2454 ? Kontorovich, Michael VIC 2385 69+
17 2448 ? Goldenberg, Igor VIC 2335 113+
20 2436 ? Belin, Igor NSW 2319 117+

The mean ACF-FIDE difference for group 1 is 56.3 points. For group 2 it is 140.9 points. So this statement:


70 ACF rating points were added to each player to bring the ACF ratings in line with FIDE but the top 20 are now over 100 points (average) above FIDE.

simplifies things because a large part of the problem is the way that inactive players have become overrated, through past deflation, recent compaction or other sources of points losses at the top end.

As for the 56 points, this difference partly exists because Glicko is a more dynamic system, an approach that is justified (as I understand it) by better predictive capacity than less dynamic systems. Rogers is rated as highly as he is because during the last year he performed way above his rating in a very large sample of Australian tournament games. FIDE ratings are less responsive and would not register this to such a degree, but when a player performs above their rating over so many games it is significant. Wallace built a high ACF rating with a series of outstanding results in tournaments in NSW before leaving. Solomon, Tao and Lane are coming off recent excellent results which Glicko is more sensitive to than FIDE. Players might feel that the less dynamic FIDE system more accurately describes their standard of play at the time, but does this actually stand up to mathematical scrutiny? Apparently not.

Also measuring the difference between one list and another by taking the top players off only one list is invalid and exaggerates the problem. Even if the two samples have the same average, such a method will tell you that they don't. You need to look at the top players off both lists. For instance the top 7 on the FIDE list among the ACF-active players include three whose ACF is lower than their FIDE (Johansen, Smerdon and Zhao). If you take that set of 7 players the mean difference for the top 7 ACF-actives is only 34 points.

Finally, that 34 or 56 point difference disappears if we look at a fuller set of top players and not a miserably small sample. I'm including everyone over 2250 who is ACF and FIDE rated and ! or !!.

Rogers(+86), Johansen(-10), Lane(+70), Wallace(+95), Gluzman(+55), Solomon(+49), Tao(+49) as above

ACF, FIDE, ACF-FIDE
Smerdon 2380/2418 -38
Zhao 2380/2398 -18
Sandler 2390/2371 +19
Depasquale 2359/2354 +5
West 2383/2352 +31
Chapman 2381/2349 +32
Feldman 2272/2327 -55
Canfell 2284/2310 -26
Baron 2280/2303 -23
Barber 2159/2299 -140
Reilly 2271/2291 -20
Tindall 2301/2286 +15
Berezina-Feldman 2245/2281 -36
Hamilton 2238/2279 -41
Rujevic 2276/2273 +3
Boyd 2243/2271 -28
Levi 2295/2263 +32
Tan 2234/2258 -24
Xie 2297/2257 +40
Woodhams 2141/2254 -113
Samar 2139/2254 -115
Froehlich 2400/2397 -3
Bjelobrk 2378/2338 +40
Teichmann 2366/2320 +44
Rashid 2274/2296 -22
Smirnov omitted as there are two Vladimir Smirnovs on FIDE list.

16 active players are rated higher on ACF by total 665 points, 16 lower on ACF by total 712 points. Over the 32 players listed the mean difference is that the FIDE rating is 1.46 points. This is trivially small therefore there is parity between FIDE and ACF ratings for players over 2250 and active on the ACF list.


Rogers (currently in Iceland) asked if he could deduct the extra 70 points from his ACF rating but this was NOT PERMITTED.

Well of course not. The rating system doesn't measure how good players think they are, it measures the objective evidence of how well they are performing. And while he is performing so well, any player who beats him deserves to be fully rewarded, not short-changed.


Rogers standard of play is 2582 not 2668

I disagree totally. His Aussie performances in the past 12 months fully justify his rating. (However, a note: his performance in Aus may be inflated partly by knowing his opponents' games and styles here well. This does not prove he is a 2668-strength player at world level yet.)


Wohl standard of play is 2398 not 2561 etc.

Possibly true in Wohl's case because his FIDE rating has had a great deal of fresh data, while his ACF is way out of date.

Bill Gletsos
09-03-2004, 05:20 PM
Here is all players rated under the Glicko2 system rated 2200 and over who have been active since 1st June 2002 irrespective of rating code.


2668!! 23 NSW Rogers, Ian
2509!! 11 VIC Johansen, Darryl K
2505! 0 NSW Wallace, John-Paul
2493!! 18 NSW Lane, Gary W
2458! 0 VIC Gluzman, Michael
2446!! 18 QLD Solomon, Stephen J
2439! 11 SA Tao, Trevor
2422 0 VIC Speck, Nicholas S
2400!! 0 VIC Froehlich, Peter
2393 0 VIC Jordan, Bill
2383! 0 VIC West, Guy
2381!! 11 SA Chapman, Mark
2380!! 11 VIC Smerdon, David C
2380!! 17 NSW Zhao, Zong-Yuan
2378!! 25 VIC Bjelobrk, Igor
2371!! 14 VIC Sandler, Leonid
2366! 0 VIC Teichmann, Erik
2360 7 VIC Rutherford, Simon A
2359!! 25 VIC Depasquale, Chris J
2348 7 NSW Curtis, John
2335 0 NSW Stephens, Malcolm
2328 0 VIC Kagan, Naum
2318 0 NSW Drummond, Matthew
2316 0 QLD Laird, Craig
2301! 0 NSW Tindall, Brett
2297!! 37 NSW Xie, George
2295!! 11 VIC Levi, Eddy L
2289?? 0 NSW Hendry, John
2284!! 18 NSW Canfell, Gregory J
2280!! 26 VIC Baron, Michael
2276!! 7 VIC Rujevic, Mirko
2275 0 NSW Dauvergne, Peter
2274!! 13 VIC Rashid, Abdulwahab
2272! 0 NSW Feldman, Vladimir
2271!! 11 NSW Reilly, Tim
2269 0 QLD Allen, Andrew P
2259! 0 QLD Stephson, David J
2252! 0 NSW Cook, Roger S
2252! 0 NSW Smirnov, Vladimir
2245!! 8 NSW Berezina - Feldman, Irina
2245! 7 NSW Flatow, A (Fred)
2244! 6 NSW Kabir, Ruhul
2243!! 26 WA Boyd, Tristan
2241!! 10 SA Zaric, Srboljub
2238!! 0 VIC Hamilton, Douglas G
2236!! 6 NSW Seberry, Ralph B
2234!! 30 NSW Tan, Justin
2219 0 SA Koshnitsky, Ngan
2213 0 QLD Pizzato, Charles A
2212 0 VIC Saw, Geoffrey
2204!! 0 QLD Sorokina, Anastasia
2202 0 NSW Weeks, Manuel W
2201! 0 NSW Bouchaaya, Tony

Now here are all players rated under the old ACF ELO system rated 2200 and over who have been active since 1st June 2002.

