Author: Dann Corbit
Date: 17:25:40 12/16/99
Go up one level in this thread
On December 15, 1999 at 18:05:24, Enrique Irazoqui wrote: >On December 15, 1999 at 17:17:46, Abel Dávalos wrote: > >>Your should try with Shredder 4.0, it play a5 after 30 seconds > >On a modest PII-300 with 64MB hashtables, Shredder 4 picks a5 at ply 20, after >62 seconds. But more interesting than picking a5 or not is the evaluation. The >--15.17 of Shredder and similar by Crafty, Rebel and Zarkov don't make Torstein >wrong, no? It should be 0.00. True, the computer did not know the real reason to make the move. But after a couple correct guesses in a row things like that are often discovered. Right now, it is too far over the horizon, but it could pop into view at any time. As it stands, they chose the right move for the wrong reason. For that exact reason, I think a EPD test position should also have a target CE to know if the "answer" was found or it just stumbled onto the right move by accident. A real downside of scoring a move as "right" if the correct move is chosen is that some programs will stop instantly if they have chosen the correct move, even though the CE would tell them that they are about to die. Someone once said "It's better to be lucky than good." Personally, I don't believe in luck. But I do believe in probability. ;-)
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.