Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Alpha-beta question

Author: Sune Fischer

Date: 01:49:17 07/13/02

Go up one level in this thread


On July 13, 2002 at 04:37:31, Dann Corbit wrote:

>On July 13, 2002 at 04:02:49, Uri Blass wrote:
>
>>On July 13, 2002 at 03:39:06, Dann Corbit wrote:
>>
>>>On July 13, 2002 at 03:36:54, Uri Blass wrote:
>>>
>>>>On July 13, 2002 at 01:38:49, Russell Reagan wrote:
>>>>
>>>>>I have a question about alpha-beta. My understanding is that with perfect move
>>>>>ordering, you can get your branching factor down to the square root of the
>>>>>min-max branching factor. I did a few walk throughs of small trees using
>>>>>alpha-beta because I wanted to "see" what the tree looked like (where cutoffs
>>>>>occured), and I don't find this to be the case. I did a simple ternary tree walk
>>>>>through, and these are my numbers:
>>>>>
>>>>>Depth=1 : 4 nodes
>>>>>Depth=2 : 9 nodes
>>>>>Depth=3 : 49 nodes
>>>>>Depth=4 : 132 nodes
>>>>>
>>>>>Using a min-max search the branching factor should be 3. The branching factor
>>>>>for each of these depths was 2.22, 2.45, and 2.69 (which looks to be approaching
>>>>>3 with added depth). The square root of 3 is 1.73, so am I misinterpreting what
>>>>>I heard about the branching factor and alpha-beta?
>>>>>
>>>>>In other words...If your min-max branching factor is N, then does using
>>>>>alpha-beta with perfect move ordering give you the square root of N as the
>>>>>branching factor, or is that the lowest possible limit of the branching factor?
>>>>>
>>>>>If I understand this all correctly, that means that in chess a branching factor
>>>>>below about 6 is not possible without using forward pruning (using alpha-beta)?
>>>>
>>>>You need to assume also that hash tables are not used to prune the tree.
>>>>
>>>>I believe that pruning is very important and I exepct 2M nodes per second with
>>>>recursive null move pruning(R=3) to beat 200M nodes per second with no pruning
>>>>if you use 120/40 time control.
>>>>
>>>>I believe that programs practically can get good branching factor mainly thanks
>>>>to pruning and not thanks to hash tables(at least I do because I still do not
>>>>use hash tables to prune the tree).
>>>
>>>But you order your moves from the hash table!
>>>;-)
>>>
>>>Which came first, the chicken or the egg?
>>
>>This is only better order of moves and we assume perfect order of moves
>>for calculating the branching factor.
>
>I was talking of trying to actually achieve it.  Without a hash table, you will
>have to have some sort of genius evaluation to approach it.
>
>>It is clear that hash tables with no pruning means beanching factor of at least
>>5 in the opening position and 6 in the middle game.
>
>I wonder what the best branching factor for a program is.
>
>I think Pepito and Chess Tiger are both impressive in that regard.

Branch factors depend on the search depth, and that is not well defined.

Bob and Christophe had a recent discussion about that, and because some use
forward pruning instead of extensions, it will seem as though they search deeper
and thus has a lower BF.

-S.



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.