Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: MTD(f)

Author: Vasik Rajlich

Date: 09:53:51 06/19/05

Go up one level in this thread


On June 19, 2005 at 06:44:45, Dieter Buerssner wrote:

>On June 18, 2005 at 15:27:32, Vasik Rajlich wrote:
>
>Thanks for elaborating - I agree with everything.
>
>>>Lazy eval can be made "soft", futility pruning, too.
>>>
>>
>>As far as I can tell, a lazy eval will always cost you some softness.
>
>Of course. (Isn't "soft" really anti intuitive - I am no native English speaker,
>but I'd imagine the bound gets softer, not harder ...)
>
>>Regarding futility pruning in q-search, how will you return a good fail-soft
>>value here:
>
>This is not really futility pruning specific, I think
>

I think of standing pat in q-search as a form of "futility pruning", but I guess
that's breaking the official use of the word (going back to the Dark Thought
papers ..)

>>[D] r1bqk1nr/pppp1ppp/8/2b1p3/3nP3/2NP1N2/PPP1BPPP/R2QK2R w KQkq - 6 1
>>
>>White is down a piece and needs to get to +0.50 (that's our window, let's say).
>>We'd like q-search to return a fail-soft of -3.00. But: white tries 1. Nxd4, and
>>black fails high by standing pat. How do you avoid returning a fail soft of 0.00
>>or so?
>
>Exactly. Basic qsearch without any tricks will do this by allowing a standpat
>score. Perhaps the most important reason, we typically get back fail soft scores
>of alpha for all moves at the root besides the PV.
>

Yes, of the several reasons for bad failing soft, this might be the most common.
Maybe MTD (f) engines should disable all "pruning" in the q-search ..

Vas

>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.