Author: Bruce Moreland
Date: 14:40:40 12/15/99
Go up one level in this thread
On December 15, 1999 at 11:44:52, Torstein Hall wrote: >Take the following position and setup on your PC: > >4b3/1r4k1/p3pr2/1p1pPp2/PPpP1Pp1/2P3Pp/K6P/8 w - - 0 1 > >Every human would play 1.a5 without thinking, but your chessprogram plays >1.exf6+??? It may be wrong to call a program stupid but..... > >Torstein > >( So far I have only tried it on Fritz 6 and Hiarcs 7.32 ) This class of problem is very old, and is designed to exhibit worst-case behavior in chess programs by exercising an extremely uncommon special case. All it proves is that humans can generalize better than computers in this case. That indicates a failure for chess software in an AI sense, but doesn't necessarily indicate a failure in terms of chess strength. If you want to find a position where programs will outperform humans, just go to a book of mate in N chess problems like those composed by Sam Lloyd. Those problems are attractive to humans because they tax the human ability to generalize. But a computer would not find them interesting. bruce
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.