PDA

View Full Version : ACF September 2006 Ratings



Pages : [1] 2

Bill Gletsos
01-09-2006, 12:38 AM
The ACF September 2006 rating lists have been sent to the webmaster and State Rating Officers.

For the September 2006 rating period there were 266 Tournaments rated (100 Classic, 166 Rapid) and 19013 games of which 7495 were in the ACF Classic rating system and 11518 were in the ACF Rapid rating system.

Top Players
2613!! 28 NSW Rogers, Ian [GM]
2461!! 0 NSW Zhao, Zong-Yuan [IM]
2458! 0 NSW Wohl, Aleksander H [IM]
2455!! 0 VIC Johansen, Darryl K [GM]
2442!! 7 VIC Smerdon, David C [IM]
2416!! 0 NSW Lane, Gary W [IM]
2397!! 14 QLD Solomon, Stephen J [IM]
2380! 6 SA Chapman, Mark [IM]
2372!! 7 QLD Froehlich, Peter [IM]
2366!! 29 NSW Bjelobrk, Igor [FM]
2345!! 14 VIC Goldenberg, Igor [FM]
2323! 7 SA Tao, Trevor
2321!! 16 VIC West, Guy [IM]
2315!! 35 NSW Xie, George [FM]
2311! 0 VIC Sandler, Leonid [IM]
2294! 0 NSW Reilly, Tim [FM]
2282! 0 VIC Baron, Michael [FM]
2276!! 0 NSW Canfell, Gregory J [FM]
2269! 0 NSW Drummond, Matthew
2256! 0 NSW Feldman, Vladimir [IM]
2256!! 6 NSW Smirnov, Vladimir
2252! 8 NSW Tan, Justin
2246!! 9 VIC Rujevic, Mirko [IM]
2238! 0 VIC Depasquale, Chris J [FM]
2223!! 25 NSW Rej, Tomek
2214!! 14 NSW Ayvazyan, Armen
2207! 1 NSW Tindall, Brett [FM]
2207! 0 VIC Jordan, Bill [FM]
2206! 0 VIC Levi, Eddy L [FM]
2198! 0 VIC Chow, Samuel
2197!! 8 NSW Flatow, A (Fred) [FM]
2190!! 10 NSW Song, Raymond
2187!! 18 WA Boyd, Tristan
2180!! 7 QLD Ly, Moulthun
2179! 6 NSW Scott, Ronald
2178!! 6 QLD Humphrey, Jonathan [FM]
2173!! 29 SA Zaric, Srboljub
2168! 0 TAS Dowden, Tony
2166! 0 QLD Caoili, Arianne B [WFM]
2160! 0 NSW Berezina - Feldman, Irina [IM]
2141! 0 NSW Fuller, Maxwell L [FM]
2133! 0 WA Lakner, Jay
2126! 0 NSW Bouchaaya, Tony
2120!! 3 NSW Jones, Brian A [FM]
2118!! 42 VIC Stojic, Dusan
2116! 0 NSW Goris, Robert
2113! 5 NSW Vojvodic, Branislev
2106!! 7 NSW Yu, Ronald
2105!! 7 SA Obst, James
2104! 1 SA Sheldrick, Kevin
2104!! 15 SA Guthrie, Aaron [FM]
2100! 0 VIC Le, Tuan N
2100!! 14 NSW Halpin, Patrick
2098!! 7 NSW O'Chee, Kevin
2095!! 14 NSW Charles, Gareth
2092!! 32 NSW Wright, Neil S
2091! 3 NSW Quick, Simon
2091!! 20 NSW Bird, Andrew
2091! 0 VIC Booth, Stewart
2090!! 16 SA Cowley, Robert G (Bob)

Bill Gletsos
01-09-2006, 12:39 AM
Top Females
2166! 0 QLD Caoili, Arianne B [WFM]
2160! 0 NSW Berezina - Feldman, Irina [IM]
2081! 0 NSW Koshnitsky, Ngan [WIM]
2079! 5 NSW Eriksson, Ingela
2045! 0 NSW Dekic, Biljana [WIM]
2044!! 25 NSW Moylan, Laura A [WIM]
1989!! 3 NSW Song, Angela [WFM]
1921!! 9 NSW Musaeva, Aina
1823! 0 NSW Lane, Nancy L [WIM]
1757!! 23 ACT Oliver, Shannon [WFM]
1690!! 7 WA Maris, Natalie A
1654!! 10 WA Payne, Sophie
1649!! 36 VIC Szuveges, Narelle S [WIM]
1647! 0 NSW Harris, Rebecca
1636!! 22 QLD Jule, Alexandra
1605!! 16 NSW Reid, Vaness
1571! 0 VIC Beggs, Diana
1481!! 9 NSW Troshenkova, Irina
1463!! 8 NSW Grbin, Tereza
1454!! 19 TAS Martin, Janice

Bill Gletsos
01-09-2006, 12:40 AM
Top Under 20
2461!! 19 NSW Zhao, Zong-Yuan [IM]
2223!! 19 NSW Rej, Tomek
2198! 18 VIC Chow, Samuel
2190!! 11 NSW Song, Raymond
2180!! 14 QLD Ly, Moulthun
2166! 19 QLD Caoili, Arianne B [WFM]
2118!! 16 VIC Stojic, Dusan
2106!! 17 NSW Yu, Ronald
2105!! 15 SA Obst, James
2098!! 18 NSW O'Chee, Kevin
2077!! 15 VIC Wallis, Christopher
2072!! 17 NSW Suttor, Vincent
2071!! 14 ACT Ikeda, Junta
2053! 16 ACT Wei, Michael
2050!! 17 ACT Oliver, Gareth
2036!! 18 NSW Hu, Jason
2028!! 19 NSW Zvedeniouk, Ilia
2000!! 19 VIC Bourmistrov, Denis
1997! 16 NSW Cronan, James
1989!! 13 NSW Song, Angela [WFM]

Bill Gletsos
01-09-2006, 12:40 AM
Top Under 18
2190!! 11 NSW Song, Raymond
2180!! 14 QLD Ly, Moulthun
2118!! 16 VIC Stojic, Dusan
2106!! 17 NSW Yu, Ronald
2105!! 15 SA Obst, James
2077!! 15 VIC Wallis, Christopher
2072!! 17 NSW Suttor, Vincent
2071!! 14 ACT Ikeda, Junta
2053! 16 ACT Wei, Michael
2050!! 17 ACT Oliver, Gareth
1997! 16 NSW Cronan, James
1989!! 13 NSW Song, Angela [WFM]
1950!! 17 NSW Morris, Michael
1929!! 13 NSW Illingworth, Max
1898!! 15 VIC Vijayakumar, Rukman
1895!! 17 VIC Jager, Jesse
1881!! 12 VIC Lin, Zhigen Wilson
1880! 16 QLD Barnard, Casey T
1863!! 14 NSW Huang, Justin
1856! 15 VIC Lugo, Ruperto

Bill Gletsos
01-09-2006, 12:41 AM
Top Under 16
2190!! 11 NSW Song, Raymond
2180!! 14 QLD Ly, Moulthun
2105!! 15 SA Obst, James
2077!! 15 VIC Wallis, Christopher
2071!! 14 ACT Ikeda, Junta
1989!! 13 NSW Song, Angela [WFM]
1929!! 13 NSW Illingworth, Max
1898!! 15 VIC Vijayakumar, Rukman
1881!! 12 VIC Lin, Zhigen Wilson
1863!! 14 NSW Huang, Justin
1856! 15 VIC Lugo, Ruperto
1850!! 11 VIC Morris, James
1848!! 15 WA Donaldson, Thomas
1836!! 15 QLD Lazarus, Benjamin
1829! 14 VIC Jia, Jing Qu
1807!! 14 ACT Hoang, Khoi
1777!! 15 ACT Guo-Yuthok, Sherab
1763!! 13 ACT Brown, Andrew
1748!! 14 NSW Hoving, Eliot
1741!! 13 WA Choong, Yita

Bill Gletsos
01-09-2006, 12:41 AM
Top Under 14
2190!! 11 NSW Song, Raymond
1989!! 13 NSW Song, Angela [WFM]
1929!! 13 NSW Illingworth, Max
1881!! 12 VIC Lin, Zhigen Wilson
1850!! 11 VIC Morris, James
1763!! 13 ACT Brown, Andrew
1741!! 13 WA Choong, Yita
1670!! 13 NSW Harris, Benjamin
1662!! 13 QLD Anderson, Daniel C
1659!! 13 NSW Wu, Edwin
1656!! 10 ACT Yuan, Yi
1651!! 13 VIC Yu, Derek
1614!! 11 VIC Schon, Eugene
1587!! 12 NSW Miranda, Adrian
1560! 12 NSW Xu, William
1550!! 12 VIC Tang, Jason
1521!! 11 QLD Finke, Kelvin
1511! 13 VIC Xu, Mingda
1500! 13 NSW Zhu, Jonathan
1498!! 13 SA Zulfic, Fedja

Bill Gletsos
01-09-2006, 12:41 AM
Top Under 12
2190!! 11 NSW Song, Raymond
1850!! 11 VIC Morris, James
1656!! 10 ACT Yuan, Yi
1614!! 11 VIC Schon, Eugene
1521!! 11 QLD Finke, Kelvin
1437!! 11 ACT Chow, Justin
1419!! 11 VIC Yu, Sally
1415!! 10 ACT Guo, Emma
1397!! 10 QLD Nakauchi, Gene
1335!! 9 QLD Stahnke, Alexander
1320!! 11 ACT Xing, Edward
1297!! 11 NSW Cheung, Benjamin
1289!! 11 QLD McGarity, Liam
1247! 11 VIC Thakur, Udit
1240!! 8 QLD Liu, Yi
1231! 10 QLD Zhu, Kevin
1227! 11 VIC Antolis, Cedric
1207!! 9 NSW Lau, Joshua
1167!! 11 NSW Ren, Jonathan
1161!! 9 VIC Matheson, Laurence

Bill Gletsos
01-09-2006, 12:42 AM
Top Under 10
1335!! 9 QLD Stahnke, Alexander
1240!! 8 QLD Liu, Yi
1207!! 9 NSW Lau, Joshua
1161!! 9 VIC Matheson, Laurence
1160!! 8 NSW Wang, Oscar
1148! 8 WA Sam, Ryan
1104!! 9 NSW Tsui, Edison
1036!! 9 TAS Horton, Vincent
980!! 9 QLD van den Bergh, Lachlan
967!! 9 ACT Zhang, Taiyang

Bill Gletsos
01-09-2006, 12:43 AM
Top Females Under 20
2166! 19 QLD Caoili, Arianne B [WFM]
1989!! 13 NSW Song, Angela [WFM]
1757!! 19 ACT Oliver, Shannon [WFM]
1647! 16 NSW Harris, Rebecca
1636!! 16 QLD Jule, Alexandra
1605!! 16 NSW Reid, Vaness
1429!! 14 QLD Kinder, Jessica
1419!! 11 VIC Yu, Sally
1415!! 10 ACT Guo, Emma
1359! 15 QLD Evans, Amy L
1314!! 16 QLD Lyons, Kieran C
1312!! 16 ACT Ikeda, Miona
1301!! 14 ACT Oliver, Tamzin L
1291!! 13 NSW Ng, Deborah
1235! 16 VIC Galiabovitch, Elena
1232!! 12 ACT Russell, Luthien
1231!! 12 SA Eustace, Sophie
1225! 13 NSW Soltysik, Adelaide
1194!! 12 QLD McGarity, Molly
1192!! 15 VIC Rozenblat, Vanja

Bill Gletsos
01-09-2006, 12:43 AM
Top Seniors
2197!! 68 NSW Flatow, A (Fred) [FM]
2173!! 65 SA Zaric, Srboljub
2141! 60 NSW Fuller, Maxwell L [FM]
2023! 60 NSW Wettstein, Marcus
1990! 62 NSW Capilitan, Romeo
1988!! 67 NSW Ghenzer, Charles
1980!! 61 NSW Degroen, Mark S
1971! 68 WA Partis, Michael T
1970! 71 NSW Jens, Henk W
1963!! 65 WA Ellis, David
1916! 70 NSW Purdy, John S
1911!! 78 NSW Viner, Phillip J
1889!! 69 NSW Hutchings, Frank P
1876!! 77 WA Leonhardt, Wolfgang
1864! 63 SA Mdinaradze, Edgar
1842! 67 NSW Benson, Lachlan
1839! 62 QLD Hindley, Simon
1823!! 61 QLD Lovejoy, David
1821! 70 QLD Mehltreter, Otto
1818! 70 TAS Gibbs, Glen B

Bill Gletsos
01-09-2006, 12:43 AM
Top Improvers
1098 285 NSW Papantoniou, John
1207 215 VIC Glasson, Daniel
1843 179 SA Sykes, Alexander
1415 168 ACT Guo, Emma
1035 145 QLD Hansen, Timothy
1240 134 QLD Liu, Yi
1114 125 VIC Liu, Nicholas
1214 121 VIC Somaskanthan, Aingaran
893 118 ACT McGlynn, Daniel
1643 118 NSW Nguyen, Joseph
588 118 QLD Kanagarajah, Abbie
967 116 ACT Zhang, Taiyang
1083 114 QLD Beilby, Kieton
1326 111 QLD Boyle, Gregory
1325 110 VIC Cameron, James
1213 108 QLD Lee, Leoma
1160 107 NSW Wang, Oscar
784 107 QLD Kinder, Danielle
1350 106 VIC Sunder, Vijay
731 105 QLD Lam, Alan
1391 102 VIC Miitel, Peter J
1850 102 VIC Morris, James
1236 101 SA Milton, Anthony
616 100 NSW Koh, Clarise
1947 92 NSW Chernih, Nicholas
1289 90 QLD McGarity, Liam
1256 86 QLD McLean, Matthew
1276 85 VIC Gan, Zhong Hao
1343 82 VIC Hogan, Steven
867 82 QLD Curtis, Shaun

Bill Gletsos
01-09-2006, 12:44 AM
Busiest Players
1907 53 VIC Voon, Richard
2031 48 NSW Bolens, Johny
1429 47 QLD Kinder, Jessica
1880 46 VIC Raine, Marcus
1929 43 NSW Illingworth, Max
2118 42 VIC Stojic, Dusan
944 42 QLD Eriksson, Caleb
1504 41 NSW Greenwood, Norman
2077 41 VIC Wallis, Christopher
1614 41 VIC Schon, Eugene
1398 41 VIC Potter, Michael
1419 40 VIC Yu, Sally
1303 39 NSW Kamaras, Con
1743 38 VIC Lycett, Garry
1309 38 VIC Potter, Daniel
1083 38 QLD Beilby, Kieton
1750 37 NSW Pickering, Anthony
1396 37 QLD Grigg, Sam
1850 36 VIC Stojic, Svetozar
1655 36 NSW Saksena, Kaushik
1649 36 VIC Szuveges, Narelle S
1956 36 VIC Dizdarevic, Mehmedalija
1790 36 WA Dunlop, Gordon
2315 35 NSW Xie, George
1415 35 ACT Guo, Emma
1194 35 QLD McGarity, Molly
1651 33 VIC Yu, Derek
1814 33 QLD Weller, Tony
1763 32 ACT Brown, Andrew
2092 32 NSW Wright, Neil S
2036 32 NSW Hu, Jason
2080 32 VIC Pyke, Malcolm L
1325 32 VIC Cameron, James
1444 32 VIC Potter, Christopher
1204 32 QLD Barrett, Daniel J (Somerset)
784 32 QLD Kinder, Danielle
1535 31 NSW Christensen, Joshua
1498 31 SA Zulfic, Fedja
1316 31 VIC Doon, Steven
1276 31 QLD Ford, Daniel

Bill Gletsos
01-09-2006, 12:44 AM
Category Players Percentage
2600..2699 1 0.04
2400..2499 5 0.24
2300..2399 10 0.64
2200..2299 17 1.32
2100..2199 33 2.64
2000..2099 74 5.59
1900..1999 75 8.59
1800..1899 122 13.46
1700..1799 188 20.97
1600..1699 230 30.16
1500..1599 207 38.43
1400..1499 251 48.46
1300..1399 231 57.69
1200..1299 201 65.72
1100..1199 149 71.67
1000..1099 134 77.03
900.. 999 108 81.34
800.. 899 96 85.18
700.. 799 80 88.37
600.. 699 72 91.25
500.. 599 61 93.69
400.. 499 52 95.77
300.. 399 50 97.76
200.. 299 36 99.20
100.. 199 20 100.00

