Author: Robert Hyatt
Date: 13:54:56 05/14/99
Go up one level in this thread
On May 14, 1999 at 16:28:43, Francis Monkman wrote: >Here's another nice one, from my first game with Hiarcs 2.1 in late 94. > >2r1r1k1/3q1p2/p1bp2pp/1pn1pNb1/4P3/2P2QNP/PPB2PP1/R3R1K1 w - - 0 24 > >White's 24th is a killer, 24. Rad1 (Black has just played 23. ...g6). > >Crafty 16.6 and CSTal II both fail to get this after 30 minutes each, >preferring respectively 24. h4 and 24. Qg4, neither of which wins. > >LGG 2.0 comes up trumps again: > >depth=12 +0.23 a1d1 g6f5 g3f5 e8e6 h3h4 g5d8 f5d6 e6d6 >d1d6 d7d6 f3g4 g8h7 g4c8 d8h4 c8f5 h7g8 e1d1 d6e6 f5h5 >Nodes: 70970656 NPS: 19444015 >Time: 00:00:03.65 > >(In fact, Black played 27. ...Rb8 in the game, but if 27. ...Rxd6 >as in LGG's analysis, White should then play 28. Qg3+ and 29. Qxe5) > >I only tried these three programs so far (naturally, attempts in 94/5 >by Geniuses 2-4 and Fritz 3 failed) -- any offers, ...Jeremiah? > >Francis Just looking at what you posted, this appears to be a case of a program coming up with a random move for the wrong reason. IE the eval of +.23 is not exactly 'winning'. And while the moves from CSTal/Crafty might not win either, what makes Rad1 a winner? And if it does, do you _really_ want a program that will play such a move without having a clue why it is playing it? My opinion is that there are probably several reasonable moves here, none are outright winning or losing...
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.