Author: Don Dailey
Date: 08:59:20 07/12/98
Go up one level in this thread
>I once culled games for mates. If you do self testing >and let the games run to mate positions you are guaranteed >to be able to generate mate in N to any depth you are >capable of searching. Woops, I wish to correct this statement before someone else does. Given enough games and randomness, this may be a nearly true statement (as long as N is not a ridiculous number.) But in practice it might be very hard to find really deep mates this way because most/all of them will be based on an error which will bound the depth of the mate. If the game is a draw and then an error is made, there may suddenly be a mate in 5, but not a deeper mate (with best play.) So the depth of the longest possible checkmate is bounded by the depth of the game generated, and the last game theoretic error by the defender assuming the game ended decisively. In practice, any time you have a lost position there will be a forced mate and it could be incredibly deep if the game is only "theoretically lost" but with "lots of play" left, at least as far as humans are concerned. - Don
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.