Weird move of Stockfish

[Henk]

Looks like I am testing Stockfish 15 on hyperbullet time control.

Stockfish playing weird move. Looks like a bug. Or is this Stockfish humor? Playing cat and mouse. What to expect when machines taking over.
44. d1R??



[pgn] [Event "Computer Chess Game"] [Site "GEBRUIK-645EVFP"] [Date "2024.01.16"] [Round "-"] [White "Skipper_15_56"] [Black "Stockfish 15"] [Result "0-1"] [TimeControl "30"] [Annotator "1. +0.26 1... -0.31"] 1. e4 {+0.26/10} c5 {-0.31/21 1.6} 2. Nf3 {+0.57/10 0.5} Nc6 {-0.38/19 0.2} 3. Nc3 {+0.47/10 0.5} e5 {-0.28/21 0.7} 4. Bc4 {+0.64/9 0.5} Be7 {-0.28/20 0.2} 5. Nd5 {+0.65/10 0.5} Nf6 {-0.24/20 1.0} 6. d3 {+0.39/10 0.5} Nxd5 {-0.31/20 0.3} 7. exd5 {+0.71/10 0.5} Nd4 {-0.06/17 0.4} 8. Nxe5 {+0.59/9 0.4} d6 {-0.02/19 0.3} 9. Nf3 {+0.59/10 0.4} Bg4 {+0.33/17 0.4} 10. a4 {+0.50/10 0.4} O-O {+1.70/15 0.5} 11. O-O {+0.57/9 0.4} Bh5 {+1.70/15 0.9} 12. a5 {+0.41/9 0.4} Nxf3+ {+3.29/17 1.3} 13. gxf3 {+0.69/10 0.4} Bg5 {+3.06/20 1.1} 14. Re1 {+0.80/10 0.4} Qf6 {+3.36/19 0.9} 15. Bxg5 {+0.85/10 0.4} Qxg5+ {+3.52/17 0.4} 16. Kh1 {+0.61/10 0.4} Qf4 {+3.99/17 0.5} 17. Kg2 {+0.61/10 0.4} Rae8 {+3.75/22 2.0} 18. Re4 {+0.10/9 0.4} Rxe4 {+4.37/20 0.7} 19. dxe4 {-1.17/10 0.4} f5 {+4.73/17 0.1} 20. Bd3 {-1.14/11 0.4} c4 {+6.39/17 0.5} 21. Be2 {-1.61/11 0.4} fxe4 {+6.41/20 0.7} 22. Ra3 {-2.32/10 0.4} exf3+ {+6.62/16 0.1} 23. Bxf3 {-4.13/11 0.3} Qg5+ {+6.84/15 0.2} 24. Kf1 {-4.61/11 0.3} Rxf3 {+6.98/19 1.0} 25. Rxf3 {-3.18/12 0.3} Qg4 {+7.02/15 0.3} 26. Qe2 {-2.92/13 0.3} Qxf3 {+7.20/16 0.1} 27. Qe6+ {-3.17/12 0.3} Qf7 {+7.46/15 0.2} 28. Qc8+ {-3.10/11 0.3} Qe8 {+7.30/15 0.4} 29. Qxc4 {-3.36/11 0.3} Bf3 {+7.51/17 1.1} 30. b4 {-3.33/10 0.3} Qe5 {+8.29/17 0.4} 31. Qc8+ {-3.21/12 0.3} Kf7 {+9.51/17 0.4} 32. Qe6+ {-3.00/11 0.3} Qxe6 {+9.93/16 0.4} 33. dxe6+ {-3.25/11 0.3} Kxe6 {+9.91/16 0.3} 34. Kg1 {-3.30/9 0.3} Kd5 {+10.92/17 0.5} 35. c3 {-4.78/11 0.3} Kc4 {+11.27/16 0.4} 36. h4 {-5.70/10 0.3} Kxc3 {+11.62/16 0.1} 37. b5 {-6.64/12 0.3} d5 {+12.38/17 0.4} 38. Kh2 {-5.77/11 0.3} d4 {+13.36/16 0.1} 39. Kg3 {-6.64/11 0.3} Bh5 {+13.78/14 0.1} 40. f3 {-11.93/11 0.3} d3 {+14.42/16 0.1} 41. b6 {-14.61/12 0.3} a6 {+16.20/16 0.3} 42. Kf2 {-16.43/12 0.3} d2 {+19.45/17 0.3} 43. Ke2 {-16.88/12 0.3} Bf7 {+21.21/17 0.1} 44. Ke3 {-17.01/11 0.3} d1=R {+22.98/16 0.3} 45. f4 {-11.15/9 0.3} h6 {+25.12/18 0.8} 46. Ke2 {-11.78/10 0.2} Bh5+ {+29.24/22 0.7} 47. Ke3 {-12.28/10 0.2} Rd5 {+57.08/23 0.4} 48. Kf2 {-12.30/10 0.2} Rxa5 {+67.31/21 0.1} 49. Ke3 {-12.93/10 0.3} Ra4 {+67.66/22 0.1} 50. f5 {-14.04/11 0.2} Rxh4 {+99.61/22 0.1} 51. Kf2 {-18.64/11 0.2} a5 {+99.81/23 0.5} 52. Kf1 {-16.37/10 0.2} Rh2 {+99.87/25 0.6} 53. Ke1 {-22.83/13 0.2} Re2+ {+99.91/31 0.3} 54. Kf1 {-22.83/12 0.3} a4 {+99.93/75 0.2} 55. f6 {-22.89/10 0.3} a3 {+99.95/198 0.4} 56. fxg7 {-327.62/9 0.3} a2 {+99.97/115 0.5} 57. g8=Q {-327.64/10 0.3} a1=Q# {+99.99/245 0.5} {Xboard adjudication: Checkmate} 0-1 [/pgn]

[DrCliche]

It's not really a bug, just a random artifact of machine learning and the very low node count at hyperbullet.

[Uri Blass]

It's not really a bug, just a random artifact of machine learning and the very low node count at hyperbullet.

It is not a bug but I think that it may be a bad design even to calculate underpromotions in this case(small depth and promotion to queen is winning for the side that promoted).