+2.64 of stockfish is probably a draw

[Eelco de Groot]

Instead of 41.g3? 41.Qh3 might already be won for White:

[d]R2b1k1n/2q2p1p/2p2p2/4p2P/4P3/7Q/5PPK/8 b - - 0 1

Analysis by Stockfish dev-20251112-8e5392d7:

41...Ke7 42.Df5 Sg6 43.hxg6 hxg6 44.Dh3 Db7 45.Da3+ Ke8 46.Da4 Db6 47.Dc2 Db7 48.Ta3 Le7 49.Th3 Ld6 50.Th8+ Ke7 51.Dc4 Dd7 52.Kg1 f5 53.exf5 Dxf5 54.Dxc6 Dd7 55.De4 Db5 56.g3 Db4 57.Dc6 Dc5 58.Db7+ Kf6 59.Th7 Dc1+ 60.Kh2 Dc4 61.Kg2 Lc7 62.Df3+ Ke7 63.Th4 Db5 64.De4 Dc5 65.Th8 Dd4 66.Tb8
Tiefe: 51/129 00:01:00 3130MN, tb=784958
+- (2.06)

So the win for White was probably given away at 41.g3? already, regards

I think the diagram is not the right one but analysis is just too slow on my computer. Here you need tablebases and a lot of memory I suppose. Crystal 5 KWK also did go wrong at some time with the following line

55 1077:57 +0.65 5...Db2 6.Td7 Dc2 7.De3 Lc5 8.Tc7 Lxe3 9.Txc2 Ld4 10.g4 Ke8 11.Tc7 Lb2 12.Kf3 La1 13.Ke3 Ld4+ 14.Ke2 Lb2 15.Kd3 La3 16.Tc3 Lf8 17.g5 fxg5 18.Kc4 g4 (59.908.352.086) 926
55 1077:57 +0.80 5...Dc4 6.g4 Kg8 7.Td7 De6 8.Tc7 Kf8 9.Tc3 Kg7 10.Tc1 Lf8 11.Kg3 De7 12.Th1 Kg8 13.Dc3 Dd6 14.Kf3 Kg7 15.Dc1 Kg8 16.Dc2 Da3+ 17.Kg2 Dd6 18.Td1 De6 (59.908.352.086) 926
55 1077:57 +0.98 5...Da4 6.Tc1 Dd4 7.De2 Kg7 8.Tc4 Dd6 9.Dc2 Ld8 10.Tc6 Dd4 11.Tc5 Le7 12.Tc4 Dd6 13.Tc6 Dd4 14.Ta6 Lc5 15.Ta5 Lb4 16.Tb5 Le7 17.Dc6 f5 18.exf5 gxf5 (59.908.352.086) 926
-----
56 1100:18 +1.93 5...Db2 6.Td7 Dc2 7.Dd3 Dxd3 8.Txd3 f5 {??? this is nonsense}9.exf5 gxf5 10.Tf3 f4 11.gxf4 exf4 12.Kh3 Lc5 13.Kg4 Ld4 14.Kg5 Kg7 15.Kf5 La1 16.Txf4 Lb2 17.Tg4+ Kf8 18.Ke4 La1 (61.288.578.828) 928
55 1100:18 +0.83 5...Dc4 6.g4 Kg8 7.Td7 De6 8.Tc7 Kf8 9.Tc3 Kg7 10.Tc1 Lf8 11.Kg3 De7 12.Th1 Kg8 13.Dc3 Dd6 14.Kf3 Kg7 15.Dc1 Kg8 16.Dc2 Da3+ 17.Kg2 Dd6 18.Td1 De6 (61.288.578.828) 928
55 1100:18 +1.19 5...Da4 6.Tc1 Dd4 7.De2 Kg7 8.Tc4 Dd6 9.Dc2 Ld8 10.Tc6 Dd4 11.Tc5 Le7 12.Tc4 Dd6 13.Tc6 Dd4 14.Ta6 Lc5 15.Ta5 Lb4 16.Tb5 Le7 17.Dc6 f5 18.exf5 gxf5 (61.288.578.828) 928

Probably a bug or hash fault? It leads to a Mate in 72, included in the pgn but it is nonsense of course.

I then picked a point in your line after 58...Kf6 and that Crystal after about a day calculating can confirm seems to be winning. So rest of my analysis not very helpful I'm afraid. Need more memory


[d]7R/1Q3p2/3b1kp1/2q1p3/8/6P1/5P2/6K1 w - -

Engine: Crystal 5 KWK (1024 MB)
gemaakt door the Stockfish developers (see AUTHORS f

58 1544:13 +2.59 59.Th7 Dc4 60.Kg2 De6 61.Dc6 Ke7
62.Th8 Dd7 63.Dd5 Kf6 64.Ta8 De6
65.Dc6 Le7 66.Db5 Df5 67.Kg1 De4
68.Ta4 Df3 69.Ta6+ Kg7 70.Dxe5+ Lf6
71.Dd6 Lg5 72.Ta5 (85.824.847.501) 926

58 1544:13 +1.63 59.Tg8 Dc1+ 60.Kg2 Dc4 61.Td8 De6
62.Db5 Kg7 63.Ta8 Lc7 64.Db7 Df5
65.Kg1 Ld6 66.Db5 De6 67.Dc6 Lf8
68.Dxe6 fxe6 69.Tc8 Kf7 70.Kg2 La3
71.Kf3 Kf6 72.Ke4 (85.824.847.501) 926

