Author: Robert Hyatt
Date: 07:29:12 08/29/01
Go up one level in this thread
On August 28, 2001 at 12:04:00, Roy Eassa wrote: >On August 28, 2001 at 09:26:10, Robert Hyatt wrote: > >>On August 27, 2001 at 16:50:36, robert flesher wrote: >> >>>Tiger 12.0 running on a mere pentium 850mhz with 256 mega of ram Sees that if >>>white exchanges both rooks followed by moving the king and playing b4 and a4 >>>wins. Anyone who still has tiger 12.0 should be able to confirm this. I have >>>gambit2.0 and tiger 14.0, but as ive posted before i prefer tiger12.0. >> >> >>What I am trying to ask is this: >> >>1. from the original position, does CT understand that trading all the >>rooks is bad? IE a 5-6 ply search should let it see the rooks coming off. >>And after that PV/score is displayed, does it show white is winning? >> >>2. After playing the moves that trade the rooks off, there are no pieces >>left. It is pretty easy to search to 20 plies here, and that is more than >>enough to see that white queenside majority turning into a white outside >>passed pawn. Does it realize the outside passer is winning? >> >>If it doesn't realize white is winning after searching just deeply enough to >>see the pieces coming off, then it is going to have trouble in these kinds of >>endings, which are pretty common. If it does realize white is winning in case >>2, that means it probably understands the outside passer, or at least it knows >>that one side having a passer is probably winning. >> >>The more interesting question is "does it understand that the white majority is >>_also_ winning if there are no pieces left, because the majority will turn into >>a distant passer later in the game"??? > > >My brief experiments show that the answer is "No" to 1 and "Yes" to 2 for both >current Tigers and for Fritz 6 & Junior 6. OK... so they don't understand pawn majorities but they do understand either passed pawns or outside passed pawns. That matches with what I have seen on ICC...
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.