Author: Robert Hyatt
Date: 05:53:20 02/10/98
Go up one level in this thread
On February 10, 1998 at 06:58:59, Bernhard Bauer wrote: cute. The point, of course, is that KNN vs KB is drawn in 99.9999999999% of the cases. Just that this one is wrong. But the search cuts off at instant draws to save time. I do this knowing that once I start using KNN vs KP this won't ever happen because the mate (or draw) is found before the Evaluate() procedure ever sees the position... :) >Hallo everybody, > >As its said Bruce Moreland is using an EGTB for the 5-man ending KNNKP. >This is a simple and in some sense a brutal method of takling such a >problem >which uses in addition a lot of disk space. > >IMHO the method of Bob Hyatt is much smarter and needs no disk space at >all. >This method works of course not so well if crafty has the KNN side. >Crafty >will in some very rare cases not be able to win,if its possible, >however, if Crafty has the weaker part Crafty will draw even in hopeless >situations. This is achieved by a new extremly tricky method. > >But look yourself. >In the following position I (white) had good hope to win. >--> FEN: k1K/8/2N/2N/8/8/1p/8 w >The only difficulty IMNSHO was whether to move Na6 or Ne6. Since Ne6 >centralizes the knight more than the weak edge position after Na6 I >moved >Ne6 hoping to win this game. >However I completely overlooked the tricky answer of Crafty. >Crafty played the great move b1=B !!!!! >Now it became obvious that white could'nt win. Because the endgame >KNNKB is a general draw, isn't it? > >The output given by Crafty follows somewhat modified. >BTW you may need a grain of salt -)) > > >Kind regards >Bernhard > >----------------------------------------------------------------- > >Crafty v14.9 > >White(1): ponder off >pondering disabled. >White(1): book off >book file disabled. >White(1): noise 0 >noise level set to 0. >White(1): hash 3M >hash table memory = 3M bytes. >White(1): hashp 640K >pawn hash table memory = 640K bytes. >White(1): time cpu >using cpu time >White(1): st 60 >search time set to 60.00. >White(1): setboard k1K/8/2N/2N/8/8/1p/8 w > > +---+---+---+---+---+---+---+---+ > 8 | *K| | K | | | | | | > +---+---+---+---+---+---+---+---+ > 7 | | | | | | | | | > +---+---+---+---+---+---+---+---+ > 6 | | | N | | | | | | > +---+---+---+---+---+---+---+---+ > 5 | | | N | | | | | | > +---+---+---+---+---+---+---+---+ > 4 | | | | | | | | | > +---+---+---+---+---+---+---+---+ > 3 | | | | | | | | | > +---+---+---+---+---+---+---+---+ > 2 | | *P| | | | | | | > +---+---+---+---+---+---+---+---+ > 1 | | | | | | | | | > +---+---+---+---+---+---+---+---+ > a b c d e f g h > >White(1): d >White(1): search Ne6 >White(1): move >end-game phase > clearing hash tables > time surplus 0.00 time limit 1:00 (1:00) > depth time score variation (1) > 1 0.00 -4.01 1. Ne6 b1=Q > 1-> 0.00 -4.01 1. Ne6 b1=Q > 2 0.01 ++ 1. Ne6!! > 2 0.01 0.00 1. Ne6 b1=B > 2-> 0.02 0.00 1. Ne6 b1=B > 3 0.02 0.00 1. Ne6 b1=B > 3-> 0.03 0.00 1. Ne6 b1=B > 4 0.03 0.00 1. Ne6 b1=B > 4-> 0.03 0.00 1. Ne6 b1=B > 5 0.04 0.00 1. Ne6 b1=B > 5-> 0.04 0.00 1. Ne6 b1=B > 10-> 0.07 0.00 1. Ne6 b1=B > 30-> 0.19 0.00 1. Ne6 b1=B > 35-> 0.22 0.00 1. Ne6 b1=B > 40-> 0.24 0.00 1. Ne6 b1=B > 45-> 0.26 0.00 1. Ne6 b1=B > 50-> 0.28 0.00 1. Ne6 b1=B > 55-> 0.31 0.00 1. Ne6 b1=B > 60-> 0.33 0.00 1. Ne6 b1=B > time: 0.35 cpu:9% mat:6 n:588 nps:10000 > ext-> checks:134 recaps:0 pawns:240 1rep:3 > predicted:0 nodes:588 evals:8 > endgame tablebase-> probes done: 0 successful: 0 > hashing-> trans/ref:102% pawn:50% used:w0% b0% > >White(1): Ne6 > time used: 0.35 > >execution complete. >---------------------------------------------------------------------
This page took 0.01 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.