Null Move Beta=0

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

D Sceviour
Posts: 570
Joined: Mon Jul 20, 2015 5:06 pm

Re: Null Move Beta=0

Post by D Sceviour »

Has anyone tried the Null Move beta=0? Any success, failure, indifference or questions?
Engin
Posts: 918
Joined: Mon Jan 05, 2009 7:40 pm
Location: Germany
Full name: Engin Üstün

Re: Null Move Beta=0

Post by Engin »

would that be not easier to let null move only be allow at beta != 0 ?
D Sceviour
Posts: 570
Joined: Mon Jul 20, 2015 5:06 pm

Re: Null Move Beta=0

Post by D Sceviour »

Engin wrote:would that be not easier to let null move only be allow at beta != 0 ?
The last time I tried beta!=0, it resulted in a slowdown of the search. However, there have been a great many changes since then. This time beta!=0 showed good results in the general search! Therefore, beta!=0 might be the simplest answer for all positions. Next experiment with beta!=0 is king pawn endgames. The idea may only be effective if the engine returns NULL for repeated positions.

My engine can only solve this position with beta!=0 condition.

[d]1k3b1q/pP1KpBp1/P3P1Pp/7P/8/8/8/8 w - - 16 1