Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The Scalable Search Test

Author: Gian-Carlo Pascutto

Date: 08:11:42 06/18/00

Go up one level in this thread


On June 18, 2000 at 10:55:10, Steve Maughan wrote:

>Gian-Carlo,
>
>
>>Also, if anyone has other positions like this (preferrably even harder/worse),
>>please share them.
>
>You may like to have a look at this collection of positions
>
>http://www.maughan.clara.net/bbc.htm
>
>There are quite a few that take I would regard as 'even harder'

I'll be checking those out, but I suspect that you misunderstood what I meant.
I am not seeking positions that are hard to 'solve', but rather positions
that cause the search to misbehave. I.e. positions were programs can find the
good move, but then fail to get a reasonable score within reasonable time.

The problem I posted is very trivial...Rebel solved it within a second, most
amateurs can do it as quickly too. But getting a reasonable score back can
take way longer...

For example, Sjeng 7.3 finds Rxh5 in 5 seconds, fails high on it after 7
seconds with a +2.5 score (so far so good), but then takes more than half
an hour to fail high again, and it doesn't resolve the score within an
hour. This is the kind of behaviour I'm looking for.

--
GCP



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.