Bill Gletsos
01-09-2006, 12:45 AM
Normal Tournaments Rated
ACT
19/07/2006 40 9 2006 ACT Junior Championship
23/07/2006 78 7 2006 ANU Open
07/07/2006 10 9 2006 Belconnen First Division
07/07/2006 10 9 2006 Belconnen Premier
07/07/2006 10 9 2006 Belconnen Second Division
07/07/2006 14 4 2006 Swanson Sprint
02/06/2006 17 5 2006 Voldemort Goblet
02/08/2006 10 7 ANU Winter Tournament
02/08/2006 12 7 Cold Comfort Cup 2006
NSW
12/06/2006 44 7 2006 NSW Open Champs
12/06/2006 40 7 2006 NSW Open U1600
25/07/2006 34 10 2006 NSWCA Grade Matches Open
20/07/2006 74 9 2006 NSWCA Grade Matches U1400
20/07/2006 69 9 2006 NSWCA Grade Matches U1600
26/07/2006 38 10 2006 NSWCA Grade Matches U1800
18/07/2006 26 9 2006 NSWCA Grade Matches U2000
06/06/2006 8 6 2006 Wollongong KO
11/04/2006 8 2 2006 Wollongong KO Preliminary 1
25/04/2006 16 2 2006 Wollongong KO Preliminary 2
06/06/2006 12 6 2006 Wollongong KO Swiss
19/05/2006 14 6 Albury Rated Tournament
02/07/2006 60 7 CJS Purdy Centenary Tournament 2006
25/07/2006 10 7 Central Coast District Championship 2006
01/08/2006 14 1 Central Coast Lagues CC Versus Coast 2006
06/08/2006 27 7 Coal City Open 2006
09/07/2006 22 7 Fischerbooks Over 40's Tournament 2006
03/07/2006 17 9 Hakoah Cup 2006
19/06/2006 2 3 Manly Club Champs C Grade 2006 Play-off
14/07/2006 52 8 NSW JUNIOR CHAMPIONSHIP - UNDER 12
14/07/2006 50 8 NSW JUNIOR CHAMPIONSHIP - UNDER 18
30/07/2006 51 7 NSWCA July Weekender 2006
08/08/2006 16 7 Wgong 2006 Grading Swiss
QLD
29/06/2006 21 5 2006 BCC Autumn Swiss
25/05/2006 20 4 2006 BCC Lasker Memorial
27/07/2006 21 4 2006 BCC Tal Memorial
26/06/2006 26 7 2006 Gold Coast Open
26/06/2006 56 7 2006 Gold Coast Under 1600
06/08/2006 13 7 2006 NQ Open
12/06/2006 8 7 2006 Queensland Championship
30/06/2006 22 8 2006 Queensland Junior Championship
12/06/2006 44 7 2006 Queensland Reserve Championship
19/04/2006 16 7 2006 Tal Memorial
28/02/2006 19 6 2006 Townsville Summer Swiss
12/06/2006 13 8 Brisbane Young Champions
20/08/2006 10 5 Gold Coast Adults Only 2006
07/06/2006 38 8 Gold Coast Club Championship 2006
07/05/2006 48 7 Gold Coast Junior Championships 2006
12/06/2006 11 8 Gold Coast Junior Masters T2 A 2006
12/06/2006 12 8 Junior Masters T2 2006
30/07/2006 8 7 Mackay Open 2006
20/08/2006 12 5 Qld Vets and Disabled Chps 2006
09/07/2006 10 7 Queensland Girl's Championship 2006
06/06/2006 17 6 SCC Memorial Challenge
12/06/2006 10 9 Super Junior Masters T2 2006
18/04/2006 6 5 Townsville Autumn Round Robin A Grade
18/04/2006 6 5 Townsville Autumn Round Robin B Grade
18/04/2006 6 5 Townsville Autumn Round Robin C Grade
29/05/2006 18 5 Townsville Winter Swiss 2006
SA
09/07/2006 44 7 2006 Lidums Checkmate Open
22/08/2006 8 7 City of Adelaide Major 2006
22/08/2006 19 7 City of Adelaide Minor 2006
27/06/2006 83 9 Interclub 2006
11/06/2006 17 6 Queen'S Birthday 2006
29/04/2006 31 8 SA State Junior U18 Championships - 2006
TAS
28/08/2006 15 15 2006 HICC Championship
14/06/2006 17 17 Burnie Club Championship 2006
12/06/2006 25 6 Tasmanian Open 2006
VIC
23/06/2006 64 9 2006 Box Hill Club Championship
26/06/2006 40 9 2006 City of Melbourne Open
25/06/2006 54 7 2006 Victorian Junior Under 12 Championship
29/06/2006 38 9 2006 Victorian Junior Under 18 Championship
12/06/2006 68 7 2006 Victorian Open
20/08/2006 16 15 Ballarat Club Championship
09/03/2006 13 7 Ballarat Spielvogel Memorial
22/08/2006 10 9 Box Hill FIDE Rated Tournament
26/08/2006 6 5 Box Hill Tuesday Grades Tournament (A Grade)
22/08/2006 6 5 Box Hill Tuesday Grades Tournament (B Grade)
22/08/2006 7 5 Box Hill Tuesday Grades Tournament (C Grade)
22/08/2006 13 5 Box Hill Tuesday Grades Tournament (D Grade)
20/06/2006 29 7 Box Hill Winter Swiss
15/06/2006 14 5 Croydon Fischer Tournament
24/08/2006 12 5 Croydon Kasparov Tournament
28/06/2006 56 7 Dandenong Autumn Open
30/08/2006 10 9 Dandenong Grades Tournament (A Grade)
30/08/2006 10 9 Dandenong Grades Tournament (B Grade)
30/08/2006 10 9 Dandenong Grades Tournament (C Grade)
30/08/2006 15 9 Dandenong Grades Tournament (D Grade)
30/08/2006 10 9 Dandenong Grades Tournament (Premier Grade)
05/06/2006 18 7 Hobsons Bay/Yarraville Cup
05/08/2006 12 11 Latrobe Valley Club Championship
21/08/2006 12 7 Melbourne Under 1600
21/08/2006 16 7 Melbourne Under 2000
15/06/2006 11 7 Mentone Open
WA
27/08/2006 46 9 2006 WA State Championship
04/06/2006 8 5 Foundation Day 2006
30/05/2006 14 7 Fremantle Open 2006
14/07/2006 4 3 Huntingdale Challengers 2006 Part 2
26/06/2006 4 3 Huntingdale Challengers 2006 part 1
07/07/2006 10 9 Huntingdale Masters 2006
06/07/2006 36 7 Metro Open 2006

Bill Gletsos
01-09-2006, 12:46 AM
Rapid Tournaments Rated
ACT
18/06/2006 113 7 2006 ACT Primary Allegro
06/08/2006 63 9 2006 ACT Under 10 Championship
27/08/2006 76 9 2006 ACT Under 12 Championship
25/06/2006 35 7 2006 ACT Under 8 Championship
14/06/2006 13 7 2006 Autumn Rapid
12/07/2006 35 7 2006 Chess Chicks Championship
30/06/2006 33 8 2006 Norths Juniors Term 2 Rapid
29/06/2006 36 9 2006 Souths Juniors Term 2 Rapid
01/04/2006 12 7 Street Chess 1 April 2006
01/07/2006 10 7 Street Chess 1 July 2006
10/06/2006 16 7 Street Chess 10 June 2006
15/07/2006 10 7 Street Chess 15 July 2006
17/06/2006 14 7 Street Chess 17 June 2006
23/03/2006 14 7 Street Chess 23 March 2006
24/06/2006 12 7 Street Chess 24 June 2006
06/05/2006 10 7 Street Chess 6 May 2006
08/04/2006 14 7 Street Chess 8 April 2006
08/07/2006 10 7 Street Chess 8 July 2006
NSW
14/07/2006 27 9 2006 Australian Uni Rapid
13/08/2006 58 7 2006 NSW Teams Challenge Individual 2
08/08/2006 48 9 2006 St. George Allegro
29/08/2006 31 9 2006 St. George Rapid
25/06/2006 9 9 CommonMan 2006
03/06/2006 10 5 Ettalong Coldie 2006
26/08/2006 8 7 Ettalong Deep Winter Chill`2006
26/08/2006 6 5 Ettalong Deep Winter Chill`2006 div A
08/07/2006 10 7 Ettalong Wet and Coldie 2006
10/07/2006 12 11 Liverpool John Edmondson Memorial Rapid RR 1
10/07/2006 12 11 Liverpool John Edmondson Memorial Rapid RR 2
30/08/2006 73 9 Metropolitan Leagues Rapid Play 2006
16/07/2006 40 7 NSW Rapid Championship 2006
24/08/2006 15 7 Parramatta August Rapid 2006
08/08/2006 7 7 Smithfield Rapid Championship Part 1
08/08/2006 7 7 Smithfield Rapid Championship Part 2
QLD
25/05/2006 78 8 2006 Brisbane High Schools Open Division Round 2
24/05/2006 119 7 2006 Brisbane Primary Schools Open A Grade Round 2
02/08/2006 28 7 2006 Central Qld High Schools Championship Rd 2 Premier
01/07/2006 47 6 2006 Day Of Knights Rapid
30/07/2006 39 7 2006 Fastrain Under 10 Tournament
30/07/2006 28 7 2006 Fastrain Under 12 Tournament
30/07/2006 19 7 2006 Fastrain Under 14 Tournament
30/07/2006 21 7 2006 Fastrain Under 18 Tournament
29/06/2006 23 8 2006 Queensland Junior Under 10
29/06/2006 19 8 2006 Queensland Junior Under 12
27/06/2006 15 9 2006 Queensland Under 8 Championship
01/08/2006 32 7 2006 Wide Bay Burnett High Schools Championship Rd 2 Premier
13/08/2006 52 7 Alex Jess Rapid
21/06/2006 101 8 Gold Coast High Schools Teams Chess R2
21/06/2006 48 7 Gold Coast High Schools Teams R2 Championship Individual
31/05/2006 98 8 Gold Coast Primary Schools Teams Central A Grade R2
01/06/2006 48 7 Gold Coast Primary Schools Teams Championship Individuals R2
30/05/2006 65 7 Gold Coast Primary Schools Teams North A Grade R2
01/06/2006 88 8 Gold Coast Primary Schools Teams South A Grade R2
28/05/2006 63 6 Gold Coast Rapid Championship 2006
03/07/2006 16 8 July Camp 030706adv
03/07/2006 19 8 July Camp 030706int
04/07/2006 19 9 July Camp 040706adv
04/07/2006 17 9 July Camp 040706int
05/07/2006 17 8 July Camp 050706adv
05/07/2006 19 8 July Camp 050706int
07/07/2006 16 5 July Camp 070706adv
07/07/2006 19 5 July Camp 070706int
06/07/2006 131 8 Runaway Bay Tourney June 2006
18/08/2006 31 5 Somerset Junior House 2006
11/08/2006 34 5 Somerset Senior House 2006
26/03/2006 17 7 Sunshine Coast Womens & Girls 2006
02/06/2006 15 4 fr020606adv
04/08/2006 9 5 fr040806adv
04/08/2006 12 4 fr040806int
09/06/2006 6 4 fr090606adv
09/06/2006 11 4 fr090606int
11/08/2006 9 5 fr110806adv
11/08/2006 12 5 fr110806int
14/07/2006 11 4 fr140706adv
14/07/2006 8 4 fr140706int
18/08/2006 11 4 fr180806adv
18/08/2006 10 4 fr180806int
21/07/2006 10 4 fr210706adv
21/07/2006 12 4 fr210706int
28/07/2006 10 4 fr280706adv
28/07/2006 11 4 fr280706int
16/06/2006 8 5 fri160606adv
16/06/2006 9 5 fri160606int
23/06/2006 11 5 fri230606adv
05/06/2006 9 4 mon050606adv
05/06/2006 9 4 mon050606int
07/08/2006 16 4 mon070806adv
07/08/2006 8 4 mon070806int
10/07/2006 15 5 mon100706adv
14/08/2006 15 4 mon140806adv
14/08/2006 12 4 mon140806int
17/07/2006 16 4 mon170706adv
17/07/2006 11 4 mon170706int
19/06/2006 13 4 mon190606adv
19/06/2006 8 4 mon190606int
24/07/2006 17 4 mon240706adv
24/07/2006 10 4 mon240706int
29/05/2006 14 4 mon290506adv
29/05/2006 12 4 mon290506int
31/07/2006 16 4 mon310706adv
31/07/2006 10 4 mon310706int
01/06/2006 11 4 th010106adv
01/06/2006 12 4 th010606int
03/08/2006 14 4 th030806adv
03/08/2006 14 4 th030806int
08/06/2006 13 4 th080606adv
08/06/2006 10 4 th080606int
10/08/2006 14 4 th100806adv
10/08/2006 13 4 th100806int
13/07/2006 10 4 th130706adv
13/07/2006 12 4 th130706int
15/06/2006 11 4 th150606adv
15/06/2006 11 4 th150606int
17/08/2006 14 4 th170806adv
17/08/2006 12 4 th170806int
20/07/2006 11 4 th200706adv
20/07/2006 14 4 th200706int
22/06/2006 8 4 th220606adv
22/06/2006 13 4 th220606int
27/07/2006 14 4 th270706adv
27/07/2006 14 4 th270706int
01/08/2006 6 4 tu010806adv
01/08/2006 7 4 tu010806int
06/06/2006 14 4 tu060606adv
15/08/2006 14 4 tu150806adv
18/07/2006 8 4 tu180706adv
18/07/2006 8 4 tu180706int
25/07/2006 8 4 tu250706adv
25/07/2006 7 4 tu250706int
30/05/2006 8 4 tu300506adv
30/05/2006 11 4 tu300506int
13/06/2006 8 4 tue130606adv
13/06/2006 11 4 tue130606int
08/08/2006 8 4 tues080806adv
08/08/2006 7 4 tues080806int
11/07/2006 9 5 tues110706adv
20/06/2006 13 4 tues200606adv
02/08/2006 14 4 wed020806adv
02/08/2006 13 4 wed020806int
07/06/2006 16 4 wed070606adv
07/06/2006 12 4 wed070606int
09/08/2006 16 4 wed090806adv
09/08/2006 11 4 wed090806int
12/07/2006 14 4 wed120706adv
12/07/2006 8 4 wed120706int
14/06/2006 11 4 wed140606adv
14/06/2006 18 4 wed140606int
16/08/2006 17 4 wed160806adv
19/07/2006 21 4 wed190706adv
19/07/2006 11 4 wed190706int
21/06/2006 12 4 wed210606adv
21/06/2006 10 4 wed210606int
26/07/2006 17 4 wed260706adv
26/07/2006 13 4 wed260706int
31/05/2006 12 5 wed310506adv
31/05/2006 8 5 wed310506int
SA
29/04/2006 18 10 SA State Junior U12 Championships - 2006
04/07/2006 23 5 Winter Allegro 2006
VIC
16/06/2006 64 7 Boroondara Secondary Interschool Term 2 Championship
29/08/2006 69 7 Boroondara Secondary Interschool Term 3 Championship
21/07/2006 32 8 Box Hill Rapidplay Tournament
06/08/2006 35 7 Canterbury Rookies & Queens Cup (August)
09/07/2006 47 7 Canterbury Rookies & Queens Cup (July)
23/05/2006 34 6 Canterbury Term 2 Junior Coaching Group Tournament
27/07/2006 41 8 Dark Horse Junior Club Championship
12/08/2006 61 9 Waverley-Monash Primary Interschool Championship

Bill Gletsos
01-09-2006, 01:06 AM
Some regular BB posters ratings:

Ian Rogers - 28 2613!!
Igor Goldenberg - 14 2345!!
George Xie - 35 2315!!
Leonid Sandler - 0 2311!
Michael Baron - 0 2282!
jeffrei - 0 2278?
Greg Canfell - 0 2276!!
drug - 6 2256!!
Brett Tindall- 1 2207!
Tristan Boyd - 18 2187!!
Comrade Zukovsky - 6 2179!
Brian Jones - 3 2120!!
Ronald Yu - 7 2106!!
Gareth Charles - 14 2095!!
Andrew Bird - 20 2091!!
macavity - 32 2080!!
paulb - 7 2075!!
D Dragicevic - 22 2045!!
Laura Moylan - 25 2044!!
Jason Hu - 32 2036!!
bobby1972 - 15 2025!!
Goughfather - 15 2010!!
Lee Jones - 0 1999!!
Kerry Stead - 0 1999!!
Jason Chan - 7 1968!!
Kevin Bonham - 19 1966!!
Chernz - 15 1947!!
firegoat7 - 0 1942!!
rob - 7 1897!!
pballard - 5 1869!
EZbeet - 14 1803!
GeorgeL - 21 1792!!
elevatorescapee - 0 1785!
Boris - 8 1778!
Barry Cox - 21 1752!!
PhilD707 - 16 1747!!
Ian Rout - 21 1742!!
pax - 0 1739?
Amiel Rosario - 0 1715!!
DavidFlude - 23 1703!!
Shaun Press - 7 1662!
The_Wise_Man - 5 1651!!
Belthaser - 0 1647
JGB - 0 1640
David Richards - 0 1639
Candy Cane - 16 1605!!
Dozy - 1 1565!!
starter - 0 1565!
heaviestknight - 24 1544!!
Liberace - 0 1541!!
skip - 0 1534
altecman - 30 1514!!
antichrist - 0 1492?
Frosty - 8 1485!
Phil Bourke - 7 1470!
Trent Parker - 29 1469!!
Howard Duggan - 7 1460!!
Arrogant-One - 7 1435!
bergil - 30 1405!!
Scott Colliver - 5 1404
themovingman - 8 1397!!
Paul Sike - 8 1393!!
AES - 7 1328!
Careth - 0 1325!!
Garvin Gray - 18 1279!!
PHAT - 14 1272!!
EGOR - 3 1242!!
qpawn - 7 1222!
watto- 16 1160!!
alana - 16 1033!!
bunta - 0 1021

The_Wise_Man
01-09-2006, 09:36 AM
Bill, this must be one of your quickest ratings notifications yet!

Well done! :clap: :clap: :clap:

Alana
01-09-2006, 04:09 PM
:P :) :P YAY!!! MY RATING IS OVER 1000!!! :P :owned: (finally... :eek: ) but my rapid only went up 3 points... :doh: :wall:

Bill Gletsos
02-09-2006, 12:35 AM
September ratings are up on the ACF website.

bergil
02-09-2006, 12:35 AM
September ratings are now up on the ACF website.About bloody time! :P (1day late)

WhiteElephant
02-09-2006, 11:07 AM
September ratings are up on the ACF website.

When I click on Ratings, I still get the June ones. Am I looking in the wrong place?

Desmond
02-09-2006, 11:15 AM
Category Players Percentage
2600..2699 1 0.04
2400..2499 5 0.24
2300..2399 10 0.64
2200..2299 17 1.32
2100..2199 33 2.64
2000..2099 74 5.59
1900..1999 75 8.59
1800..1899 122 13.46
1700..1799 188 20.97
1600..1699 230 30.16
1500..1599 207 38.43
1400..1499 251 48.46
1300..1399 231 57.69
1200..1299 201 65.72
1100..1199 149 71.67
1000..1099 134 77.03
900.. 999 108 81.34
800.. 899 96 85.18
700.. 799 80 88.37
600.. 699 72 91.25
500.. 599 61 93.69
400.. 499 52 95.77
300.. 399 50 97.76
200.. 299 36 99.20
100.. 199 20 100.00


Not a bad looking bell graph there. The only blip being the 1500-1599 category.

Basil
02-09-2006, 12:30 PM
When I click on Ratings, I still get the June ones. Am I looking in the wrong place?

You may need to hit "refresh" on your browser to clear the cache - especially if you had been looking at the July ratings list recently.

WhiteElephant
02-09-2006, 07:03 PM
You may need to hit "refresh" on your browser to clear the cache - especially if you had been looking at the July ratings list recently.

Aaah yes, thanks.

Wow, I lost nearly 100 rating points after coming =3rd in a tournament where I was seeded 2nd.

Bill Gletsos
02-09-2006, 07:03 PM
The CV Ratings officer has informed me that the Croydon Kasparov Tournament was submitted with an incorrect result and some forfeits coded as actual results.

I have rerun the ratings and the only players affected with their corrected normal ratings are the following CV players.

2147 5 Anderson, Tim
1645 16 Braham, Peter
1340 12 Frost, Jared
1482 7 Frost, Stephen
1210 7 Gabrys, Marek
2g 4 Gardiner, Jason
1398 5 Jaske, Jan
1289 11 Kloprogge, Ion
1649 4 Lacey, David
898 4 Sawkins, Cliff
1440 10 Smith, Mike
1742 10 Waller, A (Tony)

Basil
02-09-2006, 10:40 PM
Lee Jones - 0 1999!!
Kerry Stead - 0 1999!!
I think being a point shy of the double millenium would give me the irits.

