Stockfish question

Discussion of chess software programming and technical issues.

Moderator: Ras

jwes
Posts: 778
Joined: Sat Jul 01, 2006 7:11 am

Stockfish question

Post by jwes »

What would it take to have stockfish do tt cutoffs at pv nodes? I tried copying the code from search to searchpv, but it didn't seem to work right. Is there something else I would need to do? The reason I want to do this is for reverse analysis.
zamar
Posts: 613
Joined: Sun Jan 18, 2009 7:03 am

Re: Stockfish question

Post by zamar »

jwes wrote:Is there something else I would need to do? The reason I want to do this is for reverse analysis.
Yes, you need to remember that in PV nodes alpha != beta -1.
So ok_to_use_TT needs to be written differently. And it's not too obvious how this should be done, there are many alternatives.
Joona Kiiski
jwes
Posts: 778
Joined: Sat Jul 01, 2006 7:11 am

Re: Stockfish question

Post by jwes »

zamar wrote:
jwes wrote:Is there something else I would need to do? The reason I want to do this is for reverse analysis.
Yes, you need to remember that in PV nodes alpha != beta -1.
So ok_to_use_TT needs to be written differently. And it's not too obvious how this should be done, there are many alternatives.
It appears that ok_to_use_TT treats exact nodes as both upper and lower so it returns true if it is deep enough so just removing the test for PvNode seems to work.