2660 23 NSW Rogers, Ian [GM]
2510 11 VIC Johansen, Darryl K [GM]
2507 0 NSW Wallace, John-Paul [IM]
2492 18 NSW Lane, Gary W [IM]
2458 0 VIC Gluzman, Michael [IM]
2445 18 QLD Solomon, Stephen J [IM]
2437 11 SA Tao, Trevor
2423 0 VIC Speck, Nicholas S
2387 14 VIC Sandler, Leonid [IM]
2383 0 VIC Jordan, Bill [FM]
2383 17 NSW Zhao, Zong-Yuan [IM]
2380 0 VIC West, Guy [IM]
2379 11 VIC Smerdon, David C [IM]
2379 11 SA Chapman, Mark [IM]
2377 0 QLD Laird, Craig [FM]
2375 7 NSW Curtis, John [FM]
2374 25 VIC Bjelobrk, Igor
2367 0 VIC Teichmann, Erik [FM]
2357 0 VIC Froehlich, Peter [IM]
2354 0 NSW Drummond, Matthew
2352 25 VIC Depasquale, Chris J [FM]
2346 7 VIC Rutherford, Simon A [FM]
2346 0 VIC Kagan, Naum [FM]
2343 0 NSW Feldman, Vladimir [IM]
2336 0 NSW Ayvazyan, Armen
2334 0 NSW Stephens, Malcolm [FM]
2308 0 NSW Cook, Roger S
2300 0 NSW Tindall, Brett [FM]
2298 26 VIC Baron, Michael [FM]
2297 37 NSW Xie, George
2295 11 VIC Levi, Eddy L [FM]
2284 11 NSW Reilly, Tim [FM]
2283 0 QLD Allen, Andrew P [FM]
2277 18 NSW Canfell, Gregory J [FM]
2271 7 NSW Flatow, A (Fred) [FM]
2266 7 VIC Rujevic, Mirko [FM]
2265 19 WA Barber, Haydn J [FM]
2265 6 NSW Seberry, Ralph B
2262 0 QLD Stephson, David J
2262 6 NSW Kabir, Ruhul
2260 0 NSW Bouchaaya, Tony
2258 0 NSW Dauvergne, Peter
2258 13 VIC Rashid, Abdulwahab [FM]
2250 0 VIC Delion, Leon
2248 0 VIC Hamilton, Douglas G [FM]
2243 26 WA Boyd, Tristan
2238 8 NSW Berezina - Feldman, Irina [IM]
2235 10 SA Zaric, Srboljub
2227 0 VIC Saw, Geoffrey [FM]
2218 0 NSW Vojvodic, Branislev
2214 0 NSW Smirnov, Vladimir
2213 0 NSW Jakirlic, Nermin
2213 0 QLD Pizzato, Charles A
2208 30 NSW Tan, Justin
2207 0 NSW Wettstein, Marcus
2204 7 NSW Scott, Ronald
2204 0 NSW Weeks, Manuel W [FM]
2204 11 VIC Booth, Stewart
2200 0 QLD Sorokina, Anastasia [WIM]

Cat
09-03-2004, 10:23 PM
Here is all players rated under the Glicko2 system rated 2200 and over who have been active since 1st June 2002 irrespective of rating code.


2668!! 43 NSW Rogers, Ian
2509!! 44 VIC Johansen, Darryl K
2505! 27 NSW Wallace, John-Paul
2493!! 39 NSW Lane, Gary W
2458! 36 VIC Gluzman, Michael
2446!! 40 QLD Solomon, Stephen J
2439! 26 SA Tao, Trevor
2422 30 VIC Speck, Nicholas S
2400!! 104 VIC Froehlich, Peter
2393 46 VIC Jordan, Bill
2383! 45 VIC West, Guy
2381!! 40 SA Chapman, Mark
2380!! 19 VIC Smerdon, David C
2380!! 17 NSW Zhao, Zong-Yuan
2378!! 21 VIC Bjelobrk, Igor
2371!! 41 VIC Sandler, Leonid
2366! 42 VIC Teichmann, Erik
2360 28 VIC Rutherford, Simon A
2359!! 42 VIC Depasquale, Chris J
2348 51 NSW Curtis, John
2335 32 NSW Stephens, Malcolm
2328 57 VIC Kagan, Naum
2318 38 NSW Drummond, Matthew
2316 50 QLD Laird, Craig
2301! 26 NSW Tindall, Brett
2297!! 18 NSW Xie, George
2295!! 51 VIC Levi, Eddy L
2289?? 104 NSW Hendry, John
2284!! 33 NSW Canfell, Gregory J
2280!! 27 VIC Baron, Michael
2276!! 104 VIC Rujevic, Mirko
2275 38 NSW Dauvergne, Peter
2274!! 104 VIC Rashid, Abdulwahab
2272! 47 NSW Feldman, Vladimir
2271!! 36 NSW Reilly, Tim
2269 31 QLD Allen, Andrew P
2259! 104 QLD Stephson, David J
2252! 104 NSW Cook, Roger S
2252! 104 NSW Smirnov, Vladimir
2245!! 38 NSW Berezina - Feldman, Irina
2245! 66 NSW Flatow, A (Fred)
2244! 104 NSW Kabir, Ruhul
2243!! 19 WA Boyd, Tristan
2241!! 63 SA Zaric, Srboljub
2238!! 62 VIC Hamilton, Douglas G
2236!! 37 NSW Seberry, Ralph B
2234!! 18 NSW Tan, Justin
2219 104 SA Koshnitsky, Ngan
2213 23 QLD Pizzato, Charles A
2212 22 VIC Saw, Geoffrey
2204!! 23 QLD Sorokina, Anastasia
2202 37 NSW Weeks, Manuel W
2201! 104 NSW Bouchaaya, Tony

Now here are all players rated under the old ACF ELO system rated 2200 and over who have been active since 1st June 2002.

2660 23 NSW Rogers, Ian [GM]
2510 11 VIC Johansen, Darryl K [GM]
2507 0 NSW Wallace, John-Paul [IM]
2492 18 NSW Lane, Gary W [IM]
2458 0 VIC Gluzman, Michael [IM]
2445 18 QLD Solomon, Stephen J [IM]
2437 11 SA Tao, Trevor
2423 0 VIC Speck, Nicholas S
2387 14 VIC Sandler, Leonid [IM]
2383 0 VIC Jordan, Bill [FM]
2383 17 NSW Zhao, Zong-Yuan [IM]
2380 0 VIC West, Guy [IM]
2379 11 VIC Smerdon, David C [IM]
2379 11 SA Chapman, Mark [IM]
2377 0 QLD Laird, Craig [FM]
2375 7 NSW Curtis, John [FM]
2374 25 VIC Bjelobrk, Igor
2367 0 VIC Teichmann, Erik [FM]
2357 0 VIC Froehlich, Peter [IM]
2354 0 NSW Drummond, Matthew
2352 25 VIC Depasquale, Chris J [FM]
2346 7 VIC Rutherford, Simon A [FM]
2346 0 VIC Kagan, Naum [FM]
2343 0 NSW Feldman, Vladimir [IM]
2336 0 NSW Ayvazyan, Armen
2334 0 NSW Stephens, Malcolm [FM]
2308 0 NSW Cook, Roger S
2300 0 NSW Tindall, Brett [FM]
2298 26 VIC Baron, Michael [FM]
2297 37 NSW Xie, George
2295 11 VIC Levi, Eddy L [FM]
2284 11 NSW Reilly, Tim [FM]
2283 0 QLD Allen, Andrew P [FM]
2277 18 NSW Canfell, Gregory J [FM]
2271 7 NSW Flatow, A (Fred) [FM]
2266 7 VIC Rujevic, Mirko [FM]
2265 19 WA Barber, Haydn J [FM]
2265 6 NSW Seberry, Ralph B
2262 0 QLD Stephson, David J
2262 6 NSW Kabir, Ruhul
2260 0 NSW Bouchaaya, Tony
2258 0 NSW Dauvergne, Peter
2258 13 VIC Rashid, Abdulwahab [FM]
2250 0 VIC Delion, Leon
2248 0 VIC Hamilton, Douglas G [FM]
2243 26 WA Boyd, Tristan
2238 8 NSW Berezina - Feldman, Irina [IM]
2235 10 SA Zaric, Srboljub
2227 0 VIC Saw, Geoffrey [FM]
2218 0 NSW Vojvodic, Branislev
2214 0 NSW Smirnov, Vladimir
2213 0 NSW Jakirlic, Nermin
2213 0 QLD Pizzato, Charles A
2208 30 NSW Tan, Justin
2207 0 NSW Wettstein, Marcus
2204 7 NSW Scott, Ronald
2204 0 NSW Weeks, Manuel W [FM]
2204 11 VIC Booth, Stewart
2200 0 QLD Sorokina, Anastasia [WIM]

Are you saying these would be their current ELO ratings? If so they are remarkably similar!

