A test position from lczero-berserk

[Uri Blass]

[d]8/p2k2p1/r1p2p2/P1P5/1P6/2BK4/8/8 b - - 1 44 am g6

I wonder how many engines play this blunder and if you can reproduce it with some engine or with the game.

https://tcec-chess.com/#div=gm&game=18&season=26

Berserk planned Ke6 when it played the previous move with evaluation of 0.00 and I guess it decided that g6 also leads to a draw so there is no difference and it suggest that it is a mistake to evaluate draws as 0.00 and there are better draws and worse draws.

[Jouni]

45. -g6 seems to be only losing move. Even simple -Rxa5 is draw. Very difficult zugzwang:

Analysis by The Huntsman 1:

45.Bxf6 Ke6 46.Bg7 Kf5 47.Kd4 g5 48.Bh6 g4 49.Bc1 Ke6 50.Bf4 Kf5 51.Bg3 Ke6 52.Ke4 Kf6 53.Bf2 Kg6 54.Ke5 Kf7 55.Kd6 Rxa5 56.bxa5 Kg6 57.Bg3 a6 58.Kxc6 Kf5 59.Kb6 Ke4 60.Bh4 Ke5 61.Kxa6 Kd5 62.Kb6 Kc4 63.a6
+- (#47) Depth: 37/39

[Jouni]

Correction: there are more losing moves. But 44. -g6 don't make any sense. Beginner move .