Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What is the average nodes per second for minimax?

Author: leonid

Date: 03:07:06 06/25/00

Go up one level in this thread


On June 25, 2000 at 02:22:16, blass uri wrote:

>On June 24, 2000 at 21:40:19, leonid wrote:
>
>>On June 24, 2000 at 15:11:46, blass uri wrote:
>>
>>>On June 24, 2000 at 15:02:09, blass uri wrote:
>>>
>>><snipped>
>>>>>Alpha-beta can give you very often as much as 15 moves of equal value in the
>>>>>first half of usual game. I speak all the time about what I call "positional
>>>>>logic".
>>>>
>>>>Alpha beta never gives you more than one move with a value and other moves that
>>>>you know that they are not better and you have no idea if they are equal or
>>>>worse.
>>>
>>>To be more exact alpha beta is a function that givees you a number (alpha if the
>>>evaluation of the position after search is worse than alpha,beta if the
>>>evaluation after search is bigger than beta and a score between alpha and beta
>>>in other cases.
>>>
>>>
>>>If the evaluation of Na3 after search is 0.00 you are not intereted in the exact
>>>value of e4 if it is not better than 0.00 and your calculation will give 0.00
>>>even if e4 is losing material  so you will have no way to know if e4 is losing
>>>or leads to an equal position.
>>>
>>>Uri
>>
>>Ura, probably all confusion is about the ply number. I speak about upper ply.
>>Ply where you are forced to see all the moves. It is the place where multiple
>>moves you can have to chose later.
>>
>>Leonid.
>
>I understand that you calculate only material evaluation for all the legal moves
>and if the result is equal you calculate decide which move to search first by
>positional score so you can avoid 1.Na3 and play 1.e4 because the positional
>evaluation(not material evaluation) of 1.e4 is better than the positional
>evaluation of 1.Na3.

So, now I see that it was more confusion about words that some real difference
in programming. I even started to think that maybe I do my program differently.

Your naming of the part of the program, that is usual, is also helpful. Will try
to speak in usual terms when speaking about programming.

In my program I give name as "positional logic" to the part of program that look
for positional move. Partially it is found by material echange section and
partly by section of "general orientation". I consider both sections as part of
general positional logic. Only first part "material echange" is just too weak to
deliver complet positional move. Second part come to remediate innate weakness
in calulating power of the "material echange". Otherwise, material echange could
stay as monolitic piece of unique positional logic.


>This is better than only material evaluation but still worse than
>doing the things in one step because if a move is losing something positional
>and you do not see it in the first ply you will never see it if you do not see
>that it loses material.

Probably what you say is that not only moves (in upper ply) that are equal in
value that must be regarded in second section. Moves that are close to the final
value, in upper ply, must be also left for verification for better final move.

>I do not see the reason to do 2 steps when you can calculate it in one step.

I still can't see how you can have everything calulating in one step without
making sometime useless calcualtions. You will do those calculation all the time
when second step is absolete. It happened when only one clear winning move is in
sight.

Leonid.

>Uri



This page took 0.01 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.