Bill Gletsos
02-09-2006, 10:44 PM
I think being a point shy of the double millenium would give me the irits.It does however allow them to enter U2000 events. ;)

Basil
02-09-2006, 10:49 PM
It does however allow them to enter U2000 events. ;)
25 HCDs !! [not a classic post] but certainly gets "best post of the hour" prize :)

Note the double exclamation mark after the HCDs. Means that's a realiable marking!

MichaelBaron
03-09-2006, 04:53 PM
Aaah yes, thanks.

Wow, I lost nearly 100 rating points after coming =3rd in a tournament where I was seeded 2nd.

You got the Australian rating system to thank for that :D

Rincewind
03-09-2006, 05:15 PM
You got the Australian rating system to thank for that :D

I think George's effort were at least partially to blame. I mean had he come 1st he might have even won some rating points. :)

MichaelBaron
03-09-2006, 06:28 PM
I think George's effort were at least partially to blame. I mean had he come 1st he might have even won some rating points. :)


It should be impossible to lose 100 points in one event. For example if i play in a fide rated event with an average rating of 2100 or so, i am expected to score 6.5/9 or so to preserve my fide rating. If i "want" to lose 100 points ..i have to score 0/9!

Anyway...there have been enough discusion of our "glorious" rating system on this site..lets not start yet another debate about it :)

Rincewind
03-09-2006, 07:19 PM
It should be impossible to lose 100 points in one event. For example if i play in a fide rated event with an average rating of 2100 or so, i am expected to score 6.5/9 or so to preserve my fide rating. If i "want" to lose 100 points ..i have to score 0/9!

The FIDE system actually tries to slow down rating movements with arbitrary K factors based on rating. Iagree that is probably would not happen under the FIDE system, but I don't accept that that means the FIDE system is objectively any better.

The first step in such a debate would be to agree to the objectives of the rating system. Has this ever been done? What are the objectives of the ACF rating system? Once they are agreed to by everyone then we could discuss what is good and bad in a meaningful way.

For example, if the objective was to keep everyone happy then a system where people never lose points would be good. However, I think the objective if to provide an as-predictive-as-possible measure of relative chess playing strength. Then the Glicko system makes more sense.


Anyway...there have been enough discusion of our "glorious" rating system on this site..lets not start yet another debate about it :)

I didn't bring it up. I was just pointing out it is easy to blame the rating system, however, no one loses rating points from winning a game. ;)

Bill Gletsos
03-09-2006, 07:56 PM
I think George's effort were at least partially to blame. I mean had he come 1st he might have even won some rating points. :)George hadnt played a rated game since 1997 until he played in the March 2005 rating period. In that period he performed at the 1995 level and his 1890?? rating went to 1942. He hadnt played since then and at the start of this rating period his rating was 1942?. In this period he scored 5.5 from 9 and performed at the 1801 level and the system dropped his rating to 1855.

Oepty
04-09-2006, 05:58 PM
While the rating system does cause players to lose a lot of points if they play well below their previous level on their return to chess, it does allow them to gain a lot of points back if they start playing well again. Alex Sykes is an expample of this on the rating list this time. He lost a lot of points because he was very rusty when returned after a break but his return to form in the Interclub and City of Adelaide Major have been rewarded with a 179 point increase. Having a more responsive rating system works both ways.
Scott

Kevin Bonham
04-09-2006, 06:47 PM
Scott, I see there is a Colin Cloudsdale listed from SA in the latest rating list. A player by that name won the 1972 and 73 Tas Champs and tied for first in the 76 Tas Open with Haydn Barber and James Henri. In 1978 he tied for first in the Tas Champs tournament but wasn't eligible for the title on residency grounds. I'm assuming it's the same player. If so his involvement in Tassie chess was before my time but he is frequently mentioned as an outstanding player by players of that era.

Rhubarb
04-09-2006, 07:11 PM
While the rating system does cause players to lose a lot of points if they play well below their previous level on their return to chess, it does allow them to gain a lot of points back if they start playing well again.This could well be the most concise explanation ever from a non ratings officer. Thank you, Scott, this is exactly what happens.

peter_parr
06-09-2006, 03:06 PM
It should be impossible to lose 100 points in one event. For example if i play in a fide rated event with an average rating of 2100 or so, i am expected to score 6.5/9 or so to preserve my fide rating. If i "want" to lose 100 points ..i have to score 0/9!

In a 2 day Sydney weekender in July 2006 RGH rated 1860? Scored 3.5/7. His expected score was 4.5/7. He lost 94 rating points (he would have lost 15 rating points on a FIDE rating system).

In the same event IPP rated 2266? Score 4/7.
He lost 217 rating points (he would have lost 39 rating points on a FIDE rating system).

All players with a ?? instead of ? would lose considerably more.

Only Australia out of 150 FIDE countries uses the Glicko rating system.

A democratic referendum (Gletsos one vote) would I am sure reveal that the clear majority of players would prefer to use a system similar to the FIDE rating system as used world wide. (No penalty for inactivity).

Gletsos opinion is well known.

When a player returns to competitive chess and finds he loses over 200 rating points in 2 days we are risking losing him altogether.

Only last week another inactive player over 2000 rated sold me his entire extensive library – sad but true.

arosar
06-09-2006, 03:17 PM
Only last week another inactive player over 2000 rated sold me his entire extensive library – sad but true.

Pete, first of all, thanks for selling me those old CIAs. I love them.

Now here is a question. Did the inactive player sell you his library because he was unhappy with the rating system?

Anyway, I'm not too au fait with the mathematics of Glicko - but to a fair-minded bloke like me, so long as the behaviour of the rating system is symmetrical (for both good and bad performances), then it should be OK. All RGH and IPP have to do is basically play well again to recover their lost points.

Anyone know what the required performances have to be for IPP and RGH to recover the points?

Cheers,

AR

bergil
06-09-2006, 03:33 PM
I lost 150 points thereabouts one comp by giving queen odds to juniors - give me 135 back and we will call squareHow about a slap in the back of the head for being so stupid and call it even? :P

Bill Gletsos
06-09-2006, 03:45 PM
How strong a player was 8 years ago isnt relevant. It is how strong he is now.

RGH has a performance rating of around 1633. Prior to this event he last played in June 2000 when he played 6 games. Prior to that he had previously played in December 1996.
IPP had a performance rating of around 1863. Prior to this event he last played in in Dec 2004 when he played 6 games. Prior to that was 1 game in Dec 2001 and prior to that back in October 1998.

bergil
06-09-2006, 04:21 PM
Only Australia out of 150 FIDE countries uses the Glicko rating system. Of the 150 countries how many use ELO? And can you please list them and cite a reference to them all?


When a player returns to competitive chess and finds he loses over 200 rating points in 2 days we are risking losing him altogether. If a player is out as long as the 2 you cited did we have them? Richard Gastineau-Hills 6 games in a decade and Ian P Parsonage 7 games in 8 years before the CJS Purdy. :hmm:

Bill Gletsos
06-09-2006, 04:37 PM
In fact having a closer look at IPP, prior to that 1 game in Dec 2001 he had only played one game in October 1998 since October 1994. In 1994 he played a total of 10 games.
Interestingly that means that under the ACF Elo system he played 11 games in 7 years (1994-2000), and under Glicko he has played 14 games (Dec 2000-sep 2006).

bergil
06-09-2006, 04:56 PM
In fact having a closer look at IPP, prior to that 1 game in Dec 2001 he had only played one game in October 1998 since October 1994. In 1994 he played a total of 10 games.
Interestingly that means that under the ACF Elo system he played 11 games in 7 years (1994-2000), and under Glicko he has played 14 games (Dec 2000-sep 2006).Hang on does this mean under the dodgy argument of Mr Parr, that Glicko has attracted Ian back to chess based on his games played under ELO (10) compared to games played under Glicko (14)? :owned:

Brian_Jones
06-09-2006, 05:46 PM
Only last week another inactive player over 2000 rated sold me his entire extensive library – sad but true.

What! You bought both books? :cool:

bergil
06-09-2006, 05:51 PM
What! You bought both books? :cool:Got him coming and going ;)

Oepty
06-09-2006, 05:55 PM
Scott, I see there is a Colin Cloudsdale listed from SA in the latest rating list. A player by that name won the 1972 and 73 Tas Champs and tied for first in the 76 Tas Open with Haydn Barber and James Henri. In 1978 he tied for first in the Tas Champs tournament but wasn't eligible for the title on residency grounds. I'm assuming it's the same player. If so his involvement in Tassie chess was before my time but he is frequently mentioned as an outstanding player by players of that era.

Kevin, I am fairly sure this would be the same player. It certainly is the same player who was one of SA's top players in the early eighties. He did have a master list rating well over 2100 before he returned to playing a few years ago.
Scott

Rincewind
06-09-2006, 06:02 PM
In a 2 day Sydney weekender in July 2006 RGH rated 1860? Scored 3.5/7. His expected score was 4.5/7. He lost 94 rating points (he would have lost 15 rating points on a FIDE rating system).

In the same event IPP rated 2266? Score 4/7.
He lost 217 rating points (he would have lost 39 rating points on a FIDE rating system).

All players with a ?? instead of ? would lose considerably more.

Only Australia out of 150 FIDE countries uses the Glicko rating system.

A democratic referendum (Gletsos one vote) would I am sure reveal that the clear majority of players would prefer to use a system similar to the FIDE rating system as used world wide. (No penalty for inactivity).

Gletsos opinion is well known.

When a player returns to competitive chess and finds he loses over 200 rating points in 2 days we are risking losing him altogether.

Only last week another inactive player over 2000 rated sold me his entire extensive library – sad but true.

This is merely a few of argumenta ad populum cobbled together. The number of countries which use Elo is not relevant (unless the objective is for Australia to be like everywhere else) and what the majority of players want is also not relevant (unless keeping players happy is an objective).

This debate can never advance unless the protagonists can agree to a set of objectives for the ACF rating system.

Garvinator
06-09-2006, 06:16 PM
Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

antichrist
07-09-2006, 03:51 PM
Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

But that is not really a relevant point. If returning players kept their old rating for longer when opposing players won against them would not they gain more points because from a higher player?

Bill Gletsos
07-09-2006, 04:28 PM
But that is not really a relevant point. If returning players kept their old rating for longer when opposing players won against them would not they gain more points because from a higher player?But that gain is not justified as the returning players ratings is not a true measure of their current strength.
As such the one not making a relevant point is you. ;)

MichaelBaron
07-09-2006, 05:16 PM
Lets resolve the debate once and for all: how about an national referendum whether we want a Glicko system or an elo-like system.:hmm:

arosar
07-09-2006, 05:29 PM
Lets resolve the debate once and for all: how about an national referendum whether we want a Glicko system or an elo-like system.:hmm:

What, you think the general population will know what they're talking about to make an informed choice? It's a very complex issue, you know. This is one of those things that I much prefer to leave to experts. So long as they are communicative and are ultimately accountable, what's the big deal?

Just play chess.

AR

Alan Shore
07-09-2006, 05:54 PM
Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

I wonder what the record is for most points gained in a tournament? Surely I'd have to be close.

Kevin Bonham
07-09-2006, 07:17 PM
Lets resolve the debate once and for all: how about an national referendum whether we want a Glicko system or an elo-like system.:hmm:

This is one of those issues where nine tenths of players won't care and eight tenths of those who care won't understand sufficiently to cast an informed vote. As such spending money on a referendum would be absolutely pointless.

I'll make a post discussing what in my view should be the exact objectives of a rating system (I agree with Rincewind that agreement on this would further the debate, though I doubt that it will ever happen) sometime when I have more time. However to foreshadow that slightly, the primary aim of a rating system must be accurate prediction of player performance and it is a fact that the Glicko system is a vastly superior predictor to ELO.

(I use the word "must" because this is axiomatic IMO. If the main aim of a "rating system" isn't accurate prediction of performance then it's not a rating system but a statistical ego-massage unit.)

Oepty
07-09-2006, 07:30 PM
Barry, Kevin. The objective of the ACF rating system is stated in the ACF constitution. In section 2 (statement of purposes) vii it states,

to maintain a national rating system to enable comparison of the chess
playing abilities of Australian chess players;


Scott

Basil
07-09-2006, 08:58 PM
... because this is axiomatic
Yellow Card

The objective of the ACF rating system is stated in the ACF constitution. In section 2 (statement of purposes) vii it states,to maintain a national rating system to enable comparison of the chess playing abilities of Australian chess players
But Scott, isn't the Constitution satisfied in that a ratings system is in place? We have now turned to the issue of which is a better system [which is not the concern of the Constitution].

Desmond
07-09-2006, 09:03 PM
Yellow CardAt least he didn't say "metric" ;)

Alan Shore
07-09-2006, 11:40 PM
At least he didn't say "metric" ;)

I miss starter.

Bereaved
08-09-2006, 12:02 AM
I miss starter.

I saw him tonight, Belthasar, and he seemed quite well,

Take care and God Bless, Macavity

arosar
08-09-2006, 09:42 AM
For those who want to say hi to starter, just go and visit Matt's board. But you can provoke him into making a post here by complaining against his dearest beloved club's practice of "divisionalised intermingling SWISS".

AR

Kevin Bonham
08-09-2006, 11:10 AM
Barry, Kevin. The objective of the ACF rating system is stated in the ACF constitution. In section 2 (statement of purposes) vii it states,

to maintain a national rating system to enable comparison of the chess
playing abilities of Australian chess players;

Yes, but that's far too vague to help you decide between the merits of different proposed systems.

Kevin Bonham
08-09-2006, 11:15 AM
For those who want to say hi to starter, just go and visit Matt's board. But you can provoke him into making a post here by complaining against his dearest beloved club's practice of "divisionalised intermingling SWISS".

Just don't quote from or link to any anonymous sprays that might have been written by Happy Friend. (I won't punish anyone who does so innocently but I will certainly delete the post).

Bill Gletsos
09-09-2006, 12:38 AM
I've been informed that the SA State Junior U18 Championships was actually a rapid, not a normal.

Oepty
09-09-2006, 04:09 PM
I've been informed that the SA State Junior U18 Championships was actually a rapid, not a normal.

Bill. As far as I know whoever informed you is wrong.
Scott

Oepty
09-09-2006, 04:42 PM
Yes, but that's far too vague to help you decide between the merits of different proposed systems.

Kevin. Yes it is vague, but it is an objective. I would class your proposed objective of a rating system as more of a test for a rating system. The rating system should show how players have been performing in the recent past. The fact that most players performances do not jump or fall by 100s points in a week means that prediction of performances is a good test and most likely the best test of whether the rating system is doing its job.
Scott

Bill Gletsos
09-09-2006, 04:56 PM
Bill. As far as I know whoever informed you is wrong.
ScottAlan Goldsmith told me.

Oepty
09-09-2006, 05:02 PM
Alan Goldsmith told me.

Okay. Then I was wrong sorry
Scott

Denis_Jessop
09-09-2006, 08:50 PM
Yes, but that's far too vague to help you decide between the merits of different proposed systems.

Nor is it intended or appropriate that it should. Either the Glicko or ELO systems would satisfy the Constitutional but to write either of them into the Constitution would be pretty inadvisable hwving regard to how hard it is to change it.

In an earlier post, Kevin, you said this:


the primary aim of a rating system must be accurate prediction of player performance and it is a fact that the Glicko system is a vastly superior predictor to ELO.

I have a little nit pick with it in that a rating system can only be an indicator of performance, not an accurete prediction, as it is based on existing data. Even though it may be able to be used as a predictor of possible future performance I doubt if any system could be an accurate predictor of future performance. This may not be as nit-picky as I first thought, either, as it may explain why a lot of people have problems with a rating system of any kind - they expect it to reveal with certainty the performance of their opponents when it is incapable of doing so.

DJ

Basil
09-09-2006, 09:05 PM
In an earlier post, Kevin, you said this:


the primary aim of a rating system must be accurate prediction of player performance and it is a fact that the Glicko system is a vastly superior predictor to ELO.
and then Denis, you just said this

I have a little nit pick with it in that a rating system can only be an indicator of performance, not an accurete prediction

And now I have a nitpick. Perhaps Kevin was not as tight with his words as he could have been and he in fact meant 'indicator' instead or 'predictor'. I appreciate Mother can speak for himself; it just seems your nitpick was premature.

Kevin Bonham
12-09-2006, 04:05 AM
I can use "model" instead of "predict" if that helps any.

What I meant is simply that the best rating system ideally (subject to various caveats which I'll list sometime) is the one which does the most accurate job of predicting how a player will perform if they play within a few months of the list coming out. Any system will have a rather large degree of error simply because performances over few games are unpredictable, but all other things being equal the one with the lowest average prediction error is the best system.

The reason I emphasise prediction over a summary of past performance is that the purpose of a rating should be to tell you how good a player is now (which means how good would they be if they played in a tournament immediately) and not how good a player was in the past, though understanding the latter is essential to trying to model the former.

MichaelBaron
12-09-2006, 04:48 PM
What are the chances of our rating system being changed? Is there any possibility at all Australia will hav a NORMAL rating system in the near future.:hmm:

Garvinator
12-09-2006, 04:52 PM
What are the chances of our rating system being changed? Is there any possibility at all Australia will hav a NORMAL rating system in the near future.:hmm:
what do you mean by normal?

Vlad
12-09-2006, 06:18 PM
what do you mean by normal?

I think a short answer to your question is the one that Michael likes.:)

Denis_Jessop
12-09-2006, 09:50 PM
I can use "model" instead of "predict" if that helps any.

What I meant is simply that the best rating system ideally (subject to various caveats which I'll list sometime) is the one which does the most accurate job of predicting how a player will perform if they play within a few months of the list coming out. Any system will have a rather large degree of error simply because performances over few games are unpredictable, but all other things being equal the one with the lowest average prediction error is the best system.

The reason I emphasise prediction over a summary of past performance is that the purpose of a rating should be to tell you how good a player is now (which means how good would they be if they played in a tournament immediately) and not how good a player was in the past, though understanding the latter is essential to trying to model the former.

I fully agree with your first para but the problem I have with the second is that no system can accurately predict an individual's performance except by reference to existing data = past performance. I suppose that there are ways of predicting performance based on some kind of probability calculations but how accurate they are I don't know. Bill, where are you when we need you most! :wall:

As for Michael's reference to a "normal" rating system that really has me flummoxed. Chess is not normal; chess players are not normal - how can the rating system be normal? :D :doh:

DJ

Rincewind
12-09-2006, 09:50 PM
what do you mean by normal?

