Qsearch pruning: follow-up
authorStéphane Nicolet <cassio@free.fr>
Fri, 6 Nov 2020 18:20:27 +0000 (19:20 +0100)
committerStéphane Nicolet <cassio@free.fr>
Sun, 8 Nov 2020 08:15:34 +0000 (09:15 +0100)
commit392b529c3f52103ad47ad096b86103c17758cb4f
tree306dd371be7b2b53faf8f60f5ea2bd433296bde0
parent32edb1d009e09a9442cb7393920e072ffd08005d
Qsearch pruning: follow-up

This is a follow-up of the recent qsearch pruning patch in
https://github.com/official-stockfish/Stockfish/commit/a260c9a8a24a2630a900efc3821000c3481b0c5d

We now use the same guard condition (testing that we already have a defense with
a score better  score than a TB loss) for all pruning heuristics in qsearch().
This allows some pruning when in check, but  in a controlled way to ensure that
no wrong mate scores appear.

Tested with Elo-gaining bounds:

STC:
LLR: 2.97 (-2.94,2.94) {-0.25,1.25}
Total: 22632 W: 2433 L: 2264 D: 17935
Ptnml(0-2): 98, 1744, 7487, 1865, 122
https://tests.stockfishchess.org/tests/view/5fa59405936c54e11ec99515

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 105432 W: 4965 L: 4648 D: 95819
Ptnml(0-2): 85, 4110, 44011, 4423, 87
https://tests.stockfishchess.org/tests/view/5fa5b609936c54e11ec9952a

closes https://github.com/official-stockfish/Stockfish/pull/3221

Bench: 3578092
src/search.cpp