Author: Robert Hyatt
Date: 09:41:15 10/03/98
Go up one level in this thread
On October 03, 1998 at 00:49:36, blass uri wrote: > >On October 03, 1998 at 00:22:11, Robert Hyatt wrote: > >>On October 02, 1998 at 21:24:51, blass uri wrote: >> >>> >>>On October 02, 1998 at 20:47:47, Robert Hyatt wrote: >>> >>> >>>>because crafty doesn't evaluate KPP vs K or KP vs KP as won, in that >>>>position. >>> >>>KP vs KP is not relevant >>>and I assume that KPP vs K is evaluated as at least 1 pawn advantage in every >>>position even if the black king can get into the square of the white pawn. >>> >>>I understand now that you use crafty15.21 and not crafty15.20 so if my >>>assumption is right then crafty15.20 will play Rxb2 with tablebases at depth=12 >>>and you did an important improvement with the new extension stuff. >>> >>>Uri >> >> >> >>first, you are making a gross assumption when you say "KP vs KP is not >>relevant". Remember that this is a deep search, full-width. So this particular >>database gets hit about 10,000 times during my 12 ply search. > >I simply can see that white can force KPP vs K and there is no reason to go for >KP vs KP(you can go to it by pusing your pawn forward instead of going with the >king to catch the pawns but it is not the shortest way to win) > >Uri unfortunately, this has nothing to do with the way computers do their full-width search. They consider *every* plan that can be realized within some fixed search horizon. In this position, all of the three KPPKP, KPPK and KPKP (and KPK) are possible...
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.