Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question on History heuristic

Author: Peter Kappler

Date: 23:58:29 10/05/99

Go up one level in this thread


On October 05, 1999 at 22:10:32, Robert Hyatt wrote:

>On October 05, 1999 at 21:04:40, Peter Kappler wrote:
>
>>On October 05, 1999 at 16:36:59, William Bryant wrote:
>>
>>>When updating the history for a move that has generated a cutoff,
>>>what value to people use.
>>>
>>>Do you use ply or some function of ply which means that the closer you are to
>>>the leaves, the greater the value being used for updating?
>>>ie: History[from][to] += ply;
>>>
>>>Do you use depth or some function of depth which means the closer you are to the
>>>root, the greater the value being used for updating?
>>>ie: History[from][to] += ply;
>>>
>>>Or are there other values being used.
>>>
>>>Thank you in advance.
>>>
>>>William
>>>wbryant@ix.netcom.com
>>
>>
>>Hi William,
>>
>>I think moves that cause a cutoff near the root should get a higher bonus, as
>>they have been searched more deeply.  In Grok, I increment the history score by
>>2^depth, which greatly favors moves that cause cutoffs near the root.  I think I
>>got the exponential scoring idea from an ICCA article, but I can't recall the
>>author.
>>
>>--Pete
>
>
>That is the right value from the JICCA article, but think about what happens to
>that number when depth = 20 and beyond, which I see in endgames all the time,
>even at blitz time controls.  your 32 bit counter overflows quickly...


This is a good point, though I'm sure it's a much bigger problem for Crafty than
Grok.  :-)

I might try your depth*depth idea.  Thanks for the tip.

--Peter




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.