Not normal in the statistical sense as most systems I know of assume a logistic distribution (including Elo and both Glickos). :)

Rhubarb
12-09-2006, 10:26 PM
Not normal in the statistical sense as most systems I know of assume a logistic distribution (including Elo and both Glickos). :)I feel like a bit of a goose as I've been calling the FIDE version of Elo a normal distribution, and I had to look up the difference between normal and logistic distributions. Hope it doesn't make any difference to the (fairly simplistic) comments I've made elsewhere in the ratings arena. :uhoh:

Denis_Jessop
12-09-2006, 10:34 PM
I suppose what we need is a rating system or three designed by economists. That should produce something to get our teeth into. ;)

DJ

Bill Gletsos
12-09-2006, 10:41 PM
I feel like a bit of a goose as I've been calling the FIDE version of Elo a normal distribution, and I had to look up the difference between normal and logistic distributions. Hope it doesn't make any difference to the (fairly simplistic) comments I've made elsewhere in the ratings arena. :uhoh:No.
Also Rincewind isnt actually correct with regards FIDE ratings.
If you check the FIDE rating regulations you can see they show a pair of tables at http://www.fide.com/official/handbook.asp?level=B0210
Those tables are based on the normal distribution.

At http://www.fide.com/official/handbook.asp?level=B0214 FIDE list the logistic function but note that the tables rather than the formula are used. :doh: :wall:

Rhubarb
12-09-2006, 10:51 PM
No.
Also Rincewind isnt actually correct with regards FIDE ratings.But RW didn't specifically mention the FIDE version, just Elo.



If you check the FIDE rating regulations you can see they show a pair of tables at http://www.fide.com/official/handbook.asp?level=B0210
Those tables are based on the normal distribution.At http://www.fide.com/official/handbook.asp?level=B0214 FIDE list the logistic function but note that the tables rather than the formula are used. :doh: :wall:Okay so the official table is normally distributed, while the estimate is a logistic function. Got it now, I hope. :)

Bill Gletsos
12-09-2006, 10:58 PM
But RW didn't specifically mention the FIDE version, just Elo?I know you would realise that but unlike you the comprehension challenged may not have realised that was the case as generally they will see Elo and think FIDE. ;)
As such thats why I said he wasnt "actually correct with regards FIDE ratings" as opposed to saying he wasnt "actually correct with regards Elo ratings".

Okay so the official table is normally distributed, while the estimate is a logistic function. Got it now, I hope. :)You have it.

Rincewind
12-09-2006, 11:15 PM
I feel like a bit of a goose as I've been calling the FIDE version of Elo a normal distribution, and I had to look up the difference between normal and logistic distributions. Hope it doesn't make any difference to the (fairly simplistic) comments I've made elsewhere in the ratings arena. :uhoh:

Sorry maybe I led you astray slightly there. It seems to be the logistic distributions are prefered for reasons of tractability of the resulting mathematics but normal distributions have a nicer name. :)

Edit: I should also mention that glicko.xls (available for download you know where) compares an Elo with as Glicko rating system and using the EXCEL function NORMDIST() to do the Elo bit. But this is because I'm lazy and I know NORMDIST() works.

Bill Gletsos
13-09-2006, 12:03 AM
I fully agree with your first para but the problem I have with the second is that no system can accurately predict an individual's performance except by reference to existing data = past performance.The fact that rating systems cannot 100% accurately predict an individuals future performance is without question, however that doesnt mean they cannot predict future performance at a lesser accuracy.

I suppose that there are ways of predicting performance based on some kind of probability calculations but how accurate they are I don't know.If a player were accurately rated then their actual score would equal their expected score in a future period.

Bill, where are you when we need you most! :wall:Observing. ;)

As for Michael's reference to a "normal" rating system that really has me flummoxed.I think drug answered that question. ;)

MichaelBaron
13-09-2006, 01:51 AM
what do you mean by normal?

Normal means the one that is accepted internationally as a valid tool for measurement of one's chess strength.

Desmond
13-09-2006, 08:45 AM
Normal means the one that is accepted internationally as a valid tool for measurement of one's chess strength.
By your definition, a better system could never be implemented. For it to be accepted internationally, it would have to be used in multiple countries, which it never would be because it is not accepted internationally. Are you dizzy yet?

Brian_Jones
13-09-2006, 08:53 AM
Normal means the one that is accepted internationally as a valid tool for measurement of one's chess strength.

Now that the statisticians have all gone to sleep, I can ask only one question.

Can anyone name the countries in the world that use the "Glicko" system for ratings?

Australia did I hear you say?

Desmond
13-09-2006, 08:58 AM
Now that the statisticians have all gone to sleep, I can ask only one question.

Can anyone name the countries in the world that use the "Glicko" system for ratings?

Australia did I hear you say?
I am not sure if Australia is the only country using Glicko - it could be that you are correct. I don't know.

Even if you are correct, so what? Does that mean it is a bad system? No, I don't think it does.

EGOR
13-09-2006, 09:11 AM
Even if you are correct, so what? Does that mean it is a bad system? No, I don't think it does.
But it is a valid question ask why don't any other countries use Glicko, particaully when it's supposed to be so much better than the rest. Either it is better and Australia is just ahead of it's time, or it's not and those supporting it are fooling themselves. Either way we still need to be clear on what the rating system is meant to achieve.

Desmond
13-09-2006, 09:18 AM
But it is a valid question ask why don't any other countries use Glicko, particaully when it's supposed to be so much better than the rest.If it is the best system, it is other countries that should be examining why they are not using it.
Either it is better and Australia is just ahead of it's time, or it's not and those supporting it are fooling themselves.From what I can tell of previous times this question has been asked, the answer is the former.
Either way we still need to be clear on what the rating system is meant to achieve.
Predictive accuracy would be my highest priority. If you have an alternative, let's hear it.

Rincewind
13-09-2006, 11:32 AM
Now that the statisticians have all gone to sleep, I can ask only one question.

Can anyone name the countries in the world that use the "Glicko" system for ratings?

Australia did I hear you say?

Glicko is used on many online game servers including FICS. Does anyone know about ICC?

Kevin Bonham
13-09-2006, 11:57 AM
I fully agree with your first para but the problem I have with the second is that no system can accurately predict an individual's performance except by reference to existing data = past performance. I suppose that there are ways of predicting performance based on some kind of probability calculations but how accurate they are I don't know.

We may be talking about the same thing but just in different ways; I'm not sure. One can talk about the same figure either as a projection of expected future performance or a weighted summary of past performance. The reason I use the former is that it "System X tests future results better than System Y" is a simply and directly testable hypothesis. "System X summarises past results better than System Y" is not.

I am also allowing the possibility of ratings systems in which, for whatever reason, past performance is not the sole determinant of a player's rating.

antichrist
13-09-2006, 11:57 AM
I feel like a bit of a goose as I've been calling the FIDE version of Elo a normal distribution, and I had to look up the difference between normal and logistic distributions. Hope it doesn't make any difference to the (fairly simplistic) comments I've made elsewhere in the ratings arena. :uhoh:

Do we take that as an apology to Matt?

EGOR
13-09-2006, 01:07 PM
If it is the best system, it is other countries that should be examining why they are not using it.
You could be right, but it is still a reasonable thing for us to look at why they are not taking it up and if there reasons are relevant to us.

From what I can tell of previous times this question has been asked, the answer is the former.
If it is the best system, then let us be brave and stick with it.

Predictive accuracy would be my highest priority. If you have an alternative, let's hear it.
I don't have an alternative, I don't have a big problem with GLICKO, I don't have the knowledge to know if it's the best, I was just suggesting that it's not a good idea to dismiss what other countries do out of hand.

Kevin Bonham
13-09-2006, 01:11 PM
Glicko is quite complex to administer. It's not the sort of system that anyone who doesn't know what they are doing can muck about with, nor can you "learn" everything needed to run it as quickly as with some systems.

EGOR
13-09-2006, 01:19 PM
Glicko is quite complex to administer. It's not the sort of system that anyone who doesn't know what they are doing can muck about with, nor can you "learn" everything needed to run it as quickly as with some systems.
Is this the reason why other countries are not taking it up? How about FIDE, has it ever been suggested that they change to Glicko?

Bill Gletsos
13-09-2006, 01:23 PM
Now that the statisticians have all gone to sleep, I can ask only one question.

Can anyone name the countries in the world that use the "Glicko" system for ratings?

Australia did I hear you say?I suppose you are a fan of the BCF system.

Kevin Bonham
13-09-2006, 01:41 PM
Is this the reason why other countries are not taking it up?

I'm suggesting that it is a possible reason why countries/federations have not done so. I don't have any actual evidence for this being the case however.

EGOR
13-09-2006, 02:11 PM
I'm suggesting that it is a possible reason why countries/federations have not done so. I don't have any actual evidence for this being the case however.
Sounds reasonable. What about FIDE?

Kevin Bonham
13-09-2006, 02:25 PM
Sounds reasonable. What about FIDE?

Last time I looked FIDE was a federation of sorts so I include them in my comment above.

FIDE's rating system was very dodgy for years - clearly poorer than the systems of many national federations - although some aspects have lately been improved.

EGOR
13-09-2006, 02:32 PM
Last time I looked FIDE was a federation of sorts so I include them in my comment above.
OK, missed that.

pax
13-09-2006, 06:17 PM
If a player were accurately rated then their actual score would equal their expected score in a future period.

No, that would just mean that their rating accurately predicted their score in the given period. It could still differ substantially from the true strength of the player.

Desmond
13-09-2006, 06:30 PM
No, that would just mean that their rating accurately predicted their score in the given period. It could still differ substantially from the true strength of the player.
Is that "true" as in Truth with a capital "T" - the illusory greater concept to which we mere mortals are not privy?

Surely a player's results is precisely their true strength. Nothing more. Nothing less.

Garvinator
13-09-2006, 06:35 PM
Back slightly to the use of elo in most countries. I would imagine that this is because alot of countries believe that it is important to have an international rating and be part of the world wide system, especially in the higher standard tournaments eg Europe.

I dont see it really as a reflection they these countries believe that elo is the best rating system. In alot of cases it could be that they think elo is the only option.

From my understanding, alot of the problems in fide ratings has come from the fide rating floors.

Rincewind
13-09-2006, 06:41 PM
Is that "true" as in Truth with a capital "T" - the illusory greater concept to which we mere mortals are not privy?

Surely a player's results is precisely their true strength. Nothing more. Nothing less.

Consider the following analogy.

Imagine every player carries a bag containing slips of paper with numbers from 1 to 3000 written on them. Instead of playing chess players just sit down and randomly draw out a slip of paper and and compare it to that of their opponent's, highest number wins. If the numbers are within 10 then the result of the cmoparison is considered a draw. However the comparison is done by a third party black box so that the actual numbers on the slips of paper is never known.

A player's true strength can be defined as the mean of all the values in their bag. The purpose of the rating system is to determine what that number is based only on the results of the comparisons (win, draw and loses).

Desmond
13-09-2006, 06:49 PM
Consider the following analogy.

Imagine every player carries a bag containing slips of paper with numbers from 1 to 3000 written on them. Instead of playing chess players just sit down and randomly draw out a slip of paper and and compare it to that of their opponent's, highest number wins. If the numbers are within 10 then the result of the cmoparison is considered a draw. However the comparison is done by a third party black box so that the actual numbers on the slips of paper is never known.

A player's true strength can be defined as the mean of all the values in their bag. The purpose of the rating system is to determine what that number is based only on the results of the comparisons (win, draw and loses).
Nice idea. Remind me to get a bag full of 3000's.

First time I read your post, I assumed that every bag would contain exactly one of every number from 1-3000. Is this what you intended, or are you saying that some players may have a "heavier" bag, so to speak?

Bill Gletsos
13-09-2006, 06:52 PM
No, that would just mean that their rating accurately predicted their score in the given period. It could still differ substantially from the true strength of the player.No, it means their rating reflected their playing strength in the future period.

I suppose a player could claim his true strength is that of a Grandmaster although his performances indicate he is a patzer. ;)

A valid test of a rating system lies in its success in quantitative prediction, in forcasts in the scores of tournaments, matches or indeed the next rating period.

Rincewind
13-09-2006, 06:57 PM
Nice idea. Remind me to get a bag full of 3000's.

First time I read your post, I assumed that every bag would contain exactly one of every number from 1-3000. Is this what you intended, or are you saying that some players may have a "heavier" bag, so to speak?

Yes, every bag is individual with different distributions and do not necessarily contain all numbers from 1 to 3000. Also the numbers in individual bags change over time.

MichaelBaron
13-09-2006, 07:24 PM
By your definition, a better system could never be implemented. For it to be accepted internationally, it would have to be used in multiple countries, which it never would be because it is not accepted internationally. Are you dizzy yet?


Well, one system that comes to mind and that is accepted internationally is ELO:hmm:

pax
13-09-2006, 08:54 PM
Is that "true" as in Truth with a capital "T" - the illusory greater concept to which we mere mortals are not privy?

Surely a player's results is precisely their true strength. Nothing more. Nothing less.

No, a player's true strength is the mean of the statistical distribution of his expected future performance.

pax
13-09-2006, 09:05 PM
A valid test of a rating system lies in its success in quantitative prediction, in forcasts in the scores of tournaments, matches or indeed the next rating period.

Absolutely, and I don't dispute that.

I just don't think it is possible to say that a player is accurately rated. You can only ever say that the rating accurately predicted results. Future results can always contradict your current hypothesis.

It's like trying to estimate the mean of a statistical distribution by taking samples from the distribution. You can never say that your current estimate is accurate, only that it accurately reflects the observations that you have made so far.

Desmond
13-09-2006, 10:10 PM
Well, one system that comes to mind and that is accepted internationally is ELO:hmm:
So you are saying that you desire the most popular system, not the best one.

Rincewind
13-09-2006, 11:44 PM
I just don't think it is possible to say that a player is accurately rated. You can only ever say that the rating accurately predicted results. Future results can always contradict your current hypothesis.

But playing strength is a moving target. All you can hope for is to predict performance in the period for which the rating is published.


It's like trying to estimate the mean of a statistical distribution by taking samples from the distribution. You can never say that your current estimate is accurate, only that it accurately reflects the observations that you have made so far.

This is true of all measurement as far as I am aware. By your reasoning no measurement could be called "accurate" since there is always a measurement error.The actuality is we all live with the fact that all measurement is accurate to within the this tolerence.

MichaelBaron
14-09-2006, 12:45 AM
So you are saying that you desire the most popular system, not the best one.

I desire a system that has been tested. Right now all of us are gunea-pigs:wall:

Garvinator
14-09-2006, 12:51 AM
I desire a system that has been tested. Right now all of us are gunea-pigs:wall:
and what is wrong with that?

Desmond
14-09-2006, 08:51 AM
I desire a system that has been tested. Right now all of us are gunea-pigs:wall:
Which gets back to what I said earlier. If all countries took this view, there would be no possible improvement to the rating systems anywhere. Is this your desire?

pax
14-09-2006, 09:11 AM
But playing strength is a moving target. All you can hope for is to predict performance in the period for which the rating is published.


Do you not agree that there is an underlying strength that we are trying to measure that is distinct from the results in a particular period? I do not disagree that predicting the actual performance is the best that we can achieve but I'm just trying to illustrate the difference between a player's true strength and their rating.



This is true of all measurement as far as I am aware. By your reasoning no measurement could be called "accurate" since there is always a measurement error.The actuality is we all live with the fact that all measurement is accurate to within the this tolerence.

The distinction here is that we are estimating the mean of a statistical distribution by taking samples from that distribution. It's not the same as measuring a fixed number with an instrument that has some known error bound.

For example, I can measure your height with an instrument that is accurate to the nearest millimetre. It is true then to say that the measurement is accurate to +/- 1mm. The same is not true of ratings - in this case we can only measure a statistical likelihood based on the observations we have made. Perhaps we can say that our estimate is within 50 rating points with 95% confidence.

Vlad
14-09-2006, 09:21 AM
I wish you guys took a basic course in econometrics; there would be no argument then.

Rincewind
14-09-2006, 04:49 PM
Do you not agree that there is an underlying strength that we are trying to measure that is distinct from the results in a particular period? I do not disagree that predicting the actual performance is the best that we can achieve but I'm just trying to illustrate the difference between a player's true strength and their rating.

Yes I agree. I don;t agree that you can allocate any future rating changes to just inaccuracy of the first estimate as there is also the moving target problem to consider.


The distinction here is that we are estimating the mean of a statistical distribution by taking samples from that distribution. It's not the same as measuring a fixed number with an instrument that has some known error bound.

For example, I can measure your height with an instrument that is accurate to the nearest millimetre. It is true then to say that the measurement is accurate to +/- 1mm. The same is not true of ratings - in this case we can only measure a statistical likelihood based on the observations we have made. Perhaps we can say that our estimate is within 50 rating points with 95% confidence.

I didn't say that estimating a mean from a selecting a few members from a population and extrapolating was the same as standard measurement error. Just pointing out that all measurement has error. As such the term accurate as you used it in your post was physically meaningless.

Oepty
14-09-2006, 05:01 PM
Kevin. Yes it is vague, but it is an objective. I would class your proposed objective of a rating system as more of a test for a rating system. The rating system should show how players have been performing in the recent past. The fact that most players performances do not jump or fall by 100s points in a week means that prediction of performances is a good test and most likely the best test of whether the rating system is doing its job.
Scott

Upon further thought I have come to the conclusion what I wrote here is nonsense.
Scott

pax
14-09-2006, 05:30 PM
I didn't say that estimating a mean from a selecting a few members from a population and extrapolating was the same as standard measurement error. Just pointing out that all measurement has error. As such the term accurate as you used it in your post was physically meaningless.

It's not meaningless at all. In one case, a measurement has a known error bound - in that case you can guarantee accuracy within the bound. In the statistical case you CANNOT guarantee accuracy within any given bound.

pax
14-09-2006, 05:36 PM
Yes I agree. I don;t agree that you can allocate any future rating changes to just inaccuracy of the first estimate as there is also the moving target problem to consider.

Yes, the moving target is an issue. But we have to assume an upper bound on the rate of change of a player's true strength to build any rating system. Otherwise you would just rate players according to their most recent performance.

An example of what I am talking about. A player performs at 2000 for nine games in one period, and then at 1600 over twenty games in the next period. Has the player suddenly got weaker by 400 points? Probably not - more likely is that his initial performance was an exceptional given his ability.

Rincewind
14-09-2006, 11:37 PM
It's not meaningless at all. In one case, a measurement has a known error bound - in that case you can guarantee accuracy within the bound. In the statistical case you CANNOT guarantee accuracy within any given bound.

I didn't say they weren't different but you said "You can never say that your current estimate is accurate". Just pointing out that this same phrase can apply to any measurement.

The truth of the matter is that statistics work and is in fact one of the more practical areas of mathematics. Where I come from applied statisticians do a lot more consulting work than, say, applied mathematicians who solve differential equations. (And this is coming from one of the DE breed).