Bill Gletsos
09-03-2004, 10:24 PM
Are you saying these would be their current ELO ratings? If so they are remarkably similar!
I thought I made it clear what I was saying.

Cat
09-03-2004, 10:35 PM
So where's the innate superiority of Glicko over ELO? If they have essentially produced almost identical results, why have placed your entire argument about ratings on the strength of the Glicko to the point where it has suffocated all intellegent debate. It seems to me you are trying to have it both ways!

Bill Gletsos
09-03-2004, 10:43 PM
So where's the innate superiority of Glicko over ELO? If they have essentially produced almost identical results, why have placed your entire argument about ratings on the strength of the Glicko to the point where it has suffocated all intellegent debate. It seems to me you are trying to have it both ways!
Given the praise CL has heaped upon you in another thread as being so wise and intelligent, I would have thought you would have been the first to realise that a rating system has to cater to all types, not just the elite. :hand:

Guess not. :whistle:

Cat
09-03-2004, 11:01 PM
[QUOTE=Bill Gletsos]Given the praise CL has heaped upon you in another thread as being so wise and intelligent, I would have thought you would have been the first to realise that a rating system has to cater to all types, not just the elite. :hand:

QUOTE]

Too true, but these results are almost identical! Strikingly so! Sheer co-incidence or the hand of God?

Bill Gletsos
09-03-2004, 11:07 PM
Too true, but these results are almost identical! Strikingly so! Sheer co-incidence or the hand of God?
Perhaps if you thought about it for a minute(or maybe you would need longer) you could understand why they are likely to be similar.

As for the last part I wont dignify that with a response. :mad:

Kevin Bonham
09-03-2004, 11:13 PM
So where's the innate superiority of Glicko over ELO? If they have essentially produced almost identical results,

Only at the top level where the top-player tweak operates. Elsewhere there would be more differences.

Bill Gletsos
09-03-2004, 11:22 PM
Only at the top level where the top-player tweak operates. Elsewhere there would be more differences.
Ah, you make it too easy for him.
I wanted to see if he could work it out for himself.

Cat
09-03-2004, 11:26 PM
So you are actually admitting that the Glicko results for over 2200 have been manipulated to be brought into line with their expected ELO performance?

chesslover
09-03-2004, 11:28 PM
Ah, you make it too easy for him.
I wanted to see if he could work it out for himself.

I can assure you 100% that David has the brains to work out anything and everything that you throw at him. This is proven by the fact that he is a Doctor, unlike us all.

In all posts that I have read of him, I have walked away with a great deal of knowledge as a result of reading David's post.

I resent you making these snide remarks that he is not smart enough to work that out

chesslover
09-03-2004, 11:32 PM
So you are actually admitting that the Glicko results for over 2200 have been manipulated to be brought into line with their expected ELO performance?

well done :clap: :clap:

I think you got him. almost a knock out blow. The emperor has no cloths, when it comes to inactive players and the rating adjustment for them!!!!

I think Greg Canfell was the first in this thread to spot it, and now even our greatest player Ian Rogers and you have shown the glaring weakness of Bill's practice

The question is he, will he admit this fault, and try to work out teh solution or will he still try to defend the indefensible?

Bill Gletsos
09-03-2004, 11:33 PM
I can assure you 100% that David has the brains to work out anything and everything that you throw at him. This is proven by the fact that he is a Doctor, unlike us all.
That affords him respect with regards Medical matters, nothing more.

As for your assurances, well I take them with a grain of salt.


In all posts that I have read of him, I have walked away with a great deal of knowledge as a result of reading David's post.
Good for you.


I resent you making these snide remarks that he is not smart enough to work that out
Hey dipshit what about his snide remarks about hand of god.

Go back to carrying on about BB post prizes you goose.

Bill Gletsos
09-03-2004, 11:35 PM
So you are actually admitting that the Glicko results for over 2200 have been manipulated to be brought into line with their expected ELO performance?
You really are stupid arent you.

The whole situation with regards the change made for ratings over 2200 was explained at length on the old BB.

It had nothing to do bring them into line with ELo performances then and it still has nothing to do with them.

Bill Gletsos
09-03-2004, 11:40 PM
well done :clap: :clap:

I think you got him. almost a knock out blow. The emperor has no cloths, when it comes to inactive players and the rating adjustment for them!!!!
Pull your head out of his ass CL, its not a pretty sight.


I think Greg Canfell was the first in this thread to spot it, and now even our greatest player Ian Rogers and you have shown the glaring weakness of Bill's practice
You really have no clue about this you goose.
Look at Bonhams comments, look at Matts.


The question is he, will he admit this fault, and try to work out teh solution or will he still try to defend the indefensible?
Why would I admit to a fault that is no fault.
Of course expecting a goose like you to understand it is I suspect expecting a bit too much.

Cat
09-03-2004, 11:41 PM
You really are stupid arent you.

The whole situation with regards the change made for ratings over 2200 was explained at length on the old BB.

It had nothing to do bring them into line with ELo performances then and it still has nothing to do with them.

So you manipulate the dynamics of the Glicko and it just so happens that the outcome remarkably co-incides with the ELO predicted performance? Sheer co-incidence, nothing more?

chesslover
09-03-2004, 11:47 PM
So you manipulate the dynamics of the Glicko and it just so happens that the outcome remarkably co-incides with the ELO predicted performance? Sheer co-incidence, nothing more?

God does work in mysterious ways ;)

Bill Gletsos
09-03-2004, 11:57 PM
So you manipulate the dynamics of the Glicko and it just so happens that the outcome remarkably co-incides with the ELO predicted performance? Sheer co-incidence, nothing more?

Perhaps you should re-read the old thread on this on the old BB. You see they are unrelated.

Any similarity is just sheer co-incidence. In fact if you had looked closer you would have noticed there are a number that are quite different. Just look at Froelich, Feldman, Curtis, Laird, Cook, Barber etc.

Kevin Bonham
10-03-2004, 12:39 AM
I can assure you 100% that David has the brains to work out anything and everything that you throw at him. This is proven by the fact that he is a Doctor, unlike us all.

I must be a doctor too, the National Association Of Forest Industries (http://www.nafi.com.au/news/view.php3?id=857) certainly seem to think so.

(I am a doctor, but of "philosophy", not medicine. The letter printed on the site was a cruel and twisted joke I played on the poor Doctors for Forests. I am not sorry, and will be doing such things again. Whether NAFI were aware that I am not a medical doctor and chose to depict me as one for propaganda purposes, or were also sucked in, I have no idea.)


I resent you making these snide remarks that he is not smart enough to work that out

Actually Bill was reserving judgement. David's subsequent reply suggests that Bill was being too kind, as the top player tweak had nothing to do with ELO parity.

Alan Shore
10-03-2004, 01:24 AM
Wow, you all missed the most interesting part - how many 104 year olds there are on the best list! (on the post titled "Here is all players rated under the Glicko2 system rated 2200 and over who have been active since 1st June 2002 irrespective of rating code.")

Word of advice Bill, you only make yourself look stupid with language like that.

Cat
10-03-2004, 09:15 AM
Ah, you make it too easy for him.
I wanted to see if he could work it out for himself.

