Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test your programs for a #18

Author: Dieter Buerssner

Date: 12:32:41 05/24/05

Go up one level in this thread


On May 24, 2005 at 15:10:08, Marc D wrote:

>How long would it take yace without tbs?

I fear, it will not be able to show a mate in any reasonable time (say in a
week). However, it will show you the way to win without TBs in under one second.
Ke8 is found in 1/2 s with score 3.8. After 5 minutes, score is +13.

With the backwards analysis features inside Yace, you might be able to prove a
mate in reasonable time. But it will need some user interaction (although it
might be very obvious, what to do; I hope to do it automatically some time in
the future). To prove a shortest mate in such a position - looks very difficult
to me. To really prove a mate, an engine would also need to use more
sophisticated hashing schemes, than the typical chess engines use. Typical chess
engines will store a hash key somewhere inside the table. When the hash key of
the current position is equal to the stored hash key, they will assume, that the
positions are identical. This is almost always the case (really almost always -
for example you would not expect one failure during a year or longer, when
storing a 64 bit hash key). But you still would not have a real proof. I assume
Chest and company store the whole position inside the HT, to be really sure.

Regards,
Dieter



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.