Rincewind
14-09-2006, 11:40 PM
Yes, the moving target is an issue. But we have to assume an upper bound on the rate of change of a player's true strength to build any rating system. Otherwise you would just rate players according to their most recent performance.

An example of what I am talking about. A player performs at 2000 for nine games in one period, and then at 1600 over twenty games in the next period. Has the player suddenly got weaker by 400 points? Probably not - more likely is that his initial performance was an exceptional given his ability.

Performance ratings jump all around the place as they are generally based on a small number of games which are less statistically significant than the rating in general. Well designed model should be able to give the players "true strength" to with within certain bounds with acceptable confidence levels.

pax
15-09-2006, 12:45 PM
I didn't say they weren't different but you said "You can never say that your current estimate is accurate". Just pointing out that this same phrase can apply to any measurement.


No. In one case you can say "my estimate is accurate to 1mm", and in the other case you cannot. You can at best say "my estimate is accurate to 1mm with a confidence of 95%". It's not the same thing.



The truth of the matter is that statistics work and is in fact one of the more practical areas of mathematics. Where I come from applied statisticians do a lot more consulting work than, say, applied mathematicians who solve differential equations. (And this is coming from one of the DE breed).

Huh? What has any of that got to do with the argument? Of course statistics work - that's the whole point. But the guarantees you get from a statistical argument are necessarily probabilistic.

Rincewind
15-09-2006, 01:59 PM
No. In one case you can say "my estimate is accurate to 1mm", and in the other case you cannot. You can at best say "my estimate is accurate to 1mm with a confidence of 95%". It's not the same thing.

They are more alike then they are different. Both are talking about a measurement within precise tolerances. Neither have absolute accuracy. One includes a confidence level, that's all.


Huh? What has any of that got to do with the argument? Of course statistics work - that's the whole point. But the guarantees you get from a statistical argument are necessarily probabilistic.

The relevance is that a stocastic model is appropriate for this (and a lot of) real world problems. You seem to have a bias towards deterministic model.

pax
15-09-2006, 02:37 PM
The relevance is that a stocastic model is appropriate for this (and a lot of) real world problems. You seem to have a bias towards deterministic model.
You aren't listening to a word I'm saying. Of course it's a stochastic model. That's the whole point.. :wall::wall::wall:

pax
15-09-2006, 02:38 PM
ac·cu·rate (ăk'yər-ĭt) pronunciation
adj.

1. Conforming exactly to fact; errorless.
2. Deviating only slightly or within acceptable limits from a standard.
3. Capable of providing a correct reading or measurement: an accurate scale.
4. Acting or performing with care and precision; meticulous: an accurate proofreader.

Rincewind
15-09-2006, 05:54 PM
Yes and you said


It's like trying to estimate the mean of a statistical distribution by taking samples from the distribution. You can never say that your current estimate is accurate, only that it accurately reflects the observations that you have made so far.

I believe a deviation and confidence level is accurate according to definition (2) above. So what the heck have you been talking about?

pax
15-09-2006, 06:25 PM
Yes and you said



I believe a deviation and confidence level is accurate according to definition (2) above. So what the heck have you been talking about?

'Within acceptable limits' to me implies a bound. A confidence limited error is not a bound. Anyway, as much as I'd love to keep playing these semantic games with you, I've got better things to do. Have a nice weekend.

Rincewind
15-09-2006, 09:11 PM
'Within acceptable limits' to me implies a bound. A confidence limited error is not a bound. Anyway, as much as I'd love to keep playing these semantic games with you, I've got better things to do. Have a nice weekend.

:lol:

Cat
15-09-2006, 11:41 PM
No, a player's true strength is the mean of the statistical distribution of his expected future performance.

Precisely, but there's nothing to say that the distribution should be normal. That could only be determined through field observation. Assuming a normal distribution may be erroneous without empirical data.

pax
16-09-2006, 12:33 AM
Precisely, but there's nothing to say that the distribution should be normal. That could only be determined through field observation. Assuming a normal distribution may be erroneous without empirical data.
What is this, tag-team baiting? Where did I assume anything was normal? I merely assumed it had a mean and a non-zero variance.


Says he who said he had better things to do :doh:

Bill Gletsos
16-09-2006, 12:52 AM
Precisely, but there's nothing to say that the distribution should be normal. That could only be determined through field observation. Assuming a normal distribution may be erroneous without empirical data.The many performances of an individual will be normally distributed when evaluated on an appropriate scale.

Extensive investigation by Elo in 1965 and McClintock in 1977 bore the validity of that statement out.

GrimReaper
17-09-2006, 08:07 PM
The many performances of an individual will be normally distributed when evaluated on an appropriate scale.

Extensive investigation by Elo in 1965 and McClintock in 1977 bore the validity of that statement out.
Hi Bill,

I played in a tournament today and was wondering if I could get an early peak at what my new rapid rating will be?

Bill Gletsos
17-09-2006, 08:11 PM
Hi Bill,

I played in a tournament today and was wondering if I could get an early peak at what my new rapid rating will be?Not possible.
Apart from the fact that we dont do that otherwise everyone would be requesting it, your next rating is dependant on the performances of your opponents during the next rating period.
As such what there performances may be is as yet unknown.

Cat
17-09-2006, 10:59 PM
The many performances of an individual will be normally distributed when evaluated on an appropriate scale.

Extensive investigation by Elo in 1965 and McClintock in 1977 bore the validity of that statement out.

That's not quite the same thing as we're talking about - good try anyway, Bill!

Garvinator
18-09-2006, 11:12 AM
Bill,

Do you know what is the record for the greatest loss of rating points in one period under glicko?

four four two
18-09-2006, 12:24 PM
Bill,

Do you know what is the record for the greatest loss of rating points in one period under glicko?

Some players have lost over 200 points from just 7 games.:whistle:

pax
18-09-2006, 01:42 PM
Some players have lost over 200 points from just 7 games.:whistle:

That's not necessarily unreasonable - especially if their previous rating was based on a few games a long time ago..

Denis_Jessop
18-09-2006, 04:34 PM
Bill,

Do you know what is the record for the greatest loss of rating points in one period under glicko?

There's a new list for your ratings page, Bill - a biggest losers list as well as a most improved list. Note that I don't suggest a most inactive players list as it would be rather long :doh:

DJ

Denis_Jessop
18-09-2006, 08:07 PM
And while we are at it may I observe that the ACF Master File contains the names of many dead players. I know of at least three ACT players on the list who have been dead for 5 to 15 years or so.

Now this is not a matter within the control of the ACF Ratings Officers as they can't be expected to make themselves aware of such occurrences.

My eminently practical suggestion is that each player should include a clause in his or her will instructing the Executor of the estate to notify the ACF Ratings Officers of their demise.

I am not sure what the statistical effect (if any) of dead men's rating points being in the pool is. But I have a really wizard idea if the RO's will buy it. A deceased player could bequeath his rating points by will.

Some suggested forms of bequest:

"I give, devise and bequeath to my dear friend X all my ACF Glicko 2 rating points existing at my death to be distributed to him (her) at the rate of 100 points per rating period beginning at the date of the grant of probate of this my last will..."

" I give, devise and bequeath to the ACF Ratings Officers all my ACF Glicko 2 rating points existing at the date of my death to be distributed by them to other ACF-rated players as they think fit, provided that none of the said points is to be given to a junior."

:cool: :clap: :hmm: :owned:

DJ

Cat
18-09-2006, 10:52 PM
And while we are at it may I observe that the ACF Master File contains the names of many dead players. I know of at least three ACT players on the list who have been dead for 5 to 15 years or so.

Now this is not a matter within the control of the ACF Ratings Officers as they can't be expected to make themselves aware of such occurrences.

My eminently practical suggestion is that each player should include a clause in his or her will instructing the Executor of the estate to notify the ACF Ratings Officers of their demise.

I am not sure what the statistical effect (if any) of dead men's rating points being in the pool is. But I have a really wizard idea if the RO's will buy it. A deceased player could bequeath his rating points by will.

Some suggested forms of bequest:

"I give, devise and bequeath to my dear friend X all my ACF Glicko 2 rating points existing at my death to be distributed to him (her) at the rate of 100 points per rating period beginning at the date of the grant of probate of this my last will..."

" I give, devise and bequeath to the ACF Ratings Officers all my ACF Glicko 2 rating points existing at the date of my death to be distributed by them to other ACF-rated players as they think fit, provided that none of the said points is to be given to a junior."

:cool: :clap: :hmm: :owned:

DJ

Great idea, very imaginitive Dennis. But why wait 'til death, why not trade points on an ACF stock exchange? This would create an internal market, the kids would love it!

Ian Rogers could go out hunting rating points then on-sell to the Gold Coast rich kids, he'd make a fortune. Suddenly chess would become lucrative and glamorous, players would make real money and interest in the game would sky-rocket. Professional players would become Glicko-miners. This could be your presidential legacy - go for it!

Garrett
19-09-2006, 06:55 AM
ACF Master File contains the names of many dead players. I know of at least three ACT players on the list who have been dead for 5 to 15 years or so.

yes there's a Queenslander on there too that has been gone for around twenty years, I think his rating went up too.

Denis_Jessop
19-09-2006, 03:41 PM
yes there's a Queenslander on there too that has been gone for around twenty years, I think his rating went up too.

Very likely he received 220 points (150 + 70) though I'm not sure about the latter.

Seriously, it is a bit of a problem presentationally as there's no way at that the ratings officers can get to know this. They would need more than hearsay as it would be embarrassing if a reputedly dead player suddenly appeared and started playing again. And if you think that's weird, I could tell you a relevant story from my days as a young solicitor in Melbourne but it's too long.

The whole thing may not have any practical effect on the operation of the ratings system any more than the huge number of inactive players, including myself, unfortunately.

DJ

Garvinator
19-09-2006, 04:21 PM
Bill,

Do you know what is the record for the greatest loss of rating points in one period under glicko?
so much for what was actually a serious question :(

ER
19-09-2006, 07:08 PM
....it would be embarrassing if a reputedly dead player suddenly appeared and started playing again. And if you think that's weird, I could tell you a relevant story from my days as a young solicitor in Melbourne but it's too long.

Please Denis, do tell even in point form! :)
Cheers and good luck!

ER
19-09-2006, 07:10 PM
... and by the way could someone please give me a fast tutorial on how to quote properly? Thanks in advance!
Cheers and good luck!

Bill Gletsos
19-09-2006, 07:38 PM
... and by the way could someone please give me a fast tutorial on how to quote properly? Thanks in advance!
Cheers and good luck!Hit the edit button on your last post and see how I changed it to put Debi's comments in quotes. ;)

Denis_Jessop
19-09-2006, 08:01 PM
Please Denis, do tell even in point form! :)
Cheers and good luck!

OK - it's not all that long but has nothing to do with chess.

A client came to the office with the problem that (and I think it was the husband but that's not relevant; also it was about 40 years ago) their spouse had disappeared many years before. There was then a legal principle that if one's spouse had disappeared for more than 7 years one could (on establishing all the necessary facts) obtain a court order of presumed death which had the effect of dissolving the marriage. The client hadn't done that but had instead gone to the local Registrar of Births, Deaths and Marriages who had obligingly given the client a death certificate in respect of the spouse based on 7 years' absence. That shouldn't have been done, but relying on it the client re-married. As you might guess, not long afterwards, the supposedly dead spouse re-appeared! I don't have a punch line as I was working for an advice bureau so my advice was see a solicitor quickly. That wasn't very big o' me (apologies to Groucho for that). But it just goes to show that you're not dead until the fat lady sings :rolleyes:

DJ

ER
19-09-2006, 08:14 PM
Good one Denis thanks :) As for the fat lady she has already done all her singing for Vic AFL clubs this year! :)

ER
19-09-2006, 08:18 PM
HI Bill thanks. It looks classy!
Cheers and good luck!

MichaelBaron
19-09-2006, 08:54 PM
And while we are at it may I observe that the ACF Master File contains the names of many dead players. I know of at least three ACT players on the list who have been dead for 5 to 15 years or so.

Now this is not a matter within the control of the ACF Ratings Officers as they can't be expected to make themselves aware of such occurrences.

My eminently practical suggestion is that each player should include a clause in his or her will instructing the Executor of the estate to notify the ACF Ratings Officers of their demise.

I am not sure what the statistical effect (if any) of dead men's rating points being in the pool is. But I have a really wizard idea if the RO's will buy it. A deceased player could bequeath his rating points by will.

Some suggested forms of bequest:

"I give, devise and bequeath to my dear friend X all my ACF Glicko 2 rating points existing at my death to be distributed to him (her) at the rate of 100 points per rating period beginning at the date of the grant of probate of this my last will..."

" I give, devise and bequeath to the ACF Ratings Officers all my ACF Glicko 2 rating points existing at the date of my death to be distributed by them to other ACF-rated players as they think fit, provided that none of the said points is to be given to a junior."

:cool: :clap: :hmm: :owned:

DJ

No need to wait for me to die. If somebody wants my ACF rating points (that are meaningless anyway, since the Fide rating is the only one recognized internationally) let me know..I can give them to you as a gift. Hm...or shall I sell my rating ? How much would you pay for 1000 rating points:whistle: ?

Cat
19-09-2006, 10:35 PM
No need to wait for me to die. If somebody wants my ACF rating points (that are meaningless anyway, since the Fide rating is the only one recognized internationally) let me know..I can give them to you as a gift. Hm...or shall I sell my rating ? How much would you pay for 1000 rating points:whistle: ?

5 million roubles

harry
19-09-2006, 11:12 PM
"No need to wait for me to die. If somebody wants my ACF rating points (that are meaningless anyway, since the Fide rating is the only one recognized internationally) let me know..I can give them to you as a gift. Hm...or shall I sell my rating ? How much would you pay for 1000 rating points ?" Baron

Next time we play can i buy a win :D

EGOR
20-09-2006, 08:34 AM
If somebody wants my ACF rating points (that are meaningless anyway, since the Fide rating is the only one recognized internationally) let me know..
Meaningless to you! My ACF rating (1242 don't laugh.:evil: ) is the only one I've got. Thanks for thinking of the little guy.:whistle:

peter_parr
20-09-2006, 02:05 PM
Bill,

Do you know what is the record for the greatest loss of rating points in one period under glicko?

One player lost over 700 rating points from one game using the Glicko rating system.

When I showed Gletsos he made a manual adjustment. There maybe others.

The maximum he would have lost on the FIDE rating system would be 15.
Losing or gaining up to 15 rating points in one game is sensible and has been standard with FIDE since ratings first started.

The brilliant mathematician John Nunn (the youngest in 500 years to enter Oxford University) and top grandmaster recently wrote a lengthy article about why players should NOT be penalised for inactivity.

Australia is the only country in the world using the Glicko rating system.
Only in Australia can a player lose (or gain?) hundreds of rating points in a few days.

30 years ago I entered a lift in my hotel at the Hafia Olympiad and there was Prof Elo, Prof Euwe and Ed Lasker!

What would they think of Glicko ratings?

Garvinator
20-09-2006, 02:45 PM
My post 48 from this thread.

Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

--------------------------------------------------------------------------
Just repeating the question since Peter Parr didnt reply to it the first time.

Bill Gletsos
20-09-2006, 03:01 PM
One player lost over 700 rating points from one game using the Glicko rating system.Who.

When I showed Gletsos he made a manual adjustment.I dont recall making any manual adjustment.

The maximum he would have lost on the FIDE rating system would be 15.
Losing or gaining up to 15 rating points in one game is sensible and has been standard with FIDE since ratings first started.

The brilliant mathematician John Nunn (the youngest in 500 years to enter Oxford University) and top grandmaster recently wrote a lengthy article about why players should NOT be penalised for inactivity.Actually I think you will find that he argued why points should not be automatically deducted from inactive players on say an annual basis simply for being inactive.

Australia is the only country in the world using the Glicko rating system.So , England is the the only country using the BCF system.

Only in Australia can a player lose (or gain?) hundreds of rating points in a few days.Yes well thats Glicko's main benefit. thanks for reminding everyone. :hand:

30 years ago I entered a lift in my hotel at the Hafia Olympiad and there was Prof Elo, Prof Euwe and Ed Lasker!

What would they think of Glicko ratings?That is a question no one knows the answer to. however I would like to think that they all would recognise that a players rating should indicate their current strength and not their strength from 8 years ago.

Bill Gletsos
20-09-2006, 03:03 PM
My post 48 from this thread.

Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

--------------------------------------------------------------------------
Just repeating the question since Peter Parr didnt reply to it the first time.Perhaps Peter feels he cannot perform at his current listed ACF rating.

EGOR
20-09-2006, 03:11 PM
Who.
I dont recall making any manual adjustment.
Actually I think you will find that he argued why points should not be automatically deducted from inactive players on say an annual basis simply for being inactive.
So , England is the the only country using the BCF system.
Yes well thats Glicko's main benefit. thanks for reminding everyone. :hand:
That is a question no one knows the answer to. however I would like to think that they all would recognise that a players rating should indicate their current strength and not their strength from 8 years ago.
Hi Bill,
I dream of a time when you respond to a post you disagree with without chopping it into little bits and attacking each little bit with a carving knife.:naughty:

peter_parr
20-09-2006, 03:16 PM
My post 48 from this thread.

Peter,

I have been reading your comments for quite a while now and it is always on the same point, when a returning player doesnt play to their old rating and loses 'alot of points' under Glicko2.

I never see you making the same complaints when a player gains a lot of points.

--------------------------------------------------------------------------
Just repeating the question since Peter Parr didnt reply to it the first time.

Yes I agree a player should not gain a lot of rating points from not many games.

I have stated this (in my last post 155).

A player should not go up by more then 15 rating points from one game if he wins.

Basil
20-09-2006, 03:20 PM
Hi Bill,
I dream of a time when you respond to a post you disagree with without chopping it into little bits and attacking each little bit with a carving knife.:naughty:
Egor, are you aware that a number of the 'dark side' trolls (humour and licence used) practice this behaviour and have it down to a fine art? In truth, Bill is not anywhere near the worst offender :hmm: Caaaarn, fair's fair.

EGOR
20-09-2006, 03:23 PM
Egor, are you aware that a number of the 'dark side' trolls (humour and licence used) practice this behaviour and have it down to a fine art? In truth, Bill is not anywhere near the worst offender :hmm: Caaaarn, fair's fair.
This I cannot deny.:(

pax
20-09-2006, 03:48 PM
The maximum he would have lost on the FIDE rating system would be 15.
Losing or gaining up to 15 rating points in one game is sensible and has been standard with FIDE since ratings first started.


It's debatable whether this is sensible. Many people believe K=15 ELO is too slow, especially in a national rating system where many people don't play all that often. For example, if you only play 20 games or so per year it could take years for your rating to catch up with your playing strength if you have a sudden jump or dip in strength.

