The longest middlegame attacks in computer games?

[LevyRook]

What are the longest middle game attacks in computer games? Preferrably down material. Preferrably the ones which have been proven to be forced or could be proven to be forced.

I would assume someone may be studying this, since it looks like a cool thing to study. (Would you agree?)

My knowledge of computer games is very limited, here's a couple of examples I know of:
[pgn][Event "Stoofvlees vs. Igel"]
[FEN "2kr2nr/1pqbn3/p4p2/3p2pp/Pb6/2NB1N2/1BP2QPP/R4R1K w - - 0 1"]
[Variant "From Position"]
[SetUp "1"]

1. Nb5!! axb5 2. axb5 b6 3. Ra6 Kb7 4. Bd4 Nc8 5. c3 Ba5 6. c4 g4 7. Rxa5!! bxa5 8. b6 Qd6 9. c5 Qc6 10. Nd2 a4 11. Nb3!! axb3 (11... Qe6 12. Na5+ Ka8 13. b7+ Kb8 14. Qf4+ Nd6 15. cxd6 Nh6 16. Ba7+ Kxa7 17. Qd4+ Kb8 18. Qb6 Bb5 19. Qc7+ Ka7 20. Bxb5 Rxd6 21. b8=Q#) 12. Ra1 Nd6 13. Ra7+ Kb8 14. cxd6 Qc1+ 15. Bf1 Nh6 16. Qe2 Qc6 17. Qa6 b2 (17... Rhf8?? 18. b7 Qxa6 19. Bxa6 Bc8 20. Ra8#) (17... Rhe8 18. Qa5! Re6?? 19. b7) 18. Qa2 b1=Q 19. Qxb1 Nf5 20. b7 Nxd6 21. Ra6 Nxb7 22. Rxc6 { White gains material advantage. } 22... Bxc6 23. Bxf6 [/pgn]
In the game Stoofvlees vs. Igel, there was a 21-move middle game attack while down material.
[pgn][Event "Torch vs. Leela Chess 0"]
[FEN "1r3rk1/1bqp1np1/p2Rp3/2p1PP2/6Pp/2N5/PPP5/2K1Q1RB w - - 0 1"]
[Variant "From Position"]
[SetUp "1"]

1. fxe6!! Nxd6 2. Nd5!! Bxd5 3. Bxd5 Rf4 4. exd7+ Kf8 5. Qxh4 Nf7 6. Bxf7 Kxf7 7. Qh5+ Kg8 8. Qe8+ Rf8 9. Qe6+ Rf7 10. Rd1 Kf8 11. g5 g6 12. Qxg6 Ke7 13. Qe4 Kd8 14. Rh1 Rxd7 15. Rh8+ Ke7 16. Qh7+ Ke6 17. Qh6+ Kd5 18. c4+ Kxc4 19. Qe6+ Rd5 20. Rh4+ Kb5 21. Qxd5 { White gains material advantage. }[/pgn]
In the game Torch vs. Leela Chess 0, there was a 20-move middle game attack while down material.
[pgn][Event "Computer Match"]
[Site "London ENG"]
[Date "2018.01.18"]
[EventDate "?"]
[Round "?"]
[Result "1-0"]
[White "AlphaZero (Computer)"]
[Black "Stockfish (Computer)"]
[ECO "A17"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "133"]

1. Nf3 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 O-O 5. a3 Bxc3 6. Qxc3
a5 7. b4 d6 8. e3 Ne4 9. Qc2 Ng5 10. b5 Nxf3+ 11. gxf3 Qf6
12. d4 Qxf3 13. Rg1 Nd7 14. Be2 Qf6 15. Bb2 Qh4 16. Rg4 Qxh2 {White loses the second pawn.}
17. Rg3 f5 18. O-O-O Rf7 19. Bf3 Qh4 20. Rh1 Qf6 21. Kb1 g6
22. Rgg1 a4 23. Ka1 Rg7 24. e4 f4 25. c5 Qe7 26. Rc1 Nf6
27. e5 dxe5 28. Rhe1 e4 29. Bxe4 Qf8 30. d5 exd5 31. Bd3 Bg4
32. f3 Bd7 33. Qc3 Nh5 34. Re5 c6 35. Rce1 Nf6 36. Qd4 cxb5
37. Bb1 Bc6 38. Re6 Rf7 39. Rg1 Qg7 40. Qxf4 Re8 41. Rd6 Nd7
42. Qc1 Rf6 43. f4 Qe7 44. Rxf6 Nxf6 45. f5 Qe3 46. fxg6 Qxc1
47. gxh7+ Kf7 48. Rxc1 Nxh7 49. Bxh7 {