Author: Andrew Walker
Date: 19:35:52 03/15/04
Go up one level in this thread
On March 14, 2004 at 18:14:32, Joachim Rang wrote: >On March 14, 2004 at 12:42:13, Jouni Uski wrote: > >>... >>This is only reminder for all who have missed this. In Nickel's correspondence >>match >>engines get average 2,5 days for one move = 3600 minutes = equal to 1200*3 >>minutes. This means same performance as >>2,5 THz PC in 3 minutes play!! In another way 1 Gnps for Fritz/Junior!! So is >>this near >>perfect play now? In game 6 / Pluto engine is said to missed winning tactical >>16. g4!! move. >>In game 3 / Jupiter engine has played some dubious moves and is fighting for >>draw. In other >>games no fear for lose to programs I think. >> >>Jouni > >yes I follow this match with great interest. Intersting observation, that this >timecontrol is equal to a 2.5 GHz - CPU and 3 minutes per move. >So far it looks like the engines do pretty well. OTOH in game 6 Pluto missed the >winning 16.g4!! which shows clearly that this isn't perfect chess. It's also >clear that Pluto isn't Junior since Junior finds this move quite quickly. In >game 3 Arno simply found a very deep tactical combination which the program was >also unable to spot. In game 1 Arno also almost outplayed the engine but with >exact play the engine was able to defend very well. But in three games (!) the >engines were able to achieve better positions - this is quite amazing >considering the fact that Arno also may use engines. > >regards Joachim > For these long time control searches, wouldn't changing the selectivity settings be able to make the engine play stronger? Have any experiments been done with this? I'd like to know especially how this setting works for the engine I have, Hiarcs 9. Andrew
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.