Author: Bruce Moreland
Date: 21:26:19 12/21/00
Go up one level in this thread
On December 21, 2000 at 17:06:55, Jeroen van Dorp wrote: >On December 21, 2000 at 16:03:09, Uri Blass wrote: > >>The question is if it can win with Rxh2. > >That's a valid question. >Remeber however it wasn't the question "will black win with this move" but "will >black *find* the move", and it did; >If Rxh2 loses badly it can also be an indication of extremely bad calculating. > >If. If. > >>If it find Rxh2 but is losing with this move against other programs then I do >>not consider it as a solution. > >That remark is bungling in the air a bit. > >What would be the reason for that? That statement is more speculative than the >one made in the original post. What is relevance of the remark? That goes for >anything: if the shoepolish machine starts polishing your shoes, but stops in >between, your shoes won't be polished and I won't consider it a shoe polish >machine. >But *why* should the shoe polish machine doesn't do it's work properly? Is there >any indication the shoe polish machine will stop, without seeing it working -and >failing- a lot of cycles? > >Based on the way a chess engine works -including CS Tal II- the chances any >chess program finding a winning move and winning with it are bigger than the >chances it finds a (*any*) winning move and still loses. >A winning move is in most cases the beginning of a tactical sequence and no >doubt an engine will calculate that right with greater chance than failing along >the way; > >Anyhow, regardless of your own ideas about it, my message simply stated CS Tal >found the move asked for in 44 seconds, what no other engine did so far. > >Jeroen ;-} If a test suite is so difficult that no program can solve any of the positions, a program that picks a move at random will have the best score, since it will solve a few percent of them by random chance. I don't insult CST by saying that it is doing this, but I don't think it is understanding the position, either. If you take the line to the end, the last move is a tactical blunder. That's fine, it's to be expected. But if you look at what's going on in the positions near the end of that line, white has significant problems but it's not like black's pieces have fed to the k-side very well. Black is attacking with two units only, for the moment. White's defenders are not very effective but they are numerous. A gutsy decision by the program, but it doesn't have anything figured out. 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.