What is the biggest upset between engines ?

[Chessqueen]

We always see upset between two human players, but what is the biggest upset that you have witness between two engines, for instance a 3100 vs 3400 rated engine? Please post the game.

[algerbrex]

Blunder's not a top engine, and this wasn't a huge upset, but I remember from Bastiball's Mini-TCEC championship last year, Blunder 7.4.1 (~2560), beat Colossus 2021b, 2748. So about a 190 Elo gap:

[pgn][Event " Mini-TCEC Tournament: Season 1"]
[Site "CCRL"]
[Date "2021.12.22"]
[Round "1"]
[White "Colossus 2021b"]
[Black "Blunder 7.4.1"]
[Result "0-1"]
[TimeControl "0/1800+5"]
[Time "13:12:39"]
[Board "78"]
[Termination "mate"]
[ECO "B28"]
[Opening "Sicilian"]

1. e4 c5 2. Nf3 a6 {End of openingB28: Sicilian, O'Kelly variation}
3. Nc3 {+0.40/17 88584 56222760} Nc6 {+0.35/18 47628 4829813} 4. d4 {+0.41/19 106572 58531416} cxd4 {+0.40/18 46576 8136230}
5. Nxd4 {+0.38/18 42968 20832692} d6 {+0.35/17 45520 8581429} 6. Nxc6 {+0.54/18 94549 61821520} bxc6 {+0.33/19 57867 6186579}
7. Bc4 {+0.53/18 85745 50929594} e6 {+0.33/18 43191 7667650} 8. Qe2 {+0.63/17 74751 43367433} Nf6 {+0.22/18 54901 5085341}
9. Bf4 {+0.58/18 92911 53397213} d5 {+0.17/18 40989 3358020} 10. O-O-O {+0.59/17 81168 46612372} Be7 {+0.17/18 40090 5971549}
11. Rhe1 {+0.53/17 78968 42235035} Bb7 {+0.17/17 39214 4482886} 12. Qf3 {+0.48/16 62732 34040068} O-O {+0.12/17 38358 4386639}
13. e5 {+0.45/16 60840 33187707} Nd7 {+0.06/15 37523 1787968} 14. Bd3 {+0.38/17 60685 35645717} g6 {-0.10/16 36704 5544556}
15. Kb1 {+0.36/16 65776 36646575} c5 {-0.24/16 35916 7306260} 16. Qh3 {+0.23/16 53418 26418563} Qc7 {-0.44/14 35139 4503760}
17. Qg3 {+0.23/16 58870 32516844} Rab8 {-0.57/17 34392 1745728} 18. b3 {-0.24/16 80741 45595272} c4 {-0.76/15 33658 1973167}
19. Bf1 {-0.45/16 34754 15839477} Rfc8 {-0.98/18 32940 3936144} 20. Kb2 {-0.75/17 42801 21757569} Qa5 {-1.36/17 54479 3196875}
21. Nb1 {-1.13/16 67904 37611110} cxb3 {-1.43/18 40303 9269298} 22. axb3 {-0.83/14 25849 13406113} d4 {-2.03/17 66166 4555255}
23. Re2 {-0.93/16 75560 39422631} Qb6 {-2.04/16 81644 4877181} 24. Be3 {-1.02/17 48613 24475474} Bc5 {-2.14/16 58601 6551358}
25. Bf4 {-0.76/14 27058 17220528} Bd5 {-2.40/14 32934 1156603} 26. Red2 {-1.16/15 37157 23555529} Bf8 {-2.60/15 32027 3484676}
27. Qd3 {-2.37/13 19665 10980690} Nc5 {-5.24/18 31144 1335367} 28. Qxd4 {-7.61/15 17350 10193566} Qa5 {-4.84/18 30302 3302027}
29. Rd3 {-6.54/15 21438 12502683} Na4+ {-8.56/16 22684 512633} 30. Kc1 {-8.40/17 36193 20782795} Bxb3 {-8.47/17 28903 1394652}
31. R1d2 {-8.28/16 10429 5483415} Ba2 {-9.05/20 47522 5571079} 32. Na3 {-8.94/17 26860 14275275} Bxa3+ {-9.76/19 34762 2146251}
33. Kd1 {-9.71/15 9930 5415244} Rb4 {-10.13/19 19831 4235425} 34. Qd7 {-9.96/15 20896 11157191} Nc3+ {-10.35/18 19459 347905}
35. Rxc3 {-10.25/15 8232 5338170} Rb1+ {-10.59/16 32272 68276} 36. Ke2 {-10.53/15 9232 5344635} Qxc3 {-10.66/16 31114 864502}
37. Be3 {-10.77/16 27850 17472181} Bc4+ {-10.54/17 17761 3020926} 38. Rd3 {-11.44/16 52619 33466605} Bxd3+ {-11.01/17 17442 422852}
39. Qxd3 {-11.82/16 8926 5999199} Qxd3+ {-11.45/19 17135 4157992} 40. cxd3 {-12.34/19 12800 8975608} Rc2+ {-11.79/21 16827 2509858}
41. Kf3 {-13.67/20 37128 24290324} Rxf1 {-11.98/21 16538 6040187} 42. h4 {-14.24/19 11862 7941005} a5 {-12.43/17 16243 1436015}
43. h5 {-15.89/16 6840 4460166} gxh5 {-13.73/19 15963 1486284} 44. g3 {-17.51/16 6501 4313362} a4 {-15.97/23 20391 1954873}
45. g4 {-19.99/17 8658 5857421} Bb2 {-21.07/20 15302 2560523} 46. Kg2 {-21.50/15 6490 4810152} Re1 {-20.29/17 15042 390328}
47. Kg3 {-25.44/14 7218 5354457} a3 {-28.10/17 14793 3571230} 48. Bf4 {-33.08/14 9807 7067600} a2 {-25.48/15 24583 762102}
49. Kh4 {M-6/17 6561 3674798} a1=Q {M+6/14 14065 6677822} 50. gxh5 {M-5/24 6987 4013987} Rxf2 {M+5/13 13838 4886633}
51. d4 {M-4/51 5188 6451322} Rxf4+ {M+4/13 13612 4736376} 52. Kg3 {M-3/120 5024 10082628} Qa3+ {M+3/14 13404 8040430}
53. Kxf4 {M-2/120 5024 8551596} Bc1+ {M+2/15 13190 8433016} 54. Kg4 {M-1/120 5018 1640826} Re4# {M+1/16 12981 7685198}
0-1
[/pgn]

