Henk wrote:bob wrote:cdani wrote:
I do razoring before null move. Did you try it?
How would that work? Null move is done BEFORE anything is searched. Razoring applies to moves. I am not sure I understand your idea...
Maybe he means reverse razoring. Razoring for the opponent. Razoring after opponent has moved. Possible if you get last move from the stack. Don't know what's the use of it. But I also encountered "reverse futility pruning" in historic posts somewhere. I don't know.
Also stockfish does razoring before null move:
...
// Step 6. Razoring (skipped when in check)
// Step 7. Futility pruning: child node (skipped when in check)
// Step 8. Null move search with verification search (is omitted in PV nodes)
...
May be I'm not using the correct name in some concepts.
I've done a lot of testing myself, but I do not do these 1 second games since I have enough cluster hardware to use 1m + 1s as a minimum test time control anyway. I see no gain whatsoever for treating PV nodes differently, even null-move, although not doing null-move does make sense because it is an illegal move and has to fail low or fail high to avoid an illegal move in the PV...