Ok, this is how it looks to me. After strenuously defending the purity of the Glicko system, you finally awake to the fact that the ratings, at least at the elite level, are in a mess. To seek a solution you use the ELO predictions to make a determination on the dynamic activity within the Glicko system, hence the virtually identical set of results. Correct?

Kevin Bonham
10-03-2004, 11:25 AM
Ok, this is how it looks to me. After strenuously defending the purity of the Glicko system, you finally awake to the fact that the ratings, at least at the elite level, are in a mess. To seek a solution you use the ELO predictions to make a determination on the dynamic activity within the Glicko system, hence the virtually identical set of results. Correct?

No. I discussed that tweak with Bill quite extensively at the time it was being mooted (actually I was rather sceptical of it at first) and I can state with confidence that your silly conspiracy theory is wrong. But I'll let Bill explain it, if he can be bothered.

Cat
10-03-2004, 11:39 AM
No. I discussed that tweak with Bill quite extensively at the time it was being mooted (actually I was rather sceptical of it at first) and I can state with confidence that your silly conspiracy theory is wrong. But I'll let Bill explain it, if he can be bothered.

OK, but it's not conspiracy theory but an honest answer to Bill's question. I hope I'm wrong!

Bill Gletsos
10-03-2004, 12:42 PM
Ok, this is how it looks to me. After strenuously defending the purity of the Glicko system, you finally awake to the fact that the ratings, at least at the elite level, are in a mess. To seek a solution you use the ELO predictions to make a determination on the dynamic activity within the Glicko system, hence the virtually identical set of results. Correct?
What a complete load of bullshit.

You really are starting to annoy me. :evil:

For someone supposedly intelligent you sure do come across as stupid.
Either that or just damn lazy. After all all of this was thoroughly discussed on the old BB at the time and in fact you even took part in the thread.

We can but hope your memory for remembering things in better when it comes to medical matters, otherwise god help a patient if you suddenly forget a particular procdure whilst performing it. :whistle:

Now to save you the trouble of looking for what was said I'll condense it into the following:

We have been aware of a possible anomaly at the top of the the rating system for some time(since last year). This anomaly can occur in any rating system but is possibly more pronounced in some cases in the Glicko.

It was not however considered by us as big a priority as for example, implementing the Glicko2 system(to better handle rapidly improving juniors) or fixing the ACT junior problem. Over the past few months whilst working on the ACT junior issue, we experimented with correcting this anomaly.

Now you are probably asking what is this anomaly?

Put as simply as possible it is this.

The opportunity/possibility for a player who loses rating points to regain those points is less for players at the top end as opposed to players at other positions in the rating list.

A player rated 1400 beats a player rated 1700. The 1400's chance of this happening was 15%. Now the 1700 can regain those points by beating a number of players or he can get them all back in one go by beating a player rated 2000. His chance of this is 15%.

However, when a player rated 2300 loses to a player rated 2000( a 15% chance), he can get the points back by winning a number of games but his chances of getting them all back in one go by beating a 2600 rated player is 0% as there is no local player currently rated 2600 on the list.
Likewise a 2200 losing to a 1900 can get them back by beating a 2500 except only an extremely small percentage of the rating pool exeeeds or is close to 2500. The players chances therefore of being able to ven play a 2500 are virtually 0.

This therefore means that a player at the top end of the list is at a disadvantage compared to players elsewhere in the list.

Now the major criteria as far as Graham and I were concerned was that any change to the system had to have as small as possible an impact on the predictive accuracy of the standard Glicko2 system.

We tried a number of scenarios/variations that turned out to totally fail on this criteria. I even re-tried as a test a test I had run well over a year ago which simply reduced all rating changes by 50% for all players. Like when I originally ran it, its effect on predictive accuracy was completely unacceptable.

What we eventually settled on for the September 2003 list was as follows:
For ratings of 2200 and under there is no change.
For ratings of 2300 and above the rating change is 50% of what would be calculated by standard Glicko2.
For ratings above 2200 and below 2300 then use a sliding sclae to determine the % change. e.g. for a rating of 2250 its 75%.

Obviously to implement this change we went back and recalculated all rating periods from Dec 2000

The impact on predictive accuracy is negligible and is therefore acceptable to Graham and I.

Now this may be fined tuned over the next few months, as we try out variations on this theme.

As it turned out further testing revealed some changes that reduced the impact on the predictive accuracy to less than the impact of the September 2003 implementation.

This new change was implemented in the special November 2003 rating list which also included the fix for the ACT junior problem.

The above described change in the November list therefore became:
For ratings of 2200 and under there is no change.
For ratings of 2300 and above the rating change is 75% of what would be calculated by standard Glicko2.
For ratings above 2200 and below 2300 then use a sliding scale to determine the % change. e.g. for a rating of 2250 its 87.5%.

Bill Gletsos
10-03-2004, 12:47 PM
Wow, you all missed the most interesting part - how many 104 year olds there are on the best list! (on the post titled "Here is all players rated under the Glicko2 system rated 2200 and over who have been active since 1st June 2002 irrespective of rating code.")

Well spotted Bruce.
The creation of a top list of players under Glicko2 irrespective of rating code is not a standard report. I just hacked the code for one of the reports that normally includes players ages. When I get a chance I'll correct it.



Word of advice Bill, you only make yourself look stupid with language like that.
Are you referring to the fact that at times my grammer is crook, or to my abuse of morons and geese. :whistle:

arosar
10-03-2004, 01:14 PM
Listen here Bill. You need to be more respectful of people - even if you disagree with them. You're a Prez for God's sake. This is sooo conduct unbecoming of you. Sometimes I get all upset with your mannerism Bill. You're too aggressive. Be more like Kevo - he always does it with style.

AR

Garvinator
10-03-2004, 01:44 PM
Bill. You're too aggressive. Be more like Kevo - he always does it with style.

AR
are you saying that kevin abuses ppl with style :lol:

Bill Gletsos
10-03-2004, 01:44 PM
Listen here Bill. You need to be more respectful of people - even if you disagree with them. You're a Prez for God's sake. This is sooo conduct unbecoming of you. Sometimes I get all upset with your mannerism Bill. You're too aggressive. Be more like Kevo - he always does it with style.
I'm not posting in my capacity as NSWCA pres but as one of the ACF's Rating Officers.

Greg Canfells post was fine and I was respectful to Greg. Likewise he was respectful to me. I'm respectful to those that are respectful to me. I havent called you a moron or a goose AR. Likewise I havent called the majority of posters those either.

David did not ask a question, he made accusations.
Given he was involved in the thread but can no longer remember it(sounds like he has caught Alzheimers like my mate Brian in the Australian Magazine thread) then his accusations are uncalled for.
As far as I am concerned if he is going to make scurrilous accustaions then that makes him fair game.

As for CL Matt noted "We all know what a fickle dude CL is here".
CL will support whoever supports him, irrespective of their merits, until of course they no longer support him.
His support of David in this issue is meaningless.

As for my mate Kevin, yes his style is so good he made himself the goosemaster with a great avatar. :clap:

Cat
10-03-2004, 02:31 PM
It's not an accusation Bill, its an attempt to gain clarification. The problem is Bill every time questions are raised you simply go ballistic. These are valid questions and simply having a tirade is no response at all.

You have repeatedly nailed your colours firmly to the Glicko flag over the last 12 months and denounced, indeed at times lambasted, ELO as an inferior rating system.

You recognise a problem at the upper end of the rating population, for which you introduce changes to the Glicko system in November and lo & behold the outcome is almost identical to that predicted by ELO.

I have read and understood your explanations for the manipulation. The question is how can one have such confidence that the Glicko is so superior to the ELO system, when after a series of manipulations, the published results are virtually a copy of the ELO predictions. It is a matter of natural process to then question previous held beliefs about the relative benefits of Glicko v's ELO as applied in Australia, not to do so is simple complacence.