57 1544:13 +1.45 59.Td8 Dc1+ 60.Kg2 Dc7 61.Da8 Le7
62.Tg8 Dd7 63.Df3+ Df5 64.Dc6+ De6
65.Db7 Lc5 66.Tc8 Ld6 67.Da8 Df5
68.Tc3 Kg7 69.Dd5 Le7 70.Tf3 De6
71.Dxe6 fxe6 72.Te3 (85.824.847.501) 926

[pgn]
[Event "casual classical game"]
[Site "https://lichess.org/cMMxF0qB"]
[Date "2025.11.29"]
[Round "?"]
[White "LeelaRookOdds"]
[Black "BishopBoss08"]
[WhiteElo "2000"]
[BlackElo "1500"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq -"]

1. c4 e6 2. d3 d5 3. Qc2 Nf6 4. Nf3 Nc6 5. a3 a5 6. d4 Be7
7. Nc3 dxc4 8. e3 Bd6 9. Bxc4 e5 10. d5 Ne7 11. h3 c6
12. dxc6 bxc6 13. e4 Ba6 14. Bxa6 Rxa6 15. O-O Ra8 16. Bg5
Ng6 17. Bxf6 gxf6 18. Nd1 O-O 19. Ne3 Re8 20. Rd1 Qc7
21. Nh2 Rad8 22. Nhg4 Re6 23. Nh6+ Kg7 24. Nef5+ Kf8
25. Rf1 Qb6 26. h4 Bc5 27. Kh2 Red6 28. Nxd6 Rxd6 29. b4
axb4 30. axb4 Bxb4 31. Qb3 Nh8 32. h5 Bc5 33. Qh3 Qc7
34. Rc1 Bb6 35. Nf5 Rd8 36. Qh4 Rd6 37. Ra1 Qd8 38. Nxd6
Qxd6 39. Ra8+ Bd8 40. Qg4 c5 (40... Qc7 41. Qh3 (41. g3 Ke7
42. Qf5 Ng6 43. hxg6 hxg6 44. Qf3 Qd6 45. Ra7+ Ke8 46. Kg2
Qd4 47. Ra8 Kd7 48. Ra6 Ke8 49. Rxc6 Kf8 50. Rc1 Be7
51. Rd1 Qb2 52. Rd7 Qc2 53. Qd3 Qxd3 54. Rxd3 f5 {???}
55. exf5 gxf5 56. Rf3 f4 57. gxf4 exf4 58. Kh3 Bc5 59. Kg4
Bd4 60. Kxf4 Bb2 {+M72 Online Shredderbases}) 41... Ke7
42. Qf5 Ng6 43. hxg6 hxg6 44. Qf3 (44. Qh3 Qb7 45. Qa3+ Ke8
46. Qa4 Qb6 47. Qc2 Qb7 48. Ra3 Be7 49. Rh3 Bd6 50. Rh8+
Ke7 51. Qc4 Qd7 52. Kg1 f5 53. exf5 Qxf5 54. Qxc6 Qd7
55. Qe4 Qb5 56. g3 Qb4 57. Qc6 Qc5 58. Qb7+ Kf6 {About
midpoint in Peter's line.This looks very good for White.
See Crystal 5KWK analysis and diagram above.}) 44... Qd7 45. Qa3+ Ke8 46. Rb8)
41. Qc8 1-0
[/pgn]

[peter]

Instead of 41.g3? 41.Qh3 might already be won for White:

[d]R2b1k1n/2q2p1p/2p2p2/4p2P/4P3/7Q/5PPK/8 b - - 0 1

Analysis by Stockfish dev-20251112-8e5392d7:

41...Ke7 42.Df5 Sg6 43.hxg6 hxg6 44.Dh3 Db7 45.Da3+ Ke8 46.Da4 Db6 47.Dc2 Db7 48.Ta3 Le7 49.Th3 Ld6 50.Th8+ Ke7 51.Dc4 Dd7 52.Kg1 f5 53.exf5 Dxf5 54.Dxc6 Dd7 55.De4 Db5 56.g3 Db4 57.Dc6 Dc5 58.Db7+ Kf6 59.Th7 Dc1+ 60.Kh2 Dc4 61.Kg2 Lc7 62.Df3+ Ke7 63.Th4 Db5 64.De4 Dc5 65.Th8 Dd4 66.Tb8
Tiefe: 51/129 00:01:00 3130MN, tb=784958
+- (2.06)

So the win for White was probably given away at 41.g3? already

I think the diagram is not the right one

The diagram you quoted, is the one after 40...Qc7 (already deviating from the original game, where 40...c5? was played) 41.Qh3, while Uri's analysis- line startet with 41.g3?, 41.Qh3 instead would have won rather surely, even if still dependent on hardware- time and engines a little, but with about equal strength of opponents and reasonable hardware- TC (yet I wouln't like to have to play that with Black from there in a corr.- game neither) White should win that. So the discussion about WDL- probability should start latest (didn't look at the rest of the game leading there, have no idea about the opening and the opponents) at this one point, regards