In the days of ELO, 1700-1800 was a very big rating for a junior - they were probably playing at more like 2100.

I actually agree with Peter about the problems facing inactive players under Glicko, but I don't think K=15 is necessarily the answer.

EGOR
20-09-2006, 04:04 PM
Egor, are you aware that a number of the 'dark side' trolls (humour and licence used) practice this behaviour and have it down to a fine art? In truth, Bill is not anywhere near the worst offender :hmm: Caaaarn, fair's fair.
Being not very good at it isn't an excuse for doing it.:whistle:
are suggesting he needs more pactice?:hmm: :uhoh:

Basil
20-09-2006, 04:52 PM
Being not very good at it isn't an excuse for doing it.:whistle:
are suggesting he needs more pactice?:hmm: :uhoh:
No. I was suggesting you need more practice at being impartial and even-handed! :) :whistle:
Even you are clearly attempting to do so

arosar
20-09-2006, 05:21 PM
Why do we need a national rating anyways? Let's just go all FIDE. That'll fix it.

AR

Bill Gletsos
20-09-2006, 05:51 PM
Hi Bill,
I dream of a time when you respond to a post you disagree with without chopping it into little bits and attacking each little bit with a carving knife.:naughty:I dream of a time when people dont make hypocritical posts. ;)

http://www.chesschat.org/showpost.php?p=121742&postcount=42

Let he who is without sin cast the first stone. :hand: :whistle:

Bill Gletsos
20-09-2006, 05:53 PM
Why do we need a national rating anyways? Let's just go all FIDE. That'll fix it.You obviously dont realise that FIDE ratings dont currently go below 140 and there is no plan to go below 1000.
Also if people think juniors are underrated then the last system you want to follow is FIDE's.

Cat
20-09-2006, 06:04 PM
The many performances of an individual will be normally distributed when evaluated on an appropriate scale.

Extensive investigation by Elo in 1965 and McClintock in 1977 bore the validity of that statement out.

The problem is Bill, you simply don't have enough of a grasp of statistical variance to really get to grips with the issue.
Give it a rest, that's my advice!

Bill Gletsos
20-09-2006, 06:07 PM
The problem is Bill, you simply don't have enough of a grasp of statistical variance to really get to grips with the issue.The problem is cat you dont have a grasp of anything to do with ratings.

Give it a rest, that's my advice!Try doing this yourself.

Cat
20-09-2006, 07:02 PM
The problem is cat you dont have a grasp of anything to do with ratings.
Try doing this yourself.

Bill, it's a real concern when the ratings officer comes out with a statement like this in relation to the discussion Jonathon and I were having.

Just think about it, if variance was always normally distributed then ratings would never change, would they (apart from some minor stochastic variation)? It means that you are completely wasting your time publishing the ratings, if you really believe that!

Either you have no idea of what you're talking about, or you'll say anything to defend the Glicko system, or maybe both.

Variance must skew over time for changes in ratings to occur. The question then arises, is that skew predictable and measurable. If so then representing that skew mathematically would improve rating measurement.

Rincewind
20-09-2006, 07:06 PM
Just think about it, if variance was always normally distributed then ratings would never change, would they (apart from some minor stochastic variation)? It means that you are completely wasting your time publishing the ratings, if you really believe that!

This is not true as what you are measuring is also changing in time. At each point of time the performances could be normally distributed and over all time for that matter they could also be normally distributed but at discrete times the mean could be shifting and it is the mean of the next period that we are trying to predict.

Bill Gletsos
20-09-2006, 07:10 PM
Bill, it's a real concern when the ratings officer comes out with a statement like this in relation to the discussion Jonathon and I were having.

Just think about it, if variance was always normally distributed then ratings would never change, would they (apart from some minor stochastic variation)? It means that you are completely wasting your time publishing the ratings, if you really believe that!

Either you have no idea of what you're talking about, or you'll say anything to defend the Glicko system, or maybe both.The only person not having a clue what they are talking about is you as you have demonstrated on virtually every occasion.

Given you have demonstrated your cluelessness in rating matters I simply made the comment that:

The many performances of an individual will be normally distributed when evaluated on an appropriate scale.

Extensive investigation by Elo in 1965 and McClintock in 1977 bore the validity of that statement out.

If you dont agree, I dont care as you have no credibility in this area.

Cat
20-09-2006, 07:11 PM
This is not true as what you are measuring is also changing in time. At each point of time the performances could be normally distributed and over all time for that matter they could also be normally distributed but at discrete times the mean could be shifting and it is the mean of the next period that we are trying to predict.

Yes Barry, thats exactly what Jonathon and I were talking about, time variance. Bill then comes out with this crass statement - he obviously didn't grasp what we were discussing - if he doesn't understand he should say so.

He's been very rigid in his defence of the GLicko. To display this kind of ignorance requires a proper explanation. I think Bill needs to explain himself, not you.

Cat
20-09-2006, 07:13 PM
The only person not having a clue what they are talking about is you as you have demonstrated on virtually every occasion.

Given you have demonstrated your cluelessness in rating matters I simply made the comment that:

Yes but it was inappropriate Bill, completely out of context. Do you understand that? If it was a genuine error, then fess up, no big deal. If however you're struggling with these concepts, then own up for everyones sake. Be brave!

Bill Gletsos
20-09-2006, 07:19 PM
Yes but it was inappropriate Bill, completely out of context.Where you are discussing rating ever comment you make is open to question as you have continually demonstrated you dont understand even the basics.
I was simply stating something that you have never shown you understand.

Do you understand that?I understand you have continually demonstrated you are clueless.

If it was a genuine error, then fess up, no big deal. If however you're struggling with these concepts, then own up for everyones sake. Be brave!You need to stop treating rating threads as you litter box. :hand:

Cat
20-09-2006, 07:53 PM
Where you are discussing rating ever comment you make is open to question as you have continually demonstrated you dont understand even the basics.
I was simply stating something that you have never shown you understand.
I understand you have continually demonstrated you are clueless.
You need to stop treating rating threads as you litter box. :hand:


Bill, that's a really pathetic response. Stop the insults and demonstrate some spine.

You are the ratings officer and you made a fundamental and enormous gaff. Do you understand that? If you don't - own up, demonstrate magnanimity and just a little humility.

You have a responsilbility to the Australian chessplayers because they have endured some pretty heavy blustering from you about how good your rating system is. Maybe it's time for just a little more honesty. Your behavior right now is none too inspiring!

Bill Gletsos
20-09-2006, 08:00 PM
Bill, that's a really pathetic response.It is more than you deserve.

Stop the insults and demonstrate some spine.I dont know why I both wasting my time trying to enlighten the likes of you.
You are a complete waste of my time.
Then again you always have been.

You are the ratings officer and you made a fundamental and enormous gaff.No I made a comment on something you have demonstrated you have no concept about.

Do you understand that? If you don't - own up, demonstrate magnanimity and just a little humility.I understand that you are clueless.

You have a responsilbility to the Australian chessplayers because they have endured some pretty heavy blustering from you about how good your rating system is. Maybe it's time for just a little more honesty. Your behavior right now is none too inspiring!I simply pointed out something you had shown you had no concept of.

What surprised me the most was that you didnt ask who McClintock was. :whistle:

Cat
20-09-2006, 08:00 PM
What's going on Billy? Waiting for somebody to help or coach you on how to answer the question? It's still not an answer. You're now starting to look really stupid!

Bill Gletsos
20-09-2006, 08:01 PM
What's going on Billy? Somebody helping or coaching you on how to answer the question?Dream on clueless one.

Cat
20-09-2006, 08:05 PM
Dream on clueless one.


Answer the question Bill.

Kevin Bonham
20-09-2006, 08:06 PM
Hi Bill,
I dream of a time when you respond to a post you disagree with without chopping it into little bits and attacking each little bit with a carving knife.:naughty:

It's called "debate". Peter made a number of different assertions - Bill responded to each of them in turn.



The maximum he would have lost on the FIDE rating system would be 15.
Losing or gaining up to 15 rating points in one game is sensible and has been standard with FIDE since ratings first started.

Peter, why do you think 15 points per game should be the limit? Do you think it leads to the system predicting results better - if so, how? Or do you have some other reason for preferring this limit?

Suppose we are using this 15 point limit and a new player plays nine games against players of his own strength scoring 3 wins, 3 draws and 3 losses, gets a rating, and then plays one further game against a player of his own strength, winning, then has his rating processed again.

Had the player lost one of his three wins in his first batch of games his start rating would be about 80 points lower than had he won.

But had he lost the tenth game instead of winning it, his rating would only have been affected by 15 points.

That is why I have long believed 15 points maximum gain/loss is incorrect. It means a player's rating is too much influenced by the results of a player's first nine games. It can take (and I say this from experience) literally years for the player's first performances to stop unduly affecting their rating. Even 30 points gain/loss can place too much emphasis on early performances.

Some may say that the point of 15 points maximum is to be conservative. But overemphasising a player's first performances compared to their subsequent ones isn't conservative at all - it's just statistically indefensible.

Basil
20-09-2006, 08:21 PM
Bill, it's a real concern when the ratings officer comes out with a statement like this in relation to the discussion Jonathon and I were having.
David, fair suck, mate. You started the insult chain in this thread with Bill. Own up, let it go and don't flip it back on Bill. That's not cricket.

Your post #169 is the spot you want.

Cat
20-09-2006, 08:27 PM
Come on Bill, this is a serious matter. It deserves a little more than trite remarks. You must give an honest account of yourself. This is not about my behaviour. I'm asking a very clear simple question. If you don't understand, say so. If you do, answer the question properly.

Cat
20-09-2006, 08:35 PM
Okay Bill, its now 11/2 hours and you still haven't answered the question. The fact that you're floundering is clear for all to see. I'm going out for a while but I expect a proper answer by the time I get back. Otherwise I'll assume your ignorance on the matter.

Bill there will be many scratching their heads if you can't come up with a proper response on this.

Bill Gletsos
20-09-2006, 08:59 PM
Answer the question Bill.I did in post #176. :hand:

Bill Gletsos
20-09-2006, 08:59 PM
Come on Bill, this is a serious matter. It deserves a little more than trite remarks. You must give an honest account of yourself. This is not about my behaviour. I'm asking a very clear simple question. If you don't understand, say so. If you do, answer the question properly.I did in post #176. :hand:

Bill Gletsos
20-09-2006, 09:00 PM
Okay Bill, its now 11/2 hours and you still haven't answered the question. The fact that your floundering is clear for all to see. I'm going out for a while but I expect a proper answer by the time I get back. Otherwise I'll assume your ignorance on the matter.I answered it in post #176. :hand:

Bill there will be many scratching their heads if you can't come up with a proper response on this.One the clueless like you.

arosar
20-09-2006, 09:03 PM
Bill....firstly, thanks for your love PM.

Now, take a look at your last 3 posts. Did you really have to do that? Three different posts, same content. I mean, no wonder you get dubbed a Billbot.

Just stop it already. You're me president for chrissake. It's friggin' embarassing.

AR

Bill Gletsos
20-09-2006, 09:22 PM
Bill....firstly, thanks for your love PM.Your welcome. Just watch the language and try and stay out of trouble. ;)

Now, take a look at your last 3 posts. Did you really have to do that?Absolutely. The Cat has comprehension problems.

Three different posts, same content.Actually I added an extra bit to the third one just to make things clear. ;)

I mean, no wonder you get dubbed a Billbot.And you think I care. :hand:

Just stop it already. You're me president for chrissake. It's friggin' embarassing.Well you see Amiel, the Cat does not get to tell me what I meant when I make a post.
I know exactly what I meant by my post #132 and I stated it in post #176.

Rhubarb
20-09-2006, 09:30 PM
Why do we need a national rating anyways? Let's just go all FIDE. That'll fix it.Of course, this is one of Sweeney's ideas, into which he (and apparently you as well Amiel) has put zero intelligent thought. I'll try to restrain myself, however.

* As Bill says, FIDE currently doesn't go below 1400 and has no plans to go below 1000 in the long term. Perhaps you don't think a national body has a reponsibility to provide an end-to-end system?

* It would take years, probably a decade, for everyone who is currently rated over 1000 ACF to get a FIDE rating. This is because it takes several tournaments playing (and scoring) against FIDE rated players, and there aren't currently any FIDE rated players in the range that 1000 players could score against. They just don't exist yet.

* As I've stated on many occasions, FIDE's lowering of its floor can hardly be considered altruistic. The per-game fees for FIDE rated events vary, but are usually 3-4 times more expensive than ACF fees. This is on top of an annual fee that FIDE charges the ACF based on how many FIDE rated players we have, which, by the way, is on top of our annual affiliation fees.

* FIDE has strict registration and time control regulations. Half the weekenders in the country would have to lengthen their time controls for a start.

* The range of FIDE ratings used to be 2200-2700. If a junior made it on to the list, it wasn't a problem - how underrated could they possibly be? Now that the range is 1400(or 1000)-2800, and with K=15 (or initially 25), they now have a system unequipped to deal with rapidly improving juniors (I wonder if FIDE realises that yet). Personally I much prefer a system that handles juniors properly, and if it means making long-term inactive players prove their strength on their return, so much the better.

Basil
20-09-2006, 09:32 PM
Bill, now, take a look at your last 3 posts. Did you really have to do that?
Amiel, I think you needed to look at Cat's two posts prior to that (#184 and #185) to understand Bill's motive and defence for doing so. Bit of fairness please.

Cat
20-09-2006, 09:38 PM
Well you see Amiel, the Cat does not get to tell me what I meant when I make a post.
I know exactly what I meant by my post #132 and I stated it in post #176.

No, I don't need to say anything. You've said it yourself, it's plain for all to see!

Desmond
20-09-2006, 09:42 PM
No, I don't need to say anything. You've said it yourself, it's plain for all to see!
Do I need lemon juice to read it, because I have no idea wtf you are on about.

arosar
20-09-2006, 09:45 PM
Amiel, I think you needed to look at Cat's two posts prior to that to understand Bill's motive and defence for doing so. Bit of fairness please.

Yeah I'm fair. But these presidents should control themselves you know. I mean, if a bloke here like Shirty can, then should a president, you know.

AR

Bill Gletsos
20-09-2006, 10:27 PM
No, I don't need to say anything.Whats the matter, cat got your tongue. :owned:

You've said it yourself, it's plain for all to see!Correct post #176 says it all. :hand:

MichaelBaron
21-09-2006, 12:17 AM
Meaningless to you! My ACF rating (1242 don't laugh.:evil: ) is the only one I've got. Thanks for thinking of the little guy.:whistle:

Lets play an ACF rated match..you can have my rating

Garvinator
21-09-2006, 01:33 AM
Lets play an ACF rated match..you can have my rating
the acf doesnt rate matches, except for any Aus title playoffs ;) :P

Bill Gletsos
21-09-2006, 01:35 AM
the acf doesnt rate matches, except for any Aus title playoffs ;) :PNot entirely correct.
The ACf rates playoff matches for state championships and club championships.

Cat
21-09-2006, 07:31 AM
This is not true as what you are measuring is also changing in time. At each point of time the performances could be normally distributed and over all time for that matter they could also be normally distributed but at discrete times the mean could be shifting and it is the mean of the next period that we are trying to predict.

Actually I really like this discription, Barry. It reminds me of Zeno's arrow, which occupies an infinite number of stationary positions and can never quite get to it's destination.

Rincewind
21-09-2006, 08:14 AM
Actually I really like this discription, Barry. It reminds me of Zeno's arrow, which occupies an infinite number of stationary positions and can never quite get to it's destination.

Never quite reaching the destination is actually a modified version of Zeno's "Achilles and the tortoise" paradox. The arrow paradox was actually more about dividing time into discrete instantaneous points. But anyway there are three things to be said about Zeno's paradoxes in the context of this discussion.

Firstly, motion is patently possible (just as rating movement is patently possible). Secondly, if true, and a actually rating is not attainable then this is only possible by considering very small discrepancies. As ratings themselves are only publish as natural numbers, the rating and true strength could easily become the same for all practial purposes. Finally, were Zeno paradoxes to apply you would have to ignore stocastic effects which allow for overshooting and undershooting of a rating. With stocastic effects, at best a rating is always giggling around the true strength, sometimes over and sometime under and therefore paradoxes of Zeno's type would seem to be totally inapplicable.

peter_parr
21-09-2006, 10:50 AM
Gletsos does not recall making any manual adjustment and does not recollect a player losing over 700 rating points in one game.

I have Total Recall.

This matter was brought to my attention by a member of my staff.
The name was Frean - Greg Frean - State was Victoria.

He was making a welcome return to competitive chess after an absence of 20 years. He played and lost one game and Glicko/Gletsos whacked him chopping off over 700 rating points.

If his absence had been longer and his opponent lower rated he may have been whacked over 1000 rating points in one game.

I totally recall explaining this to Gletsos in my shop in the CBD and he, my staff and customers were all amazed at this large Glicko Glitch.

Gletsos after Deep Thought returned hundreds of rating points in a manual adjustment. My recollection remains crystal clear.

If Frean had beaten Rogers in this one game and Glicko added about 450 rating points to make Frean over 2650 I am 100% certain Gletsos would have whacked Glicko's calculations by about 400 rating points.

When I pointed out once again that only in Australia you can lose or gain hundreds of rating points in a few days Gletsos replies that this is Glicko's main benefit - thanks for reminding everyone.

Under FIDE regulations an inactive player and active player are treated equally for future rating calculations - this is the accepted practice by FIDE for over 30 years as approved by its 150 member countries.

Brilliant Mathematician Dr John Nunn - the youngest to enter Oxford University in 500 years and leading Grandmaster on 8th June 2005 wrote a lengthy article supporting this view and there should be no negative consequences for inactivity.

There is no positive effect and an unquestionable negative effect. Inactive players should be welcomed back without punishment.

Gletsos please note that my learned friend Dr Nunn continues - It is a strange quirk about chess players who are enthusiastic about punishing his colleagues.

BTW England uses the ECF grading system (not BCF).

In July 2006 IPP rated 2266? Scored 4/7 in 2 days and lost 217 rating points - it would be 39 rating points lost under FIDE. I believe FIDE is a sensible system as a 15 rating points loss is a lot (but fair as a maximum for losing one game).

Note FIDE rated players 2400+ lose a maximum of 10 rating points (not 15) for losing to say a 1400 rated player - sensible.

Even if this was doubled IPP would have lost 78 rating points not 217.
Players who intend to return to chess after a break must be encouraged to play again. If they are severely punished we will (and do) lose them forever.

If they do well or badly in a tournament they should be treated the same as other players in the tournament. There should be no discrimination.

One question for Gletsos.
A player rated 2405 loses to a player 1405 he loses 10 rating points and the 1405 player gains 15 rating points under FIDE.