You are going to have to do far better in way of an explanation than you have done so far. It's no good getting angry at me about this, you have dug this pit for yourself. Just calm down and think it through. Your arguments about why the manipulation has occurred are fair enough, the question is why is the outcome so similar to the ELO predictions? This you have not answered.

Oepty
10-03-2004, 03:08 PM
First of all I am surprised to see people the March 2004 ACF ratings to the January 2004 FIDE ratings, especially when dealing with top players. This is because the Australian Championship has not been rated and until the April 2004 FIDE ratings are released it seems pointlees.

Second of all, one of the major reasons Glicko2 is superior to ELO is its handling of fastly improving juniors and that is hardly revelant to the Top list.

Third of all, with the Glicko 2 ratings system, a rating is centrall to a intervall.
The interval is rating - 2*RD to rating + 2*RD. Apparently we can be then 95% confident a players strength is within these boundaries. So if we look at a person on the master list with a rating of 2500?? they have a very high RD, quite possibly even 350. So we can be 95% confident that this persons playing strength is between 1800 and 3200. Well, it is hardly likely to be 3200, but assuming a player has not died or physically or mentally declined greatly I would think that they would mostly play somewhere in the lower end of range given. I personally think that looking up the rating of a player who hasn't played for 15-20-30 years is a bit pointless and I don't really think it matters whether they are overrated or underrated as they are highly unlikely to play again. If they do and don't play to the level they previously did the rating they will get a rating alot lower on the next active list they are on.

Having said all of this though I still see the 70 point increase as being totally pointless, but it appears it not going to be changed so I will just have to live with it.
Scott

Bill Gletsos
10-03-2004, 03:52 PM
It's not an accusation Bill, its an attempt to gain clarification. The problem is Bill every time questions are raised you simply go ballistic. These are valid questions and simply having a tirade is no response at all.
Crap.
It wasnt a question, it was an accusation.
You said
Sheer co-incidence or the hand of God?
The last part implies manipulation.
I totally reject that.


You have repeatedly nailed your colours firmly to the Glicko flag over the last 12 months and denounced, indeed at times lambasted, ELO as an inferior rating system.
As I said Glicko is a better predictor. For a rating system that makes it superior.


You recognise a problem at the upper end of the rating population, for which you introduce changes to the Glicko system in November and lo & behold the outcome is almost identical to that predicted by ELO.
If you understood rating systems you would realise that is a likely outcome.


I have read and understood your explanations for the manipulation. The question is how can one have such confidence that the Glicko is so superior to the ELO system, when after a series of manipulations, the published results are virtually a copy of the ELO predictions.
There are some significant differences.
However the relationship between them at the top end isnt important, its the whole list that matters.
ELo doesnt handle under performing adults or rapidly improving adults at all well.
Also ELO doesnt handle most juniors at all well, unless the juniors not only plateau but also play many games. Even then it takes many rating periods for ELO to catch up if at all.



It is a matter of natural process to then question previous held beliefs about the relative benefits of Glicko v's ELO as applied in Australia, not to do so is simple complacence.
No its not.
It just appears that all you are doing is focusing on the top 50 odd players and ignoring what has been discussed regarding juniors, declining/improving adults etc over the past 15 mths.
It not like you have only just entered this debate.
For someone who has supposedly been following all this with a keen interest over all that time, it makes your questions look suspicious.


You are going to have to do far better in way of an explanation than you have done so far.
One could hope that your understasnding would have improved over the past 15 mths.


It's no good getting angry at me about this, you have dug this pit for yourself.
I havent dug anything.
You just seem to be being deliberately forgetful of all the previous rating dicussions.



Just calm down and think it through. Your arguments about why the manipulation has occurred are fair enough, the question is why is the outcome so similar to the ELO predictions? This you have not answered.
I would have thought it was obvious.
Firstly all the players Glicko ratings originally started from their ELo ratings on August 2000 list.
Secondly for players who perform close to their rating then the K factor value is pretty immaterial.

The old ACF ELO uses a fixed K of 15. This is applied to the difference between actual and expected scores.
Glicko has no such thing as a K factor but to put it in simplistic terms effectively multiples the difference between actual and predicted scores by a wide ranging value(an effective K).

For double !! players this value is likely to be between 15-32.
No for players over 2300 this means their effective K is somewhere between 11.25 and 24.
For players under 2200 their effective K is 15-32.

Now lets assume the difference between actual and expected 0.5.
If K is 15 thats equal to 7.5 points.
If K is 10 its 5 points.
Even if K is 32 its only 16 points.

Now for example back in April 2002 Johansen was rated 2518 on Glicko2 and 2516 on ELO. He peformed below expectations in August 2002 and his Glicko2 dropped to 2489 and his ELO to 2485. In December 2002 he again performed below expectations and his Glicko2 dropped to 2438 and his ELO to 2441.
To this point he lost more under Glicko2 than ELO but his original ACF was lower so when looking at the absolutes the difference is smaller.
Now in March 2003 he performed slightly above expectations. His Glicko2 went to 2446 and his ELO to 2446. His Glicko2 gain is more than his ELO and the ratings come back into sync.

Now look at Barber.
In August 2002 he was rated 2369 under Glicko2 and 2371 under ELO. He performed well under expectations in Dec 2002 and his Glicko2 dropped to 2263 whilst his ELO hit 2335. In June 2003 he again was well under his expected expectations. His Glicko2 dropped to 2144 and his ELO to 2282. In sept 2003 his performance again was slightly below expectations and his Glicko2 hit 2129 and his ELO 2256. In December 2003 he pretty much performed to expectations. His Glicko2 dropped very slightly to 2126 and his ELO to 2246. There is now a 120 difference between the two systems. Even in the latest period his Glicko2 is only 2159 whilst his ELo is 2265, a difference of 106.


However lets look at a junior.
Take Junta Ikeda.
His Glicko2 ratings and ELo ratings and the differences are respectively
Aug 2000 539 539 0
Dec 2000 558 562 -4
Aug 2001 608 592 16
Dec 2001 633 612 21
Apr 2002 946 820 126
Aug 2002 1300 1002 298
Dec 2002 1397 1221 176
Mar 2003 1456 1294 162
Jun 2003 1566 1458 108
Sep 2003 1629 1552 77
Nov 2003 1664 1618 46
Dec 2003 1705 1665 40
Mar 2004 1773 1732 41

Eventually the ELO started to catch up mainly due to the sheer number of games he played in the Dec 2002, June 2003 and Sep 2003 periods when he played 48, 54 and 43 games respectively in the periods.
keep in mind that the Nov 2003 period involved the ACT correction and that due to the 70 point uplift in march his march rating is virtually identical to his Dec 2003.
Even so his Glicko2 is still ahead.

Also look at James Obst another junior.
His Glicko2 ratings and ELo ratings and the differences are respectively
Apr 2002 1508 1473 35
Aug 2002 1491 1464 27
Dec 2002 1568 1537 31
Mar 2003 1532 1521 11
Nov 2003 1539 1514 25
Mar 2004 1833 1694 139

As you can see the ELo just cannot keep up with the initial improvement.

Kevin Bonham
10-03-2004, 04:16 PM
First of all I am surprised to see people the March 2004 ACF ratings to the January 2004 FIDE ratings, especially when dealing with top players. This is because the Australian Championship has not been rated and until the April 2004 FIDE ratings are released it seems pointlees.

Well spotted, Scott, and that's yet another thing wrong with Peter Parr's comparisons.

chesslover
10-03-2004, 06:44 PM
It's not an accusation Bill, its an attempt to gain clarification. The problem is Bill every time questions are raised you simply go ballistic. These are valid questions and simply having a tirade is no response at all.

