Author: Robert Hyatt
Date: 12:19:57 10/28/98
Go up one level in this thread
On October 28, 1998 at 14:44:52, Roberto Waldteufel wrote: > >Hi Bruce, > >Yes, I am familiar with Reti's famous study. In fact, this reminds me of one of >my own studies from way back - I believe it was published at the time in the >Problemist. White is to play and win (that's not a typo - White really has a >forced win here!) > >The position: White king: g3 > Black king: h5 > White pawns: g2,h2,f3,f4 > Black pawns: b5,b6,f6,g6,h6 > >I would be interested to know what programs make of this position (a) with a >full search, and especially (b) with a static evaluation only. I have shown this >position to many strong human players (including IMs) and they have always been >amazed that White is not lost, let alone winning. In fact, IM Paul Littlewood >once let his curry get cold while he analyzed this position instead of eating! > >This study, as well as Reti's, are excellent examples of the dangers of >accepting the obvious superficial assessment of passed pawns. > >Best wishes, >Roberto My static eval has no hope here... But a quick search discovers that Kh3 is best... with a score of near zero until depth=17, 10 seconds (ALR 4 proc machine) where Kh3 fails high to +.5, then at 18 it fails high to +1, and will presumably fail high again... I stopped it after 19 plies, 1 minute... static pawn race eval says 0.00 for that part of the evaluation because the white king is on move and can prevent a pawn from queening at b1... Bob
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.