Amoeba 2.7

[abulmo2]

In version 2.7, I modified the see pruning in quiescence search and forgot to disable it when in check. This caused some positions to be misevaluated, and even some losses in games: for example against Nemeton during the december blitz tourney: http://talkchess.com/forum/viewtopic.ph ... 81&t=66034
Curiously fixing this bug does not bring any Elo, so I am not in hurry to release version 2.8.

In my development version, the correct move Qxa8 is found much faster:

Code: Select all

info depth 1 seldepth 6 score cp 74 nps 835621 time 0 nodes 213 hashfull 0 pv a3b4 
info depth 2 seldepth 7 score cp 74 nps 959364 time 0 nodes 314 hashfull 0 pv a3b4 c6c7 
info depth 3 seldepth 7 score cp 98 nps 1027085 time 0 nodes 474 hashfull 0 pv a3b4 c6c7 b2c2 
info depth 4 seldepth 12 score cp 81 nps 958734 time 1 nodes 1236 hashfull 0 pv a3b4 c6c7 b4b7 c7b7 
info depth 5 seldepth 18 score cp 81 nps 1023151 time 1 nodes 1865 hashfull 0 pv a3b4 c6c7 b4b7 c7b7 b2b7 
info depth 6 seldepth 8 score cp 53 nps 1091190 time 6 nodes 7261 hashfull 0 pv a3b4 c6c7 d6e8 f8e8 d3b5 e8f8 b5d3 
info depth 7 seldepth 10 score cp 6 nps 1250128 time 7 nodes 9760 hashfull 0 pv a3b4 c6c7 d6e8 f8e8 d3b5 e8f8 b5d3 f8g8 
info depth 8 seldepth 10 score cp 322 nps 1325372 time 10 nodes 14162 hashfull 0 pv a3a8 c6a8 b2b8 a8a4 b8e8 a4e8 d6e8 f8e8 f3e3 
info depth 9 seldepth 12 score cp 315 nps 1350923 time 11 nodes 15191 hashfull 0 pv a3a8 c6a8 b2b8 a8a4 b8e8 a4e8 d6e8 f8e8 d3b5 e8f8 b5d3