Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: MTD(f)

Author: Tord Romstad

Date: 08:35:54 07/28/04

Go up one level in this thread


On July 28, 2004 at 08:55:25, Uri Blass wrote:

>On July 28, 2004 at 08:23:17, Tord Romstad wrote:
>
>>On July 28, 2004 at 07:06:22, Uri Blass wrote:
>>
>>>How much rating do you earn from dividing the evaluation by 4.
>>
>>I no longer remember exactly, but I think it was quite a significant
>>improvement.  However, it is possible that the improvement is mostly explained
>>by a horrible futility pruning bug I introduced while doing the change.  For
>>some reason, this bug seems to make the engine much stronger, and all attempts
>>to fix it have had catastrophic results.
>>
>>The bug is the following:
>>
>>In the qsearch, I prune futile captures with code that looks roughly like
>>this:
>>
>>if(approximate_eval_after_move(move) < alpha-margin)
>>   prune move
>>
>>The approximate_eval_after_move() function takes the static eval of the
>>current position and adds the change in material and piece square table
>>scores when the move is made.  When I changed from pawn=64 to pawn=32,
>>I forgot to modify the approximate_eval_after_move() function.  This
>>function still divides by 2 rather than 4, and returns scores with
>>unit pawn=64.
>
>If I understand correctly your function give twice the expected value instead of
>giving the expected value.
>
>I guess that alpha is near 0 during the game.

This can vary a lot, but OK.

>Let assume that margin is near 1 pawn

In most positions, this is not too far from the truth.  I think it is
usually slightly less.

>If approximate_eval_after_move(move)<-1 pawn was planned.
>
>If 2*approximate_eval_after_move(move)<-1 pawn was done.
>In other words
>If approximate_eval_after_move(move)<-0.5 pawn was done.
>
>I suspect that it may be better to correct the bug and also divide margin by 2.
>Did you try it?

Yes, I have spent an enormous amount of time testing all sorts of settings.
I haven't found any other setting which is even close to what I currently
use.  It's tremendously frustrating.

Perhaps I should have a look at this problem one final time before I
release Gothmog 1.0.

Tord



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.