Author: blass uri
Date: 21:01:20 05/07/00
Go up one level in this thread
On May 07, 2000 at 17:35:01, Enrique Irazoqui wrote: >On May 07, 2000 at 14:30:40, blass uri wrote: > >>On May 07, 2000 at 13:47:57, Peter Kappler wrote: >> >>>On May 07, 2000 at 09:01:57, Enrique Irazoqui wrote: >>> >>>>> F6 Rebel S4 DJ >>>>>Average opponent rating 2412 2369 2391 2297 >>>>>Games played 7 7 5 3 >>>>>Comp score 3 5 4 2 >>>>>TPR 2355 2541 2630 2430 >>>>> >>>>>Computers average TPR after 23 games: 2487 >>>> >>>>Sorry. It should be: >>>> >>>>Computers average TPR after 22 games: 2487 >>>> >>>>Enrique >>> >>> >>>Thanks for posting this, Enrique. >>> >>>2487 is almost exactly the same performance rating that Rebel achieved in the >>>20+ games it played last year against humans (Rebel challenge + match vs >>>Lithuanian team). It seems that we are finally getting a good idea of the >>>playing strength of the commercial programs. >>> >>>Do you know what hardware was used in Israel? >> >>The hardware was pIII500 in most of the cases. >> >>The games give only lower bound for the playing strength of programs because >>humans are not forced to play with the machine(the team can choose the player to >>play against the machine). >> >>The Israeli league gives only lower bound for the rating of computers because of >>another reason and the reason is that upgrade of programs is not allowed in the >>Israeli league. >> >>The GM challange gives only a lower bound because of the fact that GM's are not >>forced to play with the machine. > >I disagree. In the Israeli league programs are facing all sort of players, from >2000 to 2500+. This is more real than playing exclusively against the strongest >GMs, something that has been artificially inflating the ratings of the best >human players for a long time. Aside from this, the inability to update programs >seems quite irrelevant to me: let's say that this tpr applies to programs of 6 >months ago. > >Enrique The programs had to play hasidovski and this player(rating 2177) drew 3 programs. If the team of hasidovski had not the right to choose the player to play against programs I do not think they could get these good results. I believe that if humans learn to play against programs they can cause program to get less than 2400 but humans do not do it so practically the rating of programs should be 2550-2600. part of the players in the israeli league did not prepare against the computers inspite of the fact that the teams had the right to choose the player to play against the program. I believe that if the teams had not the right to choose the player to play against the programs the situation could be even worse for the humans. Uri
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.