Author: Uri Blass
Date: 00:24:17 02/28/05
Go up one level in this thread
On February 27, 2005 at 19:28:51, Arturo Ochoa wrote: >On February 27, 2005 at 18:54:54, Uri Blass wrote: > >>On February 27, 2005 at 18:12:48, Arturo Ochoa wrote: >> >>>On February 27, 2005 at 17:54:40, Uri Blass wrote: >>> >>>>On February 27, 2005 at 17:33:02, Arturo Ochoa wrote: >>>> >>>>>On February 27, 2005 at 14:06:58, Peter Skinner wrote: >>>>> >>>>>>On February 27, 2005 at 07:17:44, Sandro Necchi wrote: >>>>>> >>>>>>>Well, Vincent said that if an engine play Elo 3000 than the book is giving in >>>>>>>these tournaments 700 points and the program is 2300 Elo strong. >>>>>>> >>>>>>>I do not think Diep is playing Elo 3000 so the Vincent's rule cannot be used. >>>>>>> >>>>>>>Sandro >>>>>> >>>>>>Diep wasn't even playing at a 2400 level at the IPCCC. >>>>>> >>>>>>With all the games (well all games that were actually posted) from the event and >>>>>>using a start value of 2400, Diep scored: >>>>>> >>>>>>Program Elo + - Games Score Av.Op. Draws >>>>>> >>>>>> 1 Hydra : 2810 239 220 9 88.9 % 2449 22.2 >>>>>>% >>>>>> 2 Shredder : 2715 320 296 9 83.3 % 2436 11.1 >>>>>>% >>>>>> 3 Gandalf : 2566 214 205 9 61.1 % 2488 33.3 >>>>>>% >>>>>> 4 Spike : 2499 267 247 9 61.1 % 2420 11.1 >>>>>>% >>>>>> 5 Nexus : 2455 292 292 7 50.0 % 2455 14.3 >>>>>>% >>>>>> 6 Ikarus : 2447 220 220 7 50.0 % 2447 42.9 >>>>>>% >>>>>> 7 Anaconda : 2436 207 207 9 50.0 % 2436 33.3 >>>>>>% >>>>>> 8 Jonny : 2428 207 207 9 50.0 % 2428 33.3 >>>>>>% >>>>>> 9 SOS : 2418 207 207 9 50.0 % 2418 33.3 >>>>>>% >>>>>> 10 The Baron : 2417 187 183 9 55.6 % 2378 44.4 >>>>>>% >>>>>> 11 Diep : 2397 187 183 9 55.6 % 2358 44.4 >>>>>>% >>>>>> 12 Neurologic : 2218 177 194 8 37.5 % 2307 50.0 >>>>>>% >>>>>> 13 Patzer : 2208 177 195 9 33.3 % 2329 44.4 >>>>>>% >>>>>> 14 Quark : 2169 205 229 9 27.8 % 2335 33.3 >>>>>>% >>>>>> 15 IsiChess : 2150 248 286 8 25.0 % 2341 25.0 >>>>>>% >>>>>> 16 Matador : 2041 173 291 9 16.7 % 2321 33.3 >>>>>>% >>>>> >>>>>Rg. Titel Name Pkte Wtg. >>>>>1 Hydra 8 44½ >>>>>2 Shredder 7½ 43 >>>>>3 Gandalf 5½ 47 >>>>>4 Spike 5½ 42½ >>>>>5 Ikarus 5 42 >>>>>6 The Baron 5 39 >>>>>7 Diep 5 37 >>>>>8 Anaconda 4½ 44 >>>>>9 SOS 4½ 43 >>>>>10 Nexus 4½ 43 >>>>>11 Johnny 4½ 43 >>>>>12 Patzer 3 35½ >>>>>13 Neurologic 3 34 >>>>>14 Isichess 2½ 38½ >>>>>15 Quark 2½ 36 >>>>>16 Matador 1½ 36 >>>>> >>>>>http://wwwcs.upb.de/~IPCCC/IPCCC2005/r3.HTM >>>>>According to the Official Standing Table of teh Tournament, Diep was seventh and >>>>>not eleventh as you put in your table. >>>> >>>>The table is not officical standing table but performance table based on the >>>>event and the assumption that the average rating is 2400 so I do not see the >>>>problem with it. >>>> >>>>Diep played relatively weaker opponents(for example did not play hydra) so the 5 >>>>that it scored may be considered as worse than 4.5 that scored other programs. >>>> >>>>> >>>>>About your point, Diep was: >>>>> >>>>>1) 6th. in Paderborn 2004: http://wwwcs.upb.de/~IPCCC/IPCCC2004/ranking.html >>>>>2) 4th. in 4th International CSVN Tournament: >>>>>http://www.computerschaak.nl/ict4tour.html >>>>>3) 3rd. in 12th World Computer Chess Championship 2004: >>>>>http://www.cs.biu.ac.il/games/results.html >>>>>4) 1st. in Dutch Open 2004: http://www.computerschaak.nl/docc04.html. >>>>> >>>>>They are all the Official Tables and they don´t say that Diep was under 2400 >>>>>Elo. >>>> >>>>I agree that there is no proof for rating under 2400 >>>> >>>>Peter made the assumption that the average rating of the field is 2400 and the >>>>only thing that I can say is that we do not know the average rating of the >>>>field. >>>> >>>>> I do not see that your table is proving anything. >>>> >>> >>>The typical Uri Blass.... If Peter made an assumption without the complete games >>>and he put a 2400 elo as an initial value. It doesnt prove what a relative elo >>>is. >>> >>>Correction for the person who doesnt read: That is not my table. It is the >>>official table of Paderborn, Mr. Blass. The Skinner´s Table puts Diep in the >>>_11th_ place when it got the 7th. place. >> >>The skinner table is performance table and not the official table. >> >>> >>>You simply dont see because you dont read. >> >>I have no problem of reading. >>I do not confuse between official table and performance table. >>They are different. >> > >Simply, it is not valid because how he supposed the 2400 elo for this table? > >It is just useless. I agree that the 2400 is useless. We do not know rating and I think that we have no ssdf rating or WBEC rating for the participants even in part of the cases(different hardware and different versions and different books). It was possible to add 3000 elo to all engines so nobody will connect that performance with rating. The point that I responded was the place of Diep in the table. The place of Diep in peter's table is still correct or close to correct. You say that not all the games were used and it is possible that more games can change the place of diep in the table but it is clearly logical that diep will have worse performance table than place in the official table. Diep played the weakest participants and did not play hydra. Some players with 4.5 played stronger opponents so their 4.5 points give them better performance than diep. > >>I claimed nothing here about the value of book but only responded to your claim: >>"Diep was seventh and not eleventh as you put in your table." >> >> >>There is nothing wrong with performance table and there is nothing wrong with >>official table. >>They are different tables. >> >>Uri > >I dont see that the table says anything about the elo performance in numbers. >I can put 2000 or 2800 as the initial value and it doesnt prove anything. I agree. My point was about the ranking in the table and not about rating(I admit that the ranking may be wrong because of not using all the games but I see nothing illogical in place 11 for diep in performance table).
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.