[hgm]

Micro-Max once beat Glaurung in a bullet game that I had started for testing a new GUI feature.

[Chessqueen]

Micro-Max once beat Glaurung in a bullet game that I had started for testing a new GUI feature.

This is the biggest upset that I have seen between engines, after Rubichess beat Stockfish 3 about 7 times straight I returned back form the grocery store and saw this draw WOW.

[pgn][Event "Computer chess game"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.07.03"]
[Round "8"]
[White "RubiChess-20220223_x86-64-avx2"]
[Black "Stockfish-3-64-ja"]
[Result "1/2-1/2"]
[BlackElo "3145"]
[ECO "D77"]
[Opening "Neo-Grünfeld, 6.O-O dxc4"]
[Time "17:17:34"]
[WhiteElo "3465"]
[TimeControl "900+10"]
[Termination "normal"]
[PlyCount "197"]
[WhiteType "program"]
[BlackType "program"]

1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d5 6. O-O dxc4 7. Nbd2 b5 8.
Ne5 Nd5 9. b3 Nc3 10. Qe1 Qxd4 11. Nxf7 Nd5 12. Ng5 h6 13. bxc4 bxc4 14. e3
Qe5 15. Nxc4 Qxg5 16. e4 Qh5 17. exd5 Bxa1 18. d6 exd6 19. Bxa8 Bh3 20. Bg2
Bxg2 21. Qe6+ Kh7 22. Kxg2 Bg7 23. f3 Qb5 24. Rf2 Nc6 25. Bb2 Nd4 26. Bxd4
Bxd4 27. Qe7+ Kg8 28. Rc2 Qf5 29. Qe2 Rf6 30. Rd2 Re6 31. Qd1 c5 32. Na3 a6
33. h4 Qe5 34. Nc4 Qf6 35. Na3 Qe7 36. Nc2 Bg7 37. Kf2 g5 38. hxg5 hxg5 39.
Kg2 Qe8 40. Rf2 Qc6 41. Qd2 g4 42. Qf4 gxf3+ 43. Rxf3 Rf6 44. Qc4+ d5 45.
Rxf6 Bxf6 46. Qg4+ Bg7 47. Kf3 c4 48. Kf2 Qf6+ 49. Ke2 Qe5+ 50. Kf2 c3 51.
Ne1 Qc7 52. Ke2 Qc6 53. Qf5 d4 54. Kd1 Qc4 55. Nd3 Qa4+ 56. Ke1 Qc6 57. Kd1
a5 58. Ne1 a4 59. Nd3 Qh1+ 60. Ne1 Qh6 61. Qc8+ Kh7 62. Qf5+ Qg6 63. Qh3+
Kg8 64. Qc8+ Bf8 65. Qc4+ Kh7 66. Qxa4 Qh5+ 67. Kc1 Bh6+ 68. Kb1 Bg7 69.
Nd3 Qd5 70. Qa6 Qh1+ 71. Kc2 Qg2+ 72. Kb3 Qf1 73. Qc4 Qb1+ 74. Ka4 Qd1+ 75.
Kb4 Qh1 76. Ka4 Qb1 77. Nb4 Qd1+ 78. Ka3 Qg4 79. Kb3 Kh8 80. Qd5 Bf6 81.