You have repeatedly nailed your colours firmly to the Glicko flag over the last 12 months and denounced, indeed at times lambasted, ELO as an inferior rating system.

You recognise a problem at the upper end of the rating population, for which you introduce changes to the Glicko system in November and lo & behold the outcome is almost identical to that predicted by ELO.

I have read and understood your explanations for the manipulation. The question is how can one have such confidence that the Glicko is so superior to the ELO system, when after a series of manipulations, the published results are virtually a copy of the ELO predictions. It is a matter of natural process to then question previous held beliefs about the relative benefits of Glicko v's ELO as applied in Australia, not to do so is simple complacence.

You are going to have to do far better in way of an explanation than you have done so far. It's no good getting angry at me about this, you have dug this pit for yourself. Just calm down and think it through. Your arguments about why the manipulation has occurred are fair enough, the question is why is the outcome so similar to the ELO predictions? This you have not answered.

excellent excellent point David

but I do think you will never be ableto convinvce Bill of the error of his Glicko ways. Whilst Glicko 2 is Bill's "baby" and he is very protective of it, even parents have to admit that there are flaws in their children.

Bill is so one-eyed that he refuses to admit that his Glicko2 is wrong at the top end.I have no knowledge of maths or rating, but even I know that what you state makes sense to the lay person.From this thread even the number 1 player in Australia, and the multiple champion of our state Greg canfell share your concerns.

Bill should accept that no system is perfect and agree to discuss with you and the others on how to improve his "baby". Instead he even refuses to admit that there is a flaw!!!! Talk aboutseeing the world with rose colored glasses hey?

I hope that he sits down and contemplates your comments before he posts again, and posts in a manner that facilitates the fixing of the Glicko 2 problem, instead of just denying that a problem even exists.

Bill Gletsos
10-03-2004, 07:04 PM
excellent excellent point David
There is nothing excellent about it.
He has no clue what he is talking about.


but I do think you will never be ableto convinvce Bill of the error of his Glicko ways. Whilst Glicko 2 is Bill's "baby" and he is very protective of it, even parents have to admit that there are flaws in their children.
This is just your usual crap.


Bill is so one-eyed that he refuses to admit that his Glicko2 is wrong at the top end.I have no knowledge of maths or rating, but even I know that what you state makes sense to the lay person.From this thread even the number 1 player in Australia, and the multiple champion of our state Greg canfell share your concerns.
You are a total goose.

There is no comparison between David's comments and those of Greg or Ian.
David isnt even talking about the same thing as Greg or Ian you moron.

David is interested in the seeming closeness of the ELO and Glicko2 ratings of the top players.

Greg is only interested in the fact that inactives got the 70 point bonus.

Ian was suggesting that inactives should have gotten less than the 70 points anf that those that have been inactive for a long time should have gotten less than the 150 they got back in April 2000.


Bill should accept that no system is perfect and agree to discuss with you and the others on how to improve his "baby". Instead he even refuses to admit that there is a flaw!!!! Talk aboutseeing the world with rose colored glasses hey?
Why would I listen to David. He has no clue.

As for Greg and Ian, explained the issues to them.
I think they both realise it is not a simple process to just take points from inactive players in a straightforward manner.


I hope that he sits down and contemplates your comments before he posts again, and posts in a manner that facilitates the fixing of the Glicko 2 problem, instead of just denying that a problem even exists.
You idiot.
There is no Glicko2 problem.

Certainly the 70 point uplift is totally unrelated to Glicko2.

As for the similarity of ELO and Glicko2 ratings at the top thats not a flaw but is to be expected as I explained in a previous post.

Bill Gletsos
10-03-2004, 07:15 PM
Well spotted Bruce.
The creation of a top list of players under Glicko2 irrespective of rating code is not a standard report. I just hacked the code for one of the reports that normally includes players ages. When I get a chance I'll correct it.
Ok, I went and corrected that listing and replaced the ages with games played in the period.
The post on the BB has also been changed to reflect this.

Cat
10-03-2004, 10:09 PM
Thanks Bill for the reply. If I can further clarify;

1. The modified Glicko results seem closer to the ELO than the Glicko results with the old dynamics? Maybe this isn't correct, but the correlation appears to be closer superficially anyway.

2. If the modified Glicko results are closer to the ELO, do you believe they are more accurate (modified Glicko)?

3. If they are more accurate, at what point do you believe the dynamic Glicko outperforms the less dynamic system.

The point of these questions is to ;
a) understand your justification for drawing the line at 2200 or 2300, whatever it is.
b)consider the effect on high rated players in isolated regions of differing dynamics.

Rhubarb
10-03-2004, 10:13 PM
I personally think that looking up the rating of a player who hasn't played for 15-20-30 years is a bit pointless and I don't really think it matters whether they are overrated or underrated as they are highly unlikely to play again.

Well I guarantee you they won't be underrated. :)

You and Bill and Kevin might understand the ratings of people who haven't played for 15-20-30 years who are rated over say, 2400, to be unimportant (and it's true, Glicko would quickly take care of them if they ever returned), but to me it's about everyday perception. What about a player who hasn't played for 5 years who's been given the 220 points? It may well be that they play again in the future without any intrinsic loss of chess skill, so would Scott and Bill and Kevin still say: "Their rating is unimportant."

To go back to the point of my original post to this BB: Inactive players at the top end simply do not deserve to be given 220 points gratis and to have such a high number next to their names.


As for Greg and Ian, explained the issues to them.
I think they both realise it is not a simple process to just take points from inactive players in a straightforward manner.

I agree it is not a simple process to just take points from inactive players (or, more to the point, to decide which inactive players do not get free points or get less than 220 points). But does this mean you would take points away from them if you could? In other words, do you agree the problem exists?

Greg

P.S. I am, as ever, only talking about the top end of the rating scale, not because I'm elitist, but because I am sure there is a problem here, and this problem may or may not exist throughout the rest of the rating system.

Bill Gletsos
10-03-2004, 10:23 PM
Thanks Bill for the reply.
Thank you for failing to address your original accusation as well as your apparant attack of Alzheimers in regards to the original discussion on the old BB.
I think it says volumes for you credability. (A razoo emoticon would come in handy around here)


If I can further clarify;

1. The modified Glicko results seem closer to the ELO than the Glicko results with the old dynamics? Maybe this isn't correct, but the correlation appears to be closer superficially anyway.
I thought I made that obvious from my last post.


2. If the modified Glicko results are closer to the ELO, do you believe they are more accurate (modified Glicko)?
I thought I made it clear in my last post that the modified Glicko2 was statistically inferior to standard Glicko2 as far as predictive accuracy. However the loss was fairly small and therefore acceptable. That was explained on the old BB.


3. If they are more accurate, at what point do you believe the dynamic Glicko outperforms the less dynamic system.
See answer above.


The point of these questions is to ;
a) understand your justification for drawing the line at 2200 or 2300, whatever it is.
This was explained in an earlier post. It was determined by testing.
Lowering the limits had unacceptable adverse effects on predictive accuracy.


b)consider the effect on high rated players in isolated regions of differing dynamics.
You raised that back on the old BB. I and others dismissed it then as irrelevant.
I dismiss it now.

If all you are going to do is rehash what was discussed at length last year, please stop wasting my time.

Either that or I might just decide to stop answering questions that I deem a waste of my time. :whistle:

chesslover
10-03-2004, 10:33 PM
If all you are going to do is rehash what was discussed at length last year, please stop wasting my time.

Either that or I might just decide to stop answering questions that I deem a waste of my time. :whistle:

But if you do that, what happenms to your post count? ;)

Or do you feel that now you (like me yay!!!!) have custom title status, it does not matter.

