Author: Robert Hyatt
Date: 12:10:08 09/10/99
Go up one level in this thread
On September 10, 1999 at 14:38:42, Bruce Moreland wrote:
>On September 10, 1999 at 14:27:05, Robert Hyatt wrote:
>
>>On September 10, 1999 at 13:49:50, Bruce Moreland wrote:
>
>>><= Mate in N is simply not stored. <- A nasty case.
>>>>= -Mate in N is simply not stored. <- Likewise.
>
>>You used the term "a nasty case" on the last two. Did you run into problems
>>when you stored them? Or just the overwrite issue?
>
>I ran into problems when I screwed up and stored it wrong. I can't remember
>exactly what my bug was, but I probably messed up and converted it into a >=
>mate in 500 lower bound, which is of course almost exactly wrong. This can
>happen if you just look at the score and convert it to a bound, for instance "if
>(score >= mate_low) { score = mate_low; bound = lower; }"
>
>It's just a case that has to be paid attention to and disposed of properly.
>
>bruce
Aha... in my case, I have three distinct 'blocks' in the HashStore() code,
one for EXACT, one for UPPER and one for LOWER. I simply use Min() or Max()
to prevent the appropriate score from being either > MATE-300 or < -MATE+300
which was the fix needed to solve Steffen's position correctly.
this was hell to find in a 1 million line tree trace, I can tell you that. The
file was over 40 megabytes. Fortunately VI and a very fast notebook disk made
it doable.
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.