Author: Frank Quisinsky
Date: 03:38:15 09/26/02
Go up one level in this thread
On September 26, 2002 at 06:09:53, Jouni Uski wrote: >Look at this link: http://f23.parsimony.net/forum50826/messages/49884.htm >(partly in english) >I remember always, that Crafty is said to play much better with ponder, but it >seems to be wrong... > >Jouni Hi Jouni, this is a old topic. I must thinking on the messages for 3-4 years here. Prof. Robert Hyatt worte his 100 scientific work here. In the future in 3-4 years we have the same discuss in an other fora :-)) Maybe it's better for looking if we create to standard topics a new fora :-) It's complete bubble things what different people have written about permanent brain/ponder (in my opinion). An engine with played with Ponder=On have more time for other moves. After my experiments for years we have in ~ 5-8 moves/game after the opening position - start of table base position). Normaly a chess game between engines with the ~ same rating is 62 moves ... if I look in my computer chess databases with more as 50.000 games for different engine lovers. So we have with ponder a complety other game compare if we played without ponder. Right is (also after my opinion) that we have different programs which plays better with pondering. Maybe Ponder ON = ~ 25 ELO in Eng-Eng Matches. Reason: Program A can play a mistake with ponder = off (no time for find the better move). But also Programm B can play a mistake with ponder with the same reason (no time). Other case in matches if Programm A plays with ponder and programm B plays without ponder. Maximal 20-30 ponder hits but only in 5-8 other moves/game. Now we see ~ 5-8 cases with other moves: It's not clear that this moves are better as the moves without ponder. Look (extrem example): Crafty calculate without ponder 6:36 and found a "good" move. Now Crafty played with ponder and played directly after 2:00 opponent time an other move in the same position. I have seen also this possibility. But if an engine plays with ponder and the other engines plays without ponder is this after my opinion max. 100-150 ELO. Clear is, that we have to 95% the same statistics in tourneys with or without ponder but this is not important because we are all not grandmaster and can not see that the games have not the same playing strength. Best Frank
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.