I notice you still have not answered my simple rating question. Why not as everything is relative add 20 points each rating period? It will make chess players happy, and will not hurt anyone.

Bill Gletsos
10-03-2004, 10:40 PM
Well I guarantee you they won't be underrated. :)
Actually Greg they could be.
For example a junior on the list from 10-20 years ago may return to chess as an adult and be stronger than when they stopped.


You and Bill and Kevin might understand the ratings of people who haven't played for 15-20-30 years who are rated over say, 2400, to be unimportant (and it's true, Glicko would quickly take care of them if they ever returned), but to me it's about everyday perception. What about a player who hasn't played for 5 years who's been given the 220 points? It may well be that they play again in the future without any intrinsic loss of chess skill, so would Scott and Bill and Kevin still say: "Their rating is unimportant."
I think we would say their rating needs to be tested by them actually competing.


To go back to the point of my original post to this BB: Inactive players at the top end simply do not deserve to be given 220 points gratis and to have such a high number next to their names.
From a strictly mathematical view point they do.
If no one had been given the 70 points and the 150 points then the active players ratings relative to the inactive players would be the same as they are now.




I agree it is not a simple process to just take points from inactive players (or, more to the point, to decide which inactive players do not get free points or get less than 220 points). But does this mean you would take points away from them if you could? In other words, do you agree the problem exists?
Whether it is a problem or not is a matter of perception.
Some would see it as a problem, some would not.
I do however believe it is a situaton worthy of investigation and in fact is something that is being investigated.


P.S. I am, as ever, only talking about the top end of the rating scale, not because I'm elitist, but because I am sure there is a problem here, and this problem may or may not exist throughout the rest of the rating system.
I dont think I called you elist, nor implied it.
I may have used it to describe the players at the top end of the rating list, but I dont believe I was critical of them as a group during our discussion.

Rhubarb
10-03-2004, 10:40 PM
But does this mean you would take points away from them if you could? In other words, do you agree the problem exists?

Actually Bill, don't bother answering this, I just read an earlier post oy yours explaining that you were looking into this.

Bill Gletsos
10-03-2004, 10:43 PM
But if you do that, what happenms to your post count? ;)
There is lots to post about other than ratings.
After all there are always geese to corral in the other threads.


Or do you feel that now you (like me yay!!!!) have custom title status, it does not matter.
I post because I can, not because I need to.


I notice you still have not answered my simple rating question. Why not as everything is relative add 20 points each rating period? It will make chess players happy, and will not hurt anyone.
I guess I overlooked it because it seemed so silly.

On what basis would I justify the 20 points.

Feel good value is not an answer.

Bill Gletsos
10-03-2004, 10:44 PM
Actually Bill, don't bother answering this, I just read an earlier post oy yours explaining that you were looking into this.
Too late I already replied at the same time as you told me not to bother. :doh:

Cat
10-03-2004, 11:00 PM
[QUOTE]Thank you for failing to address your original accusation as well as your apparant attack of Alzheimers in regards to the original discussion on the old BB.
I think it says volumes for you credability. (A razoo emoticon would come in handy around here)

Every time you make one of your Rainman responses it's your own credibility your damaging. Quite frankly the Glicko system has recieved more patches in the last 12 months than Microsoft Windows, so don't keep trying to deflect questions with arrogance. These are ligitimate questions.


I thought I made that obvious from my last post.

Enlighten me! What are the differences, simple question, easy to answer.


I thought I made it clear in my last post that the modified Glicko2 was statistically inferior to standard Glicko2 as far as predictive accuracy. However the loss was fairly small and therefore acceptable. That was explained on the old BB.

You actually said Graham and yourself considered the predictive difference to be negligible, I was asking for clarification. Also, you were referring to your modelling, I'm really interested in your assessment of the outcomes.



This was explained in an earlier post. It was determined by testing.
Lowering the limits had unacceptable adverse effects on predictive accuracy.

In what way?

Bill Gletsos
10-03-2004, 11:33 PM
Every time you make one of your Rainman responses it's your own credibility your damaging. Quite frankly the Glicko system has recieved more patches in the last 12 months than Microsoft Windows, so don't keep trying to deflect questions with arrogance. These are ligitimate questions.
Look Doctor Dipshit,
You made an accusation that the ELO ratings I posted were manipulated. :evil:
Retract it unconditionally or you can whistle for any further answers after this post. :evil:

As for the Glicko having more patches than MS Windows, please list the patches applied to Glicko in the past 12 mths.




Enlighten me! What are the differences, simple question, easy to answer.
I have no idea what differences you refer to.
You asked:
If I can further clarify;

1. The modified Glicko results seem closer to the ELO than the Glicko results with the old dynamics? Maybe this isn't correct, but the correlation appears to be closer superficially anyway.
That asks no question about differences.
It appears you are asking if the modified Glicko2 for players above 2200 generates ratings closer to ELO than an unmodified Glicko2.
I answered that in an earlier post.



You actually said Graham and yourself considered the predictive difference to be negligible, I was asking for clarification. Also, you were referring to your modelling, I'm really interested in your assessment of the outcomes.
I think you will find what was said was "Now the major criteria as far as Graham and I were concerned was that any change to the system had to have as small as possible an impact on the predictive accuracy of the standard Glicko2 system."
That comment about negligible was made in a post that hopefully anyone could understand. Perhaps negligible was inaccurate, lets just say it was small.

As for modelling I cannot remember if we ran the modelling on a PC or a mainframe. I originally ran some modelling about 18mths prior to its implementation on a PC. I know we did modelling for the ACT problem on the mainframe and I know we were working on the change for the top end of the list at the same time. I think we did the modelling at that stage on the mainframe also.
We had the change to the standard Glicko2 cut in at various rating points and we also tried various different methods of varying the amount of the change.

What we eventually settled on satisfied our criteria that any change to the system had to have as small as possible an impact on the predictive accuracy of the standard Glicko2 system.



In what way?
I think its fairly obvious what "Lowering the limits had unacceptable adverse effects on predictive accuracy".
It means that the predictive accuracy got worse.

Rhubarb
10-03-2004, 11:54 PM
Actually Greg they could be.
For example a junior on the list from 10-20 years ago may return to chess as an adult and be stronger than when they stopped.

Yes okay, it is possible a player inactive since April 2000 could be underrated. If this actually happened to anyone rated over say, 2200, I would be stunned. I think we would both agree (intuitively, if not mathematically) that the vast majority of inactive players are overrated. IMO, this has been severely exacerbated at the top end by the 220-point addition.


I think we would say their rating needs to be tested by them actually competing.

Well of course. There would be no problem if there were no inactive players. (At least not the problem I'm arguing.)


From a strictly mathematical view point they do.
If no one had been given the 70 points and the 150 points then the active players ratings relative to the inactive players would be the same as they are now

Yes I know. But if that had been the case, then a continuously active player who was about 2300 in April 2000 would now be rated about 2100, and an inactive player who was 2300 in April 2000 would still be 2300.


Whether it is a problem or not is a matter of perception.
Some would see it as a problem, some would not.

OK, to state the obvious, I do see it as a problem. To make an emotional argument, it annoys me to see players -- some of whom IMO would be lucky to be 2100 if they had been playing regularly in the last four years -- with 2300 (or even 2400) next to their name, and to me this number has a far greater impact than the one or two question marks next to their names.


I dont think I called you elist, nor implied it.
I may have used it to describe the players at the top end of the rating list, but I dont believe I was critical of them as a group during our discussion.

No I wasn't suggesting that you implied I was elitist. I was just explaining for the benefit of others that I'm only referring to the top end because not all of my arguments necessarily apply to the rating system as a whole.

