Check extension

Discussion of chess software programming and technical issues.

Moderator: Ras

bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Check extension

Post by bob »

mjlef wrote:How about just testing what Toga uses? Extend all checks in PV nodes, and only checks that do not lose (based on on SEE) more than a pawn. A few thousand games should determine if this helps orr not, within a few percent. You can post the number of games won, lost or drawn for each version.
Let me clarify. (1) extend every check in a PV node, but what about non-PV nodes? I already extend every check everywhere. (2) using SEE is easy enough, but again, do I do this everywhere (only extend checks that are >= -1.0 according to SEE())?

So clarify and I can run the tests easily enough.
mjlef
Posts: 1494
Joined: Thu Mar 30, 2006 2:08 pm

Re: Check extension

Post by mjlef »

bob wrote:
mjlef wrote:How about just testing what Toga uses? Extend all checks in PV nodes, and only checks that do not lose (based on on SEE) more than a pawn. A few thousand games should determine if this helps orr not, within a few percent. You can post the number of games won, lost or drawn for each version.
Let me clarify. (1) extend every check in a PV node, but what about non-PV nodes? I already extend every check everywhere. (2) using SEE is easy enough, but again, do I do this everywhere (only extend checks that are >= -1.0 according to SEE())?

So clarify and I can run the tests easily enough.
Extend these:
in PV nodes, any check (whenever alpha!=(beta+1) ), even if appears to lose
in non-PV nodes, only extend moves with an SEE>= -value of one pawn

Mark
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Check extension

Post by bob »

mjlef wrote:
bob wrote:
mjlef wrote:How about just testing what Toga uses? Extend all checks in PV nodes, and only checks that do not lose (based on on SEE) more than a pawn. A few thousand games should determine if this helps orr not, within a few percent. You can post the number of games won, lost or drawn for each version.
Let me clarify. (1) extend every check in a PV node, but what about non-PV nodes? I already extend every check everywhere. (2) using SEE is easy enough, but again, do I do this everywhere (only extend checks that are >= -1.0 according to SEE())?

So clarify and I can run the tests easily enough.
Extend these:
in PV nodes, any check (whenever alpha!=(beta+1) ), even if appears to lose
in non-PV nodes, only extend moves with an SEE>= -value of one pawn

Mark
Will do. May be tomorrow or so as I currently have a slowish-game test running. But that's easy enough to do.