Breunor: According to Stockfish valuation, this game goes back and forth quite a lot but black is ahead well more than 4 in the early 50 ish moves. Move 52 unfortunately for black gives away the big lead as hcgflynn said. Here is the computer analysis:
1) -4.60 (22 ply) 52...e2 53.Rf2+ Kh3 54.Rxe2 Qf3 55.Qd2 Rd3 56.Qc2 h5 57.Rh2+ Kg3 58.Rxh5 Qxh5 59.Qxd3+ Qf3 60.Qd6+ Kh3 61.Qd2 g4 62.a3 g3 63.Qh6+ Kg2 64.Qd2+ Qf2 65.Kc3 Qxd2+ 66.Kxd2 Kf3 67.Ke1
But after 52 Qg2 ch the game should be drawn:
1) =0.00 (30 ply) 53.Kxc3 e2 54.Qd6+ Qg3+ 55.Qxg3+ Kxg3 56.Rh1 f5 57.Kd2 f4 58.Kxe2 Kg2 59.Rf1 h5 60.Rf2+ Kg1 61.Rf3 Kg2 62.Rf1 h4 63.Rf2+ Kg3 64.Rf3+ Kg2 65.Rf2+
Computer thinks black can still draw on move 57:
1) =0.00 (32 ply) 57...f5 58.Kxe2 Kg2 59.Rf1 f4 60.b4 h4 61.Rf2+ Kg3 62.Rf3+ Kg2 63.Rf2+
Interesting it prefers f5 to h4; apparently the key is that the f pawn helps keep the white king out and/or helps against horizontal checks from the rook. But Stockfish appears to be weak at endgames, you have to run it for a very long time to get something meaningful often.
Computer also doesn't like 58 Kg2;
1) +1.83 (26 ply) 58...h3 59.Kf1 g4 60.Kg1 f5 61.Re7 b5 62.Rxa7 b4 63.Rb7 f4 64.Rg7 Kh4 65.Rf7 Kg3 66.Kh1 h2 67.Rg7 f3 68.Rh7 Kf2 69.Rxh2+ Ke1 70.Rc2 g3 71.Rc1+ Kd2 72.Ra1 Ke2 73.Kg1 Kd2 74.Kf1 Ke3 75.Re1+ Kf4
After Kg2 white lead is large:
1) +3.38 (31 ply) 59...h3 60.Rf2+ Kg3 61.Rxf6 g4 62.Kf1 Kh2 63.Rg6 g3 64.Rg5 g2+ 65.Kf2 Kh1 66.Rg7 Kh2 67.Rg3 b6 68.Rg6 Kh1 69.Rg5 Kh2 70.Rg3 a6 71.Rg5 b5 72.b4 Kh1 73.Rg4 Kh2 74.Rg3 a5 75.bxa5 b4 76.a6 b3 77.a7 b2
Kibitzing
icon.