Cat
10-03-2004, 11:59 PM
Look Doctor Dipshit,
You made an accusation that the ELO ratings I posted were manipulated. :evil:
Retract it unconditionally or you can whistle for any further answers after this post. :evil:

I didn't accuse you of manipulating the ELO ratings. In fact, I didn't accuse you of anything, simply when you asked my take on why the ratings were so uncannily similar I gave you an honest answer. If it was taken as an accusation, I'm sorry.




[QUOTE]It appears you are asking if the modified Glicko2 for players above 2200 generates ratings closer to ELO than an unmodified Glicko2.
I answered that in an earlier post.

Well I missed it, I guess I'll just have to trawl your posts if I want to find out?


I think its fairly obvious what "Lowering the limits had unacceptable adverse effects on predictive accuracy".
It means that the predictive accuracy got worse.

The point is that if the predictive accuracy got worse then one would expect the 3 rating systems, Glicko 2, modified Glicko 2 and ELO, to diverge as the ratings fell. I was offering you a chance to elaborate. But don't worry, you've cleared up most of the questions and I'll find the reply you were refering to.

Cheers

Bill Gletsos
11-03-2004, 12:18 AM
I didn't accuse you of manipulating the ELO ratings. In fact, I didn't accuse you of anything, simply when you asked my take on why the ratings were so uncannily similar I gave you an honest answer. If it was taken as an accusation, I'm sorry.
You used the words "Sheer co-incidence or the hand of God? "
The Hand of God comment looks like an accusation to me.
Anyway I'm prepared to accept your apology.




Well I missed it, I guess I'll just have to trawl your posts if I want to find out?

I believe I explained it sufficiently in my post on 10-03-2004 at 04:52 PM.



The point is that if the predictive accuracy got worse then one would expect the 3 rating systems, Glicko 2, modified Glicko 2 and ELO, to diverge as the ratings fell. I was offering you a chance to elaborate. But don't worry, you've cleared up most of the questions and I'll find the reply you were refering to.

Eventually with enough games and some consistencey the ratings will start to converge to something close to a players "true strength" irrespective of the rating system used. This is especially true since all players Glicko2 ratings were originally derived from their ELO.

Rincewind
11-03-2004, 12:26 AM
You used the words "Sheer co-incidence or the hand of God? "
The Hand of God comment looks like an accusation to me.
Anyway I'm prepared to accept your apology.

Anyone with a knowledge of soccer will know that the "hand of god" was used to describe a very famous piece of cheating. However I believe the expression pre-dates that incident so an innocent interpretation is possible. However, the expression has got to have acquired a some negative connotation by association.

Kerry Stead
11-03-2004, 01:10 AM
What came first? Diego or Monty Python?
I suspect they were both pre-dated by something ... but would be an interesting fact to know anyway.

Kevin Bonham
11-03-2004, 01:33 AM
You and Bill and Kevin might understand the ratings of people who haven't played for 15-20-30 years who are rated over say, 2400, to be unimportant (and it's true, Glicko would quickly take care of them if they ever returned), but to me it's about everyday perception. What about a player who hasn't played for 5 years who's been given the 220 points? It may well be that they play again in the future without any intrinsic loss of chess skill, so would Scott and Bill and Kevin still say: "Their rating is unimportant."

Hmmm, I thought I'd said on this thread that I did believe a decay factor for inactivity was likely to be justified with further research, and also that I did think there was at least one good reason to fix these rusty ratings up (seedings for Swiss events should these players return). So while I do think that these players being overrated is not a huge deal (it's not like their overrated nature when they return will really affect selections or anyone else's rating) the above probably isn't quite doing my stated positions justice.

I also think your argument about perceptions is more than fair enough, so long as it is rigorously proved that these players really are as overrated as they seem.

Rhubarb
11-03-2004, 04:19 AM
The above probably isn't quite doing my stated positions justice.

Kevin, sorry for misrepresenting your stated positions.

Yes I think it's important to keep the issue of strength decay due to inactivity (an effect that exists for the overwhelming majority of players and exists across the entire rating system, I would suggest) separate from the question of whether inactive players (as always, at the top end ) have unfairly benefited from the blanket 150 and 70 point additions since April 2000, due, perhaps, to compaction or deflation at the top end, or for whatever reason.


I also think your argument about perceptions is more than fair enough, so long as it is rigorously proved that these players really are as overrated as they seem.

So the perception, at least amongst the players I've spoken to, is that inactive players are now significantly overrated relative to continuously active players (without including strength decay). I personally don't have the ability to rigorously prove this at this time, but at least Bill said he was investigating it.

Regards,
Greg

ursogr8
11-03-2004, 07:29 AM
But if you do that, what happenms to your post count? ;)

Or do you feel that now you (like me yay!!!!) have custom title status, it does not matter.



CL

Well done on getting the meter over 2000.

BTW did you notice ggrayggray on 999. Strange that his count has accelerated recently. And he seems to have stopped posting in the non-chess area. Odd that.

starter

Bill Gletsos
11-03-2004, 10:17 AM
BTW did you notice ggrayggray on 999. Strange that his count has accelerated recently.
I think thats because gg went hunting geese. :whistle:

Garvinator
11-03-2004, 10:58 AM
I think thats because gg went hunting geese. :whistle:
what do you mean went hunting, that is past tense. As you will see soon, im still hunting :whistle:

And Btw, i have posted in the formula one section thank you :hand:

Bill Gletsos
11-03-2004, 11:15 AM
what do you mean went hunting, that is past tense. As you will see soon, im still hunting :whistle:
Just remember the bigger the gun the better.



And Btw, i have posted in the formula one section thank you :hand:
It was starter not I who mentioned the non-chess threads.

ursogr8
11-03-2004, 11:23 AM
what do you mean went hunting, that is past tense. As you will see soon, im still hunting :whistle:

And Btw, i have posted in the formula one section thank you :hand:

You would be in big trouble if K. decided to also switch off the counter for responses to CL posts. You would probably stall before 1000. So you need K. and you need CL.

Garvinator
11-03-2004, 11:30 AM
It was starter not I who mentioned the non-chess threads.
i know, i wish i had pryamid quoting there. I knew that would look bad.

skip to my lou
11-03-2004, 12:43 PM
You can use pyramid quoting by simply manually typing the quote tags and pasting the text.

Oepty
11-03-2004, 02:12 PM
Well, I believe Greg, and I have had discussions with others, when he says inactive probably overrated relative to active players. By this I mean they were not as strong when they stopped playing chess as what their current rating would make them seem. The question is, if it is true, WHY? I think the last 70 point increase can not in any way be blamed for it. I don't know enough about why the 150 point increase was given to comment on that. Perhaps Bill can you either give an explanation or point me to a explanation.

The only way I can see inactive players getting overrated as I explained above is by the pool of active players deflating as some point in time and by a reasonable amount. I DO NOT think that this has happened since the introduction of Glicko2. So if I am correct we are starting to look back into more distant rating history to try and work how it happened and what can be done to fix it fairly. I really think it is alot to ask of our volunteer rating officers to be looking back through the 90' and perhaps even the 80's so that the ratings of players who haven't been playing for 10-30 years can be corrected. I think we are just going to have to accept that the master list will have some strange numbers in it.

Bill, another question do you take dead people off the list?
Scott

Bill Gletsos
11-03-2004, 02:22 PM
Bill, another question do you take dead people off the list?
Yes, provided someone tells me or State Rating Officer that a particular player is in fact dead.

Garvinator
11-03-2004, 04:16 PM
Yes, provided someone tells me or State Rating Officer that a particular player is in fact dead.
sorry to make a bad joke, but would a dead person be over-rated ;)