How many rating points (maximum) does the 2405 player lose and the 1405 player gain under Glicko?

Simple question - simple answer.

Kevin Bonham
21-09-2006, 11:27 AM
Note FIDE rated players 2400+ lose a maximum of 10 rating points (not 15) for losing to say a 1400 rated player - sensible.

I don't think it's sensible at all.

A 2400 player should practically never lose to a 1400 player and a 1400 player should practically never beat a 2400 player. Probably this would happen less than 1 in 200 games with such a rating difference. If such an extraordinary event occurred it could tell us far more about the players' actual playing strength than FIDE's 0.4% decrease in the rating of the stronger and 1.1% increase in the rating of the latter would suggest.


Players who intend to return to chess after a break must be encouraged to play again. If they are severely punished we will (and do) lose them forever.

I agree that the risk of mathematically justified rating point loss on return (this is not an inactivity penalty such as Nunn was talking about but occurs only if the player performs below their rating) may put some players off returning, although I think the number of players thus put off is quite small.

One solution to the problem would be to declare that a player's rating is only valid when they have been active in the last five years, and therefore that a player who is inactive for longer than that has their old rating stored and used to seed their new rating, and can have it used to seed them in tournaments (at the organisers' discretion), but is officially unrated. That way there would be no disincentive for inactive players to return, because an inactive player could not say "I have a rating of 2000 and if I return it might drop to 1700". All they could say is "I used to have a rating of 2000 when last active."

Inactive players on the master list could be listed as UNR(xxxx) where xxxx is their old rating.

There would be no disincentive to return because a player who officially has no current rating cannot lose ratings points.

pax
21-09-2006, 11:57 AM
Gletsos does not recall making any manual adjustment and does not recollect a player losing over 700 rating points in one game.

I have Total Recall.

This matter was brought to my attention by a member of my staff.
The name was Frean - Greg Frean - State was Victoria.

He was making a welcome return to competitive chess after an absence of 20 years. He played and lost one game and Glicko/Gletsos whacked him chopping off over 700 rating points.

If his absence had been longer and his opponent lower rated he may have been whacked over 1000 rating points in one game.

I totally recall explaining this to Gletsos in my shop in the CBD and he, my staff and customers were all amazed at this large Glicko Glitch.

There is a technical flaw with ordinary Glicko under which it is possible for a previously inactive player to drop to a rating lower than his performance rating for the period (if performance was lower than the previous rating). Some time ago, Bill incorporated a sanity check into the ACF ratings such that a player's rating cannot drop lower than the performance rating for the period (or vice versa).

Is it this correction that fixed Frean's rating? If so, then it was not a 'manual adjustment' at all, but rather a fix to a loophole in the system.

Bill Gletsos
21-09-2006, 01:56 PM
There is a technical flaw with ordinary Glicko under which it is possible for a previously inactive player to drop to a rating lower than his performance rating for the period (if performance was lower than the previous rating). Some time ago, Bill incorporated a sanity check into the ACF ratings such that a player's rating cannot drop lower than the performance rating for the period (or vice versa).

Is it this correction that fixed Frean's rating? If so, then it was not a 'manual adjustment' at all, but rather a fix to a loophole in the system.No, because Peters post is full of factual flaws.
You would think people who are going to make claims (even beatup claims) could get the most basic of facts right.

Frean on returning to chess in the Dec 2000 rating period played 2 rated games not one as claimed by Peter. His performance rating was around 1791.

Frean's published rating in August 2000 was 2304 which included the 150 point uplift from April 2000. Frean had not played a rated game from at least 1987. Frean's published rating in December 2000 based on his 2 rated games was 1873??. No manual change was made to Freans rating. Note that his published April 2001 rating was also 1873??.

arosar
21-09-2006, 02:29 PM
So Bill, in your expert opinion, this Glicko system is the best? And if there are any flaws we can manage them, right, with appropriate corrections?

That's all I wanna know and care about mate.

AR

Bill Gletsos
21-09-2006, 02:37 PM
So Bill, in your expert opinion, this Glicko system is the best?I've never claimed its perfect just that it is better than anything else around.

And if there are any flaws we can manage them, right, with appropriate corrections?Yes, but corrections are not made on an individual player basis but changes are made to the system as a whole.

arosar
21-09-2006, 02:46 PM
In that case Bill, I trust you on these things. Now, any other challenger should simply write a paper, like a proper academic one, outlining why their idea is better and ought to be implemented. Then we publish that paper and examine it for peer review, you know like normal.

AR

Denis_Jessop
21-09-2006, 03:57 PM
I mention that the Chess Tactics Server about which there is a separate thread also uses Glicko to rate its participants. So does FICS.

DJ

MichaelBaron
21-09-2006, 05:04 PM
The fide rating floor is now down to 1400 (i think). That should encourage all of the organizers to make their events fide-rated. That should be sufficient to attract players back to competitive chess after a long lay-off. After all, the Elo rating has been there for many years. It became recognized long before Bill became the ACF rating officer and will probably remain as the internationally recongnized rating system after he is gone. So lets not worry too much about ACF rating anyway. Rating officers come and go..but Elo rating remains!:hmm:

antichrist
21-09-2006, 05:16 PM
The fide rating floor is now down to 1400 (i think). That should encourage all of the organizers to make their events fide-rated. That should be sufficient to attract players back to competitive chess after a long lay-off. After all, the Elo rating has been there for many years. It became recognized long before Bill became the ACF rating officer and will probably remain as the internationally recongnized rating system after he is gone. So lets not worry too much about ACF rating anyway. Rating officers come and go..but Elo rating remains!:hmm:

One advantage of Glicko is that it gives overseas blow-ins an over-inflated opinion of their strength and they lose out accordingly.

pax
21-09-2006, 05:37 PM
The fide rating floor is now down to 1400 (i think). That should encourage all of the organizers to make their events fide-rated. That should be sufficient to attract players back to competitive chess after a long lay-off. After all, the Elo rating has been there for many years. It became recognized long before Bill became the ACF rating officer and will probably remain as the internationally recongnized rating system after he is gone. So lets not worry too much about ACF rating anyway. Rating officers come and go..but Elo rating remains!:hmm:

I don't know if it's worth it. FIDE rating fees are quite high (a lot more expensive than ACF for example). Time controls for FIDE rating are also more strict than ACF, meaning that most weekenders (hence most Australian tournaments) do not qualify.

MichaelBaron
21-09-2006, 07:29 PM
I don't know if it's worth it. FIDE rating fees are quite high (a lot more expensive than ACF for example). Time controls for FIDE rating are also more strict than ACF, meaning that most weekenders (hence most Australian tournaments) do not qualify.


As far as the fees are concerned, i do agree...:hmm:

Bill Gletsos
22-09-2006, 11:18 AM
Gletsos does not recall making any manual adjustment and does not recollect a player losing over 700 rating points in one game.That is because it never happened.

I have Total Recall.Not even close.
Frean played 2 rated games on his return in Dec 2000 not one game as can be seen from the published rating list which appeared in Shaun Press's magazine Australian Chess Forum (a version of which is also on Barry Cox's site). :owned:

This matter was brought to my attention by a member of my staff.
The name was Frean - Greg Frean - State was Victoria.
He was making a welcome return to competitive chess after an absence of 20 years. He played and lost one game and Glicko/Gletsos whacked him chopping off over 700 rating points.You really need to stop making stuff up.

Frean in August 2000 was rated 2304. This included the 150 point uplift from April 2000. He hadnt played a rated game since at least 1987.

Now as noted above on return to chess Frean didnt play one rated game he played two. His performance rating was around 1791.
His December 2000 rating was 1873?? as can seen in Shaun Press's magazine Australian Chess Forum.
I dont know where you learnt maths but 2304 - 1873 is 431 points and not over 700 points as you are claiming.

If his absence had been longer and his opponent lower rated he may have been whacked over 1000 rating points in one game.You really shouldnt go making things up.

I totally recall explaining this to Gletsos in my shop in the CBD and he, my staff and customers were all amazed at this large Glicko Glitch.

Gletsos after Deep Thought returned hundreds of rating points in a manual adjustment.Total rubbish.

My recollection remains crystal clear.If that is how clear you memory is your need it tested.

If Frean had beaten Rogers in this one game and Glicko added about 450 rating points to make Frean over 2650 I am 100% certain Gletsos would have whacked Glicko's calculations by about 400 rating points.More rubbish.

When I pointed out once again that only in Australia you can lose or gain hundreds of rating points in a few days Gletsos replies that this is Glicko's main benefit - thanks for reminding everyone.Yes thanks for doing it.

Under FIDE regulations an inactive player and active player are treated equally for future rating calculations - this is the accepted practice by FIDE for over 30 years as approved by its 150 member countries.That doesnt make it right.

Brilliant Mathematician Dr John Nunn - the youngest to enter Oxford University in 500 years and leading Grandmaster on 8th June 2005 wrote a lengthy article supporting this view and there should be no negative consequences for inactivity.You are taking Nunn out of context. Nunn is talking about the idea that a certain number of points are deducted for each period a player is inactive automatically.
The Glicko system does not do this at all.

There is no positive effect and an unquestionable negative effect. Inactive players should be welcomed back without punishment.They arent being punished. Their rating on return to active play is based on their performance, not their strength 8, 10, 15 or 20 years previously.

Gletsos please note that my learned friend Dr Nunn continues - It is a strange quirk about chess players who are enthusiastic about punishing his colleagues.Nunn's article was written in June 2005. He hadnt played a FIDE rated game since the July 2003 rating period.
Nunn wouldnt have a vested interest here in discrediting the ACP (Association of Chess Professionals) proposal now would he. :whistle: :whistle:

BTW England uses the ECF grading system (not BCF).Essentially the same as the British Chess Federation simply changed its name to the English Chess Federation.

In July 2006 IPP rated 2266? Scored 4/7 in 2 days and lost 217 rating points - it would be 39 rating points lost under FIDE. I believe FIDE is a sensible system as a 15 rating points loss is a lot (but fair as a maximum for losing one game).this is just a rehash of post #37. As was noted in reply #40 IPP had a performance rating of around 1863. Prior to this event he last played in in Dec 2004 when he played 6 games. Prior to that was 1 game in Dec 2001 and prior to that back in October 1998.

Note FIDE rated players 2400+ lose a maximum of 10 rating points (not 15) for losing to say a 1400 rated player - sensible.It isnt sensible at all.

Even if this was doubled IPP would have lost 78 rating points not 217.And based on his results he would be significantly overrated.

Players who intend to return to chess after a break must be encouraged to play again. If they are severely punished we will (and do) lose them forever.

If they do well or badly in a tournament they should be treated the same as other players in the tournament. There should be no discrimination.
They are not being discriminated against, however the accuracy of their 8, 10 or 15 year old rating is highly questionable compared to the rating of an active player.

One question for Gletsos.
A player rated 2405 loses to a player 1405 he loses 10 rating points and the 1405 player gains 15 rating points under FIDE.

How many rating points (maximum) does the 2405 player lose and the 1405 player gain under Glicko?

Simple question - simple answer.If the question and answer is so simple then surely a man of your expertise in ratings can do the Glicko calculations and work it out. :whistle:

antichrist
22-09-2006, 04:18 PM
Bill, have you been takin lessons from Brian Lara? This comment could cost me an 8 day suspension and probably rendition!

Cat
22-09-2006, 11:01 PM
Finally, were Zeno paradoxes to apply you would have to ignore stocastic effects which allow for overshooting and undershooting of a rating. With stocastic effects, at best a rating is always giggling around the true strength, sometimes over and sometime under and therefore paradoxes of Zeno's type would seem to be totally inapplicable.

Oh I don't know, I think the main point is that neither Zeno's arrow nor the Glicko arrow ever reach their target. Whether you represent that as a lnear progression, stochastic variation or in any other way is simply academic.

Rincewind
22-09-2006, 11:45 PM
Oh I don't know, I think the main point is that neither Zeno's arrow nor the Glicko arrow ever reach their target. Whether you represent that as a lnear progression, stochastic variation or in any other way is simply academic.

That is simply untrue. Stochatic variation means you will go past your target.

Cat
23-09-2006, 12:50 AM
That is simply untrue. Stochatic variation means you will go past your target.

The oscillations are never quite fine enough to get there.

Kevin Bonham
23-09-2006, 12:52 AM
Oh I don't know, I think the main point is that neither Zeno's arrow nor the Glicko arrow ever reach their target. Whether you represent that as a lnear progression, stochastic variation or in any other way is simply academic.

Irrelevant because (i) a rating is only ever approximate (ii) players perform both above and below their actual strength at various times so sooner or later a player will get a rating above their actual strength then bob about it.

Rincewind already explained this to you, but you were far too dense to comprehend it.

Bill Gletsos
23-09-2006, 12:54 AM
but you were far too dense to comprehend it.The cat too dense. :doh:
Say it aint so. :whistle:

Kevin Bonham
23-09-2006, 01:03 AM
The oscillations are never quite fine enough to get there.

So what? Will the sky fall if someone's rating is 10 points out from their actual playing strength at the time?

Sutek
23-09-2006, 09:41 AM
Hi Bill,

Here's a question for you.

My rating is 2240??

If I played one game against Ian Rogers (2613!!) and won what would my new rating be?

Regards
Sutek

MichaelBaron
23-09-2006, 10:42 AM
Hi Bill,

Here's a question for you.

My rating is 2240??

If I played one game against Ian Rogers (2613!!) and won what would my new rating be?

Regards
Sutek

I too got a question for Bill. The story goes as follows: "Some years ago, Nick Speck made a comeback to competitive chess after a long break and won the Australian championship with a 2600+ performance. According to Glicko, he was supposed to be 2600 on the next list. Yet, apparently he never recieved that rating due to some "manipulations" by a rating officer who did not want to ridicule the rating system by allowing Nick to be rated above Rogers."

I do not know if this story is true or not (Since i have not heard it from Nick himself but from somebody else). So could you please confirm that there were no "special adjustments" done to Speck's rating?

Bill Gletsos
23-09-2006, 11:08 AM
I too got a question for Bill. The story goes as follows: "Some years ago, Nick Speck made a comeback to competitive chess after a long break and won the Australian championship with a 2600+ performance. According to Glicko, he was supposed to be 2600 on the next list.This isnt in the least bit accurate.

Yet, apparently he never recieved that rating due to some "manipulations" by a rating officer who did not want to ridicule the rating system by allowing Nick to be rated above Rogers."

I do not know if this story is true or not (Since i have not heard it from Nick himself but from somebody else). So could you please confirm that there were no "special adjustments" done to Speck's rating?This crap has been going around for ages.
No special adjustments were made to speck's rating.

Kevin Bonham
23-09-2006, 11:10 AM
I too got a question for Bill. The story goes as follows: "Some years ago, Nick Speck made a comeback to competitive chess after a long break and won the Australian championship with a 2600+ performance. According to Glicko, he was supposed to be 2600 on the next list. Yet, apparently he never recieved that rating due to some "manipulations" by a rating officer who did not want to ridicule the rating system by allowing Nick to be rated above Rogers."

*groans*

The story is an absolutely false beatup that has been refuted on here many, many times but continues to spread like a bad email urban myth. It serves only to confirm the adage that a rumour without a leg to stand on will find some other way to get around.

It is notable that the story, absurd enough in its original form, has now grown further in the telling to the point where it was suggested that Speck would have got a rating higher than Rogers. This was never in the original version of the rumour and the obvious reason for that is that Speck's performance rating was lower than 2600.

This is what Bill said about it in April 2003:


I orginally heard the rumour that people (mainly Victorians) were theorising that speck could reach as high as number 3 on the April 2002 ACF rating list. I believe the person who told me this was Nick Kordahi. This was well before the April 2002 ratings were even submitted for rating. I remember scoffing at the idea to Nick Kordahi. Around about the same time this rumour that speck would reach the number 3 position appeared on the ACF bulletin board, as well as discussion of speck's performance rating.

I pointed out that although people were quoting speck's performance rating as 2490 something, his actual performance rating was around 2532. The reason I said this was because although the traditional means of calculating a performance rating is to take the average of your opponents ratings then add the difference determined from the probability tables based on your score percentage, this only gives an approximation. A players true performance rating is the value for which his expected score matches his actual score.

In order to quell any rumours about speck being number 3 on the list I manually calculated speck's rating as no files had been submitted for rating at this time. His rating came out at 2411. However since it is our policy not to pre-publish any players rating, I said on the Bulletin Board that speck's rating would under no circumstance exceed 2420. I certainly never mentioned a figure of 2490.

Vlad
23-09-2006, 11:32 AM
I too got a question for Bill.

Michael, you surprise me as an individual who on the one hand is continuously claiming that he is super intellectual. On the other hand your way of arguing about the ratings is the same as cat's. You are talking about specific example of Speck’s rating in 2002, right?

Take glicko code provided on one of the bulletin pages, submit the result of Speck in 2002 and find what the program gets. Then you may ask why your result is different from what Bill calculated (if it is different of course).

I have got an impression that people who can do calculations themselves are perfectly happy with the glicko. Prove me that I am wrong.:evil:

Bill Gletsos
23-09-2006, 11:33 AM
Hi Bill,

Here's a question for you.

My rating is 2240??

If I played one game against Ian Rogers (2613!!) and won what would my new rating be?

Regards
SutekHi Steve,
You havent played a rated OTB game for about 15 years.
If you were to beat Ian in that single game your rating would be around 2560??
If you were to lose to him it would be 2201?? and if you were to draw then it would be 2381??.

Kevin Bonham
23-09-2006, 11:50 AM
Hi Steve,
You havent played a rated OTB game for about 15 years.
If you were to beat Ian in that single game your rating would be around 2560??

In my view such a rise would be clearly excessive. I think sharp drops based on small numbers of games are pretty likely to be correct from a predictiveness viewpoint but I doubt that the same would apply to sharp rises of this sort.

arosar
23-09-2006, 11:55 AM
Well, this is fun! Bill, if I were to beat Mr Rogers, what would my rating be please?

AR

Bill Gletsos
23-09-2006, 11:58 AM
Well, this is fun! Bill, if I were to beat Mr Rogers, what would my rating be please?Still bugger all at around 1753. ;)

Rincewind
23-09-2006, 12:05 PM
Still bugger all at around 1753. ;)

Hey, I take offense to that remark. :)

So Amiel's rating is only one win over Rogers away from mine. :(

;)

Cat
23-09-2006, 01:09 PM
Michael, you surprise me as an individual who on the one hand is continuously claiming that he is super intellectual. On the other hand your way of arguing about the ratings is the same as cat's. You are talking about specific example of Speck’s rating in 2002, right?

Take glicko code provided on one of the bulletin pages, submit the result of Speck in 2002 and find what the program gets. Then you may ask why your result is different from what Bill calculated (if it is different of course).