Qf7 Qd1+ 82. Ka3 Bg5 83. Qf8+ Kh7 84. Qf7+ Kh6 85. Qe6+ Kh5 86. Qh3+ Kg6
87. Qe6+ Bf6 88. Nd5 Qc1+ 89. Ka4 Qd1+ 90. Ka5 Qf3 91. Nf4+ Kg5 92. Qg8+
Kf5 93. Qg6+ Ke5 94. Qe8+ Kf5 95. Qe6+ Kg5 96. Qg8+ Kf5 97. Qg6+ Ke5 98.
Qe8+ Kf5 99. Qg6+ {3-fold repetition} 1/2-1/2[/pgn]

[Chessqueen]

This is strange but I increase the time control to 1800/10 and for the 2nd game I used the 3-4-5 EGTB, and re-started the same game from move 22 , but from move 26 the game took a different direction and the game was a totally different game, but still ended up in a DRAW, which to me regardless of the T/C it is the biggest upset that I have seen between 2 engines

[pgn][Event "Computer chess game"]
[Site "DESKTOP-OFQ3C0P"]
[Date "2022.07.03"]
[Round "9"]
[White "RubiChess-20220223_x86-64-avx2"]
[Black "Stockfish-3-64-ja"]
[Result "1/2-1/2"]
[BlackElo "3220"]
[ECO "D77"]
[Opening "Neo-Grünfeld, 6.O-O dxc4"]
[Time "18:51:52"]
[WhiteElo "3465"]
[TimeControl "1800+10"]
[Termination "adjudication"]
[PlyCount "121"]


1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d5 6. O-O dxc4 7. Nbd2 b5 8.
Ne5 Nd5 9. b3 Nc3 10. Qe1 Qxd4 11. Nxf7 Nd5 12. Ng5 h6 13. bxc4 bxc4 14. e3
Qe5 15. Nxc4 Qxg5 16. e4 Qh5 17. exd5 Bxa1 18. d6 exd6 19. Bxa8 Bh3 20. Bg2
Bxg2 21. Qe6+ Kh7 22. Kxg2 Bg7 23. f3 Nc6 24. Bb2 Qc5 25. Qd7 Nd4 26. Ne3
Qb4 27. Bxd4 Qd2+ 28. Rf2 Qxd4 29. Qe6 c6 30. f4 Rf6 31. Qe7 c5 32. Rc2 Rf8
33. Kf3 Rb8 34. Qe6 a6 35. h4 Rf8 36. Rc1 Qb2 37. Rd1 Qb7+ 38. Qd5 Qc8 39.
Qxd6 Bd4 40. Nd5 Rf7 41. Re1 Qf5 42. Qe6 Qd3+ 43. Kg2 Kg7 44. Qe2 Qb5 45.
h5 Qxe2+ 46. Rxe2 gxh5 47. Kf3 Rd7 48. Ke4 Kg6 49. Rh2 Rb7 50. f5+ Kg5 51.
Nf4 Kg4 52. Rh4+ Kxg3 53. Rxh5 Rb4 54. Nd3 Ra4 55. Kd5 Kg4 56. Rxh6 Be3 57.
Nxc5 Rxa2 58. Re6 Bxc5 59. Kxc5 Kxf5 60. Rh6 a5 61. Kb5 {User Adjudication}
1/2-1/2[/pgn]