Is Alphazero's Immortal Zugzwang flawed? Does 49...Rf8 allow black to draw?

[rj329]

[pgn][Event "AlphaZero - Stockfish"]
[Site "London ENG"]
[Date "2017.12.04"]
[EventDate "?"]
[Round "?"]
[Result "1-0"]
[White "AlphaZero (Computer)"]
[Black "Stockfish (Computer)"]
[ECO "E15"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "119"]

1. Nf3 Nf6 2. c4 b6 3. d4 e6 4. g3 Ba6 5. Qc2 c5 6. d5 exd5
7. cxd5 Bb7 8. Bg2 Nxd5 9. O-O Nc6 10. Rd1 Be7 11. Qf5 Nf6
12. e4 g6 13. Qf4 O-O 14. e5 Nh5 15. Qg4 Re8 16. Nc3 Qb8
17. Nd5 Bf8 18. Bf4 Qc8 19. h3 Ne7 20. Ne3 Bc6 21. Rd6 Ng7
22. Rf6 Qb7 23. Bh6 Nd5 24. Nxd5 Bxd5 25. Rd1 Ne6 26. Bxf8
Rxf8 27. Qh4 Bc6 28. Qh6 Rae8 29. Rd6 Bxf3 30. Bxf3 Qa6 31. h4
Qa5 32. Rd1 c4 33. Rd5 Qe1+ 34. Kg2 c3 35. bxc3 Qxc3 36. h5
Re7 37. Bd1 Qe1 38. Bb3 Rd8 39. Rf3 Qe4 40. Qd2 Qg4 41. Bd1
Qe4 42. h6 Nc7 43. Rd6 Ne6 44. Bb3 Qxe5 45. Rd5 Qh8 46. Qb4
Nc5 47. Rxc5 bxc5 48. Qh4 Rde8 49. Rf6 Rf8 50. Qf4 a5 51. g4
d5 52. Bxd5 Rd7 53. Bc4 a4 54. g5 a3 55. Qf3 Rc7 56. Qxa3 Qxf6
57. gxf6 Rfc8 58. Qd3 Rf8 59. Qd6 Rfc8 60. a4 1-0[/pgn]

Would 49...Kf8 have allowed black to hold on to a draw? Here is Stockfish's analysis of the position following 49...Kf8:

Code: Select all

68/96	1:36:45	 40,404,584k	6,959k	+0.50	Bb3-d5 Re8-c8 Qh4-f4 Qh8-g8 Bd5-b7 Rc8-e8 Qf4-d6 c5-c4 a2-a4 c4-c3 Bb7-d5 Re8-d8 Qd6-c7 Kf8-e8 Qc7xc3 Qg8-f8 Rf6-f4 Re7-e6 Rf4-f3 Qf8-e7 a4-a5 Qe7-d6 Bd5-b3 a7-a6 g3-g4 Qd6-e7 Bb3-d5 Qe7-d6 Qc3-g7 Qd6-f8 Qg7xh7 Ke8-e7 g4-g5 Rd8-c8 Kg2-g3 Rc8-d8 Rf3-f4 Rd8-c8 Kg3-h3 Rc8-b8 Kh3-g2 Rb8-d8 f2-f3 Rd8-b8 Kg2-h3 Rb8-c8 Kh3-g4 Rc8-b8 Bd5-a2 Rb8-c8 Ba2-b3 Rc8-b8 Bb3-d5 Rb8-c8 Rf4-d4 Re6-d6 Rd4-e4+ Rd6-e6 Bd5xe6 d7xe6 Kg4-h4 Rc8-b8 f3-f4 Qf8-c8 f4-f5 g6xf5 Qh7xf5 Qc8-c1 Qf5-f6+ Ke7-e8 Qf6-h8+ Ke8-e7 Qh8xb8 Qc1-h1+ Kh4-g3 Qh1xe4 Qb8-c7+ Ke7-e8 Qc7-c8+ Ke8-e7 Qc8-c5+ Ke7-e8 Kg3-f2 e6-e5 g5-g6 f7xg6

[peter]

[pgn][Event "AlphaZero - Stockfish"]
[Site "London ENG"]
[Date "2017.12.04"]
[EventDate "?"]
[Round "?"]
[Result "1-0"]
[White "AlphaZero (Computer)"]
[Black "Stockfish (Computer)"]
[ECO "E15"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "119"]