I have got an impression that people who can do calculations themselves are perfectly happy with the glicko. Prove me that I am wrong.:evil:

Yeah you are wrong, and there are many who are not happy at all. You have a very shakey sense of diction, drug. Proof in this sense is what exactly? The fact that numerous individuals have post their grievances, written, etc - isn't that enough? How should it be spelt out, exactly. Maybe you could suggest a framework through which people could voice their opinion, that you would consider as legitimate and could be responded to without personalisation of the isseus.

Bill Gletsos
23-09-2006, 01:23 PM
Yeah you are wrong, and there are many who are not happy at all.Being unhappy doesnt make them right or that they know what they are talking about.

You have a very shakey sense of diction, drug.Yes, but unlike you, he isnt clueless.

Proof in this sense is what exactly?Unlike you drug knows what he is talking about.

The fact that numerous individuals have post their grievances, written, etc - isn't that enough?Not when those that have issues have no skill in statistical or ratings theory.

arosar
23-09-2006, 01:23 PM
The fact that numerous individuals have post their grievances, written, etc - isn't that enough?

I think this reasoning is not very correct Cat. Else, it's just a rating system by populism.

AR

Bill Gletsos
23-09-2006, 01:24 PM
I think this reasoning is not very correct Cat.Dont expect correct reasoning by Cat.

Else, it's just a rating system by populism.:clap: :clap:

Vlad
23-09-2006, 02:06 PM
The fact that numerous individuals have post their grievances, written, etc - isn't that enough?

That sounds similar to a strategy my 5-year-old adopts sometimes. If he wants something he starts crying. Is not what you are doing all the time? The only difference I suppose you are a bit older.

Thousands years ago people knew that the worst ruler is the one who tries to please everybody. That implies if we have a good rating officer there will always be people who post their grievances. :owned:

pax
23-09-2006, 08:29 PM
In my view such a rise would be clearly excessive. I think sharp drops based on small numbers of games are pretty likely to be correct from a predictiveness viewpoint but I doubt that the same would apply to sharp rises of this sort.

Just how likely is it that a player who is 2240 and hasn't played for 15 years is going to beat Rogers?? Even if he did, it's highly unlikely that that would be the only game he played in a period.

Kevin Bonham
23-09-2006, 08:38 PM
Just how likely is it that a player who is 2240 and hasn't played for 15 years is going to beat Rogers?? Even if he did, it's highly unlikely that that would be the only game he played in a period.

Agree completely. Although it's worth mentioning that the rating system has no idea whether somebody has played or not in the last 15 years. It only knows whether they've played rated OTB chess in (or in rare cases near) Australia.

Although it's theoretically possible for a returning player to go up hundreds of points from a single game, I don't think there are any known cases of it happening.

Cat
24-09-2006, 10:09 AM
Barry, cutting legitimate posts is a sad day. I want my response to drug's post to stay. It's especially ironic given your behavior right now

Rincewind
24-09-2006, 10:12 AM
Barry, cutting legitimate posts is a sad day. I want my response to drug's post to stay. It's especially ironic given your behavior right now

Vlad, if you really want to read this reply you can find it here (http://chesschat.org/showthread.php?t=4974).

peter_parr
25-09-2006, 03:27 PM
Gletsos post number 214 contains a litany of nonsense as has been seen before by him.

My post 202 is correct in every detail.

Frean did lose over 700 rating points as in my post 202.
As soon as the rating list came out I decided to check who had lost the most rating points knowing how Glicko works.

When Frean had lost over 700 rating points I told Gletsos in my shop in the CBD. (Thus the term he’s been Glicko’d was born).
Gletsos checked it out – agreed it was too many to lose and made a manual adjustment as stated by me. I am not a liar nor do I make up stories.

Gletsos (to his credit) makes corrections to the ratings as is required as soon as problems arise. Of course by the time the rating list is published the corrections have already been made and the amended list is different to the original list.

The fact is Frean did drop over 700 rating points which was amended by Gletsos before publication in the magazine.

The manual adjustment by Gletsos did involve hundreds of rating points.
Gletsos in post 214 claims a loss of 431 points for 2 games played and claims he made no adjustment.

Even if it were true it shows once again how absurd our rating system is if you can lose 431 rating points in 2 games (30 not 431 under FIDE).
Gletsos claims this a benefit – it is not – it is absurd.

My memory as stated is crystal clear and I do not need my memory tested – it is functioning perfectly.

Gletsos is critical of both the FIDE rating system and GM John Nunn. To suggest Nunn is writing in a biased manner is outrageous and inaccurate.

If you knew John Nunn personally you would not make such disgraceful comments.

One reason the Glicko system (invented in the USA) is not used by US Chess Federation is that due to the USCF rating floor a player cannot drop more than 200 rating points in a lifetime (unless a player makes a special application to reduce his rating – usually in old age).

I agree with comments made by Brian Jones on that subject.
As we see from IPP in the July weekender he lost 217 rating points in 2 days – more than a lifetime loss in the USA.

Each time I quote a major factor in FIDE rating system you say it isn’t sensible at all – even though it is used world wide.

You state in post 214 I can work out how much a player can gain or lose from one game under Glicko. Well I cannot. How can anyone?

I ask the question to Gletsos as the 1000 rating loss in one game may not be quite accurate – maybe less maybe more.

You advise Sutek if he played one game against Rogers (rated 373 above him) and won his rating would rise 320 rating points. If he played one game against the NSWCA Secretary rated 1179 rating points below him and lost the same mathematics would he not lose over 1000 rating points on one game?

Thus if Sutek played one game lasting a maximum of 2 hours depending on his opponent and result he could go up by 320 or drop 1000 rating points – A fluctuation of 1320 rating points in 2 hours in one game.

Don’t blame me – blame Glicko – (FIDE fluctuation would be 30 not 1320)

Baron’s post 224 is well noted and is continually discussed by many players. The FIDE system is precise – I can always work out a rating under FIDE.

Last Thursday afternoon a customer in my shop in the CBD was purchasing chess trophies and medallions for the school where he teaches chess. He joined in the conversation at the time about the ratings. He had himself been very upset that he had lost a lot of rating points from one game against a player of similar rating. After complaining to Gletsos about the loss of points from one game Gletsos agreed it was excessive and made a manual upward adjustment to his rating.

How often manual adjustments are made I do not know nor do I object to them.
What I do object to is when Gletsos launches a vicious attack against me when it is he who is not telling the truth.

The worst example of course was just before the NSWCA AGM when he denied making any agreement with Lyons over the NSWCA Presidency.

There were so many witnesses to the contrary but I have no intention of conducting a vendetta (as suffered by Sweeney).

Last Friday afternoon a former NSWCA President was in my shop in the CBD. He had played a number of events in overseas countries – he cited a major problem in the USA that some players over time were losing 100-200 rating points and winning lower rated prizes. He told me what he thought of the Glicko rating system.

The Glicko system is sometimes used in one minute, three minute or 5 minute bullet games. The rating system is not used by any other country due to it’s volatility for both active and less active players.

Our system was similar to FIDE – it should still be similar to FIDE.

If NSWCA (Gletsos) had agreed with Jamieson and all other states (who agreed with Jamieson) to have the ACF run by a commission of 8 experts (real experts) the Glicko discussion would have been finalised by the experts in a very short time.

The rating systems adopted by national federations are all very similar to the FIDE rating system. If a player competes in a tournament in Australia and it is FIDE rated the player should not go down 217 (Glicko) rating points when FIDE says 39 down for the same games.

Brian_Jones
25-09-2006, 03:39 PM
I agree with comments made by Brian Jones on that subject.

Pardon. What are you talking about. Which comments?



If NSWCA (Gletsos) had agreed with Jamieson and all other states (who agreed with Jamieson) to have the ACF run by a commission of 8 experts (real experts) the Glicko discussion would have been finalised by the experts in a very short time..

When was this?

pax
25-09-2006, 03:43 PM
The rating systems adopted by national federations are all very similar to the FIDE rating system. If a player competes in a tournament in Australia and it is FIDE rated the player should not go down 217 (Glicko) rating points when FIDE says 39 down for the same games.

The BCF/ECF system is not remotely similar to FIDE..

Kevin Bonham
25-09-2006, 04:12 PM
When Frean had lost over 700 rating points I told Gletsos in my shop in the CBD. (Thus the term he’s been Glicko’d was born).
Gletsos checked it out – agreed it was too many to lose and made a manual adjustment as stated by me. I am not a liar nor do I make up stories.

The evidence we have before us here is:
(i) from Bill Gletsos, comments about past ratings lists that can presumably be tested by checking the ratings lists in question.
(ii) from Peter Parr, claimed memories of a private conversation that nobody here has any way of testing or checking.

I'm just not interested in "evidence" of the latter kind at all. It's notable Peter brings up the Lyons presidency thing again when that also hinged on unverifiable alleged private conversations.

If Frean lost 700 ratings points it should be possible to produce ratings lists that show that this occurred.


One reason the Glicko system (invented in the USA) is not used by US Chess Federation is that due to the USCF rating floor a player cannot drop more than 200 rating points in a lifetime (unless a player makes a special application to reduce his rating – usually in old age).

This is interesting because it shows us that the reason for non-use of Glicko in the USA is not any defect of the Glicko, but a defect of tournament chess as it is run in that illustrious nation - a defect mentioned further down in your post.

The reason the USCF has ratings floors is that too many American tournaments offer greatly excessive ratings prizes (over $10000 in some cases) causing players to deliberately trash their own rating in search of $$$.

The USCF therefore cannot afford to have a highly responsive rating system because people will abuse it for personal gain.

If the US did not have such high ratings prizes then they would be able to use Glicko like we do, and quite probably would.


If he played one game against the NSWCA Secretary rated 1179 rating points below him and lost the same mathematics would he not lose over 1000 rating points on one game?

No, he wouldn't. Bill can give you the exact figure for maximum possible loss from one game but while high it is definitely in three figures. Even if a GM loses to a player rated zero they will still not lose 1000 points.

Barry's Glicko 1 Calculator (http://www.bjcox.com/modules.php?name=Glicko_Calc) is a good tool for getting a rough feel for how much difference a particular result might make, although it uses Glicko-1 and Glicko-2 is often slightly more dynamic.

Bill Gletsos
25-09-2006, 04:36 PM
Gletsos post number 214 contains a litany of nonsense as has been seen before by him.The only one sprouting nonsense is you.

My post 202 is correct in every detail.Not in the least. To start with Frean played 2 games not one as you stated.

Frean did lose over 700 rating points as in my post 202.
As soon as the rating list came out I decided to check who had lost the most rating points knowing how Glicko works.You knew nothing how Glicko worked as the Dec 2000 list was the first list ever done under Glicko.

When Frean had lost over 700 rating points I told Gletsos in my shop in the CBD. (Thus the term he’s been Glicko’d was born).Frean never lost over 700 points.

Gletsos checked it out – agreed it was too many to lose and made a manual adjustment as stated by me. I am not a liar nor do I make up stories.This is absolute rubbish. I made no manual adjustment.

Gletsos (to his credit) makes corrections to the ratings as is required as soon as problems arise. Of course by the time the rating list is published the corrections have already been made and the amended list is different to the original list.This is true because of coding errors not due to manual corrections. E.g. the latest ratings had an event coded as Normal when it should have been rapid. However in the case of the Dec 2000 list and Frean's rating no such change was made.

The fact is Frean did drop over 700 rating points which was amended by Gletsos before publication in the magazine.Total rubbish. No change was ever made to Frean's rating.

The manual adjustment by Gletsos did involve hundreds of rating points.Total rubbish. No adjustment was ever made.

Gletsos in post 214 claims a loss of 431 points for 2 games played and claims he made no adjustment.This is correct. He lost 431 points from 2 games. he never lost over 700 points as you claimed nor was any manual adjustment made.

Even if it were true it shows once again how absurd our rating system is if you can lose 431 rating points in 2 games (30 not 431 under FIDE).
Gletsos claims this a benefit – it is not – it is absurd.Nothing if about it being true at all. It is true. Your over 700 point claim is total rubbish.

My memory as stated is crystal clear and I do need my memory tested – it is functioning perfectly.Rubbish.

Gletsos is critical of both the FIDE rating system and GM John Nunn. To suggest Nunn is writing in a biased manner is outrageous and inaccurate.
If you knew John Nunn personally you would not make such disgraceful comments.Nunn's article was in response to an ACP proposal. It is plain for all to see he had a vested interest in their proposal as it have adversely affected his rating.

One reason the Glicko system (invented in the USA) is not used by US Chess Federation is that due to the USCF rating floor a player cannot drop more than 200 rating points in a lifetime (unless a player makes a special application to reduce his rating – usually in old age).You dont know what you are talking about.
That statement is incorrect and a player can drop more than 200 points in the USCF.
If they are rated above 1600 they can drop a maximum 299. because a player rated say 1899 has a floor of 1600, not 1699. Also for players rated below 1600 their rating floor is 100 not 1300. As such a player rated below 1600 can drop well over 300 points.
However the main reason why the USCF have rating floors is because of players sandbagging to win massive rating prizes.

I agree with comments made by Brian Jones on that subject.I guess wonders never cease.

As we see from IPP in the July weekender he lost 217 rating points in 2 days – more than a lifetime loss in the USA.Incorrect as I noted above.

Each time I quote a major factor in FIDE rating system you say it isn’t sensible at all – even though it is used world wide.That doesnt make it sensible.

You state in post 214 I can work out how much a player can gain or lose from one game under Glicko. Well I cannot. How can anyone?Easily if they understand Glicko by putting numbers into the Glicko formula on Glickmans website. It is just maths, not rocket science or brain surgery.

I ask the question to Gletsos as the 1000 rating loss in one game may not be quite accurate – maybe less maybe more.But you claimed yours was but a simple question with a simple answer. As such it should be simple for you to work it out.

You advise Sutek if he played one game against Rogers (rated 373 above him) and won his rating would rise 320 rating points. If he played one game against the NSWCA Secretary rated 1179 rating points below him and lost the same mathematics would he not lose over 1000 rating points on one game?No.

Thus if Sutek played one game lasting a maximum of 2 hours depending on his opponent and result he could go up by 320 or drop 1000 rating points – A fluctuation of 1320 rating points in 2 hours in one game.Your just making stuff up, yet again.

Don’t blame me – blame Glicko – (FIDE fluctuation would be 30 not 1320)It isnt 1320 at all. Stop making stuff up.

Baron’s post 224 is well noted and is continually discussed by many players. The FIDE system is precise – I can always work out a rating under FIDE.No, the FIDE system is easier to calculate. It isnt precise.

Last Thursday afternoon a customer in my shop in the CBD was purchasing chess trophies and medallions for the school where he teaches chess. He joined in the conversation at the time about the ratings. He had himself been very upset that he had lost a lot of rating points from one game against a player of similar rating. After complaining to Gletsos about the loss of points from one game Gletsos agreed it was excessive and made a manual upward adjustment to his rating.Rubbish.
Who is this supposed individual.

How often manual adjustments are made I do not know nor do I object to them.None as manual adjustments are not made.

What I do object to is when Gletsos launches a vicious attack against me when it is he who is not telling the truth.And I object to you making totally false statements regarding me and the ratings.

The worst example of course was just before the NSWCA AGM when he denied making any agreement with Lyons over the NSWCA Presidency.Thats because I made no such agreement.

There were so many witnesses to the contrary but I have no intention of conducting a vendetta (as suffered by Sweeney).Of course Jason wasnt even a member of the NSWCA at the time the AGM started. In fact according to Norm Greenwood, Jason never even attempted to pay his membership.

Last Friday afternoon a former NSWCA President was in my shop in the CBD. He had played a number of events in overseas countries – he cited a major problem in the USA that some players over time were losing 100-200 rating points and winning lower rated prizes. He told me what he thought of the Glicko rating system.Just because he thought of it doesnt mean it has any relationship to the actual situation.

The Glicko system is sometimes used in one minute, three minute or 5 minute bullet games. The rating system is not used by any other country due to it’s volatility for both active and less active players.This is nothing but supposition on your part.

Our system was similar to FIDE – it should still be similar to FIDE.

If NSWCA (Gletsos) had agreed with Jamieson and all other states (who agreed with Jamieson) to have the ACF run by a commission of 8 experts (real experts) the Glicko discussion would have been finalised by the experts in a very short time.

The rating systems adopted by national federations are all very similar to the FIDE rating system. If a player competes in a tournament in Australia and it is FIDE rated the player should not go down 217 (Glicko) rating points when FIDE says 39 down for the same games.All you are arguing for is inaccuracy over accuracy.

arosar
25-09-2006, 05:14 PM
FMD! How confusing is this? All I can say is that one of you must be a liar.

AR

antichrist
25-09-2006, 06:19 PM
FMD! How confusing is this? All I can say is that one of you must be a liar.

AR

As conversations are not recorded we can't confirm anything, even when events are recorded, i.e., JCL meeting minutes and my attendance, I still dispute them. So what hope is there.

But we can privately draw our own conclusions based on we think has been the more slippery and slimmery in the past.

MichaelBaron
25-09-2006, 06:50 PM
This debate is getting very confusing..i suggest the following.

1)If the manual adjustment has been made at least once..Bill should resign as a rating officer immidiately as he was abusing his official powers by doing so.
2) if manual adjustments have never taken place..we should move on to discuss the flows in the Glicko system itself rather than the handling of it.

:hmm:

arosar
25-09-2006, 06:58 PM
This debate is getting very confusing..i suggest the following.

1)If the manual adjustment has been made at least once..Bill should resign as a rating officer immidiately as he was abusing his official powers by doing so.

Even if he did, he might have had perfectly justifiable reasons.

It's now got to a point where we should just request bullet points. In Powerpoint, if possible.

Bill? Peter?

AR

Rincewind
25-09-2006, 07:41 PM
You dont know what you are talking about.
That statement is incorrect and a player can drop more than 200 points in the USCF.
If they are rated above 1600 they can drop a maximum 299. because a player rated say 1899 has a floor of 1600, not 1699. Also for players rated below 1600 their rating floor is 100 not 1300. As such a player rated below 1600 can drop well over 300 points.
However the main reason why the USCF have rating floors is because of players sandbagging to win massive rating prizes.

Further to this the USCF ratings committee have found their rating floors to be problematic. They have been investigating the problem of matches between players where one player is at or near his rating floor. Apparently these matches are organised because the other player can gain masses of rating points while the player close to his rating floor risks practically nothing. The outcome was to drop the ratnig floor by 100 points for the purposes of rating matches of this type. (refer USCF annual report 2006)

I can see no basis for implementing a rating floor. It would appear to be a hang over from a master points sort of system and serves no purpose other than to appease players in their declining years and build in some unneccesary inflation to the rating system.