Author: Robert Hyatt
Date: 09:42:16 04/01/00
Go up one level in this thread
On April 01, 2000 at 10:35:50, Harald Faber wrote: >On April 01, 2000 at 08:18:51, Frederic Friedel wrote: > >>Amazingly it never occurred to anyone -- Thompson, Stiller, Nalimov -- to run >>through the ending K+N+N vs K. Everyone assumed it was simply a draw in every >>case. Well, fire up your Nalimov tablebase generator and prepare for a shock! In >>over 10% of all legal positions the attacking side can forced a mate in 31 to 48 >>moves. It is difficult to do anything systematically with the tablebase after it >>has been generated (except play against them), but as far as I can tell the >>attacking side can force a win if the defending king is on one of the squares >>a1, b1 or a2 and the attacking king and at least one knight is within the square >>a1-d1-d4-a4. Naturally this can be reflected in all four corners of the board. >>Some of the mates are wonderfully intricate and involve forcing the king to >>move to a4 (or d1) and back again to a1. John Nunn is studying the ending and >>will provide a better statistical evaluation soon. > >I even found a position where KNK is won! There is exactly one position where a lone king can mate another king. but he usually gets arrested if he is caught doing that in public. :)
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.