1. Nf3 Nf6 2. c4 b6 3. d4 e6 4. g3 Ba6 5. Qc2 c5 6. d5 exd5
7. cxd5 Bb7 8. Bg2 Nxd5 9. O-O Nc6 10. Rd1 Be7 11. Qf5 Nf6
12. e4 g6 13. Qf4 O-O 14. e5 Nh5 15. Qg4 Re8 16. Nc3 Qb8
17. Nd5 Bf8 18. Bf4 Qc8 19. h3 Ne7 20. Ne3 Bc6 21. Rd6 Ng7
22. Rf6 Qb7 23. Bh6 Nd5 24. Nxd5 Bxd5 25. Rd1 Ne6 26. Bxf8
Rxf8 27. Qh4 Bc6 28. Qh6 Rae8 29. Rd6 Bxf3 30. Bxf3 Qa6 31. h4
Qa5 32. Rd1 c4 33. Rd5 Qe1+ 34. Kg2 c3 35. bxc3 Qxc3 36. h5
Re7 37. Bd1 Qe1 38. Bb3 Rd8 39. Rf3 Qe4 40. Qd2 Qg4 41. Bd1
Qe4 42. h6 Nc7 43. Rd6 Ne6 44. Bb3 Qxe5 45. Rd5 Qh8 46. Qb4
Nc5 47. Rxc5 bxc5 48. Qh4 Rde8 49. Rf6 Rf8 50. Qf4 a5 51. g4
d5 52. Bxd5 Rd7 53. Bc4 a4 54. g5 a3 55. Qf3 Rc7 56. Qxa3 Qxf6
57. gxf6 Rfc8 58. Qd3 Rf8 59. Qd6 Rfc8 60. a4 1-0[/pgn]

Would 49...Kf8 have allowed black to hold on to a draw? Here is Stockfish's analysis of the position following 49...Kf8:

Code: Select all

68/96	1:36:45	 40,404,584k	6,959k	+0.50	Bb3-d5 Re8-c8 Qh4-f4 Qh8-g8 Bd5-b7 Rc8-e8 Qf4-d6 c5-c4 a2-a4 c4-c3 Bb7-d5 Re8-d8 Qd6-c7 Kf8-e8 Qc7xc3 Qg8-f8 Rf6-f4 Re7-e6 Rf4-f3 Qf8-e7 a4-a5 Qe7-d6 Bd5-b3 a7-a6 g3-g4 Qd6-e7 Bb3-d5 Qe7-d6 Qc3-g7 Qd6-f8 Qg7xh7 Ke8-e7 g4-g5 Rd8-c8 Kg2-g3 Rc8-d8 Rf3-f4 Rd8-c8 Kg3-h3 Rc8-b8 Kh3-g2 Rb8-d8 f2-f3 Rd8-b8 Kg2-h3 Rb8-c8 Kh3-g4 Rc8-b8 Bd5-a2 Rb8-c8 Ba2-b3 Rc8-b8 Bb3-d5 Rb8-c8 Rf4-d4 Re6-d6 Rd4-e4+ Rd6-e6 Bd5xe6 d7xe6 Kg4-h4 Rc8-b8 f3-f4 Qf8-c8 f4-f5 g6xf5 Qh7xf5 Qc8-c1 Qf5-f6+ Ke7-e8 Qf6-h8+ Ke8-e7 Qh8xb8 Qc1-h1+ Kh4-g3 Qh1xe4 Qb8-c7+ Ke7-e8 Qc7-c8+ Ke8-e7 Qc8-c5+ Ke7-e8 Kg3-f2 e6-e5 g5-g6 f7xg6

To the question of the thread, you meant 49...Kf8 there too, not 49...Rf8? (in text it's correct, in title of thread probably a typo).
As for ...Kf8 proabably yes, to be drawn, yet at least 49...Rf8? was a crucial blunder of SF's way back then, that's already well known since quite a while, it was discussed here in extenso at least once afair, just can't use the search at the moment because of server's "high load".

Here's SF at the postion before 49th move of Black's after a little Forward- Backward of some lines after 49...Kf8 (only move saving the chances) in MultiPV=12 (to keep ...Rf8? in output) in depth 31 with 32G hash and 6men Syzygys:


[fen]4r1kq/p2prp1p/5RpP/2p5/7Q/1B4P1/P4PK1/8 b - - 0 1[/fen]


Analysis by Stockfish 290522:

Code: Select all

1. +/-  (1.07): 49...Kf8 50.Qf4 Qg8 51.Qc7 c4 52.Qxc4 Rd8 53.Qc3 Ke8 54.g4 Re6 55.g5 Qf8 56.Rf3 Qe7 57.Qh8+ Qf8 58.Qxh7 Ke7 59.Bxe6 dxe6 60.Ra3 a5 61.Rxa5 Ra8 62.Rc5 Kd6 63.Rc3 Rxa2 64.Qg7 Qxg7 65.hxg7 Ra8 66.Rd3+ Ke7 67.Ra3 Rxa3 68.g8Q Ra5 69.f4 Rd5 70.Qa8 Rf5 71.Qa3+ Kd7 72.Qa4+ Kd8 73.Qd4+ Ke8 74.Qh8+ Ke7 
2. +-  (8.95): 49...c4 50.Bxc4 g5 51.Qxg5+ Kf8 52.Qf5 Qg8 53.Qc5 a6 54.Bd5 Rd8 55.Bxf7 Qxf7 56.Rxf7+ Kxf7 57.Qd5+ Kf8 58.g4 Rde8 59.Qd6 Kf7 60.g5 Rf8 61.Kg3 Ke8 62.Qxa6 Ref7 63.Qa8+ Ke7 64.Qe4+ Kd6 65.f4 Kc5 66.Qc2+ Kd4 67.Qd2+ Kc5 68.Qc3+ Kb6 69.Qd4+ Kb5 70.a4+ Ka6 71.Qc4+ Kb6 72.Qb5+ Ka7 73.Qd5 Ka6 74.Qd4 Ka5 
3. +-  (8.95): 49...g5 50.Qxg5+ Kf8 51.Qf5 Qg8 52.Qxc5 a6 53.Bc4 Rd8 54.Bxf7 Qxf7 55.Rxf7+ Kxf7 56.Qd5+ Kf8 57.g4 Rde8 58.Qd6 Kf7 59.g5 Rf8 60.Kg3 Ke8 61.Qxa6 Ref7 62.Qa8+ Ke7 63.Qe4+ Kd6 64.f4 Kc5 65.Qc2+ Kd4 66.Qd2+ Kc5 67.Qc3+ Kb6 68.Qd4+ Kb5 69.a4+ Ka6 70.Qc4+ Kb6 71.Qb5+ Ka7 72.Qd5 Ka6 73.Qd4 Ka5 74.Kg4 d6 75.Qxd6 
4. +-  (16.26): 49...Rc8 50.Qf4 g5 51.Qd6 Rce8 52.Rxf7 Rxf7 53.Qxd7 Qe5 54.Bxf7+ Kf8 55.Bxe8 Qe4+ 56.Kh2 Qe7 57.Qxe7+ Kxe7 58.Bg6 Ke6 59.Bxh7 g4 60.Kg2 c4 61.Kf1 Ke5 62.Bc2 Kf6 63.Ke2 Kf7 64.Be4 Kf8 65.Bg6 Kg8 66.Ke3 Kf8 67.Kf4 Kg8 68.Bc2 Kf7 69.Kxg4 c3 70.Kf4 Ke6 
5. +-  (18.14): 49...Rb8 50.Qf4 Rxb3 51.axb3 Kf8 52.Rxf7+ Rxf7 53.Qb8+ Ke7 54.Qxh8 Ke6 55.Qc8 Kd6 56.Qa6+ Ke7 57.Qxa7 Kd6 58.Qa6+ Ke7 59.Qb6 d6 60.Qc7+ Ke6 61.Qc8+ Rd7 62.Qg8+ Rf7 63.f4 Kf6 64.Qd8+ Ke6 65.Qc8+ Rd7 66.Qg8+ Rf7 67.g4 d5 68.Qe8+ Re7 69.Qc6+ Kf7 70.Qxd5+ Ke8 71.Qxc5 Re1 72.g5 Rd1 73.Qc8+ Ke7 74.Qb7+ Kd8 75.Qxh7 Rd2+ 76.Kg3 Rd3+ 77.Kg4 
6. +-  (21.35): 49...Re6 50.Bxe6 dxe6 51.Qf4 c4 52.g4 Qxf6 53.Qxf6 Kf8 54.Qd4 a5 55.Qxc4 Rd8 56.Qc5+ Ke8 57.Qxa5 Rd5 58.Qa8+ Ke7 59.Qb7+ Kf6 60.Qc7 Rd1 61.Qf4+ Ke7 62.Qg5+ Ke8 63.Qb5+ Rd7 64.Qb8+ Ke7 65.a4 f5 66.a5 fxg4 67.Qb5 Rd6 68.Qc5 g3 69.Kxg3 Kd7 70.Qa7+ Ke8 
7. +-  (23.91): 49...d5 50.Bxd5 Rd7 51.Bc4 Ree7 52.Qf4 Kf8 53.Rxf7+ Ke8 54.Rg7 Rb7 55.Rxe7+ Kxe7 56.Qe4+ Kd6 57.Qxb7 Qd4 58.Qa6+ Kc7 59.Qxa7+ Kc8 60.Be2 Qd7 61.Qxc5+ Kd8 62.Qg5+ Kc7 63.Qa5+ Kd6 64.Qb4+ Kc7 65.Bg4 Qc6+ 66.Bf3 Qd7 67.Qc5+ Kd8 68.Bd5 Qe7 69.Qd4 Qa3 70.Qh8+ Kc7 71.Qxh7+ Kb6 72.Qb7+ Kc5 
8. +-  (29.64): 49...Ra8 50.Qf4 Kf8 51.Rxf7+ Ke8 52.Rg7 c4 53.Bxc4 Kd8 54.Qf6 Qe8 55.Bf7 Qf8 56.Rg8 Qxg8 57.Bxg8 Rb8 58.Bxh7 Rb6 59.Qg5 a6 60.Bxg6 Rxg6 61.Qxg6 Re6 62.Qg5+ Kc7 63.Qc5+ Rc6 64.Qe5+ Kb6 65.Qd4+ Ka5 66.h7 Rc2 67.Qd5+ Kb6 
9. +-  (58.60): 49...a6 50.Qf4 c4 51.Bxc4 Rf8 52.g4 d5 53.Bxd5 Rd7 54.Bb3 Ra7 55.g5 Rd7 56.Bc4 a5 57.a4 Ra7 58.Qd6 Rfa8 59.Bd5 Qxf6 60.Qxf6 Kf8 61.Bxa8 Rxa8 62.Qh8+ Ke7 63.Qxa8 f5 64.gxf6+ Kf7 
10. +-  (58.60): 49...a5 50.Qf4 c4 51.Bxc4 d5 52.Bxd5 Rf8 53.a4 Rd7 54.Bc4 Rc7 55.Bb3 Ra7 56.g4 Re7 57.g5 Rd7 58.Ba2 Ra7 59.Qd6 Re8 60.Bc4 Rea8 61.Bd5 Qxf6 62.Qxf6 Kf8 63.Bxa8 Rxa8 64.Qh8+ Ke7 
11. +-  (58.60): 49...d6 50.Qf4 c4 51.Bxc4 Rf8 52.Qxd6 Rfe8 53.Qf4 Rf8 54.g4 Rd7 55.g5 a6 56.Bb3 a5 57.a4 Ra7 58.Ba2 Rd7 59.Bc4 Ra7 60.Qd6 Rfa8 61.Bd5 Re8 62.Qc6 Rb8 63.Re6 fxe6 64.Qxe6+ Kf8 65.Qd6+ Ke8 
12. +-  (58.60): 49...Rf8 50.Qf4 d5 51.Bxd5 Rd7 52.Bc4 Re7 53.g4 Rd7 54.g5 a6 55.Bb3 c4 56.Bxc4 a5 57.a4 Ra7 58.Qd6 Rfa8 59.Bd5 Re8 60.Qc6 Rb8 61.Re6 fxe6 62.Qxe6+ Kf8 63.Qd6+ Ke8 64.Qxb8+ Ke7 

Edit: search is working again, here's one of the threads about the position regards

forum3/viewtopic.php?p=742563#p742563

[Lazy_Frank]

[quote=
Edit: search is working again, here's one of the threads about the position regards

forum3/viewtopic.php?p=742563#p742563
[/quote]

Very funny to read old threads.
I guess after 10-15 years nowadays threads will be funny as well.

[Ovyron]

I guess after 10-15 years nowadays threads will be funny as well.

Or maybe computer chess is dead (engines playing perfect chess at 1 second/move) and nobody reads anything anymore.

[Lazy_Frank]

I guess after 10-15 years nowadays threads will be funny as well.

Or maybe computer chess is dead (engines playing perfect chess at 1 second/move) and nobody reads anything anymore.

Standard chess (initial chess position) at high level indeed is dead.
As well with high probability FRC chess without book.
Still a lot of room in standard chess not covered ( i mean screwed openings, dubious lines). For example Bronstein queen sacrifice line.

[Ovyron]

Standard chess (initial chess position) at high level indeed is dead.

Luckily 5 minute games still lives on. And I don't really get what people are complaining about, if years ago chess at 2 hours/game was dead but now we can approach that quality in much faster games, then what's the problem? We're just saving 1 hour 55 minutes of our lives in time?

It's like some people think games have to be long to matter, and then a draw death is hit and they have to resort to trying to solve it with solutions like unbalanced openings, instead of just playing faster time controls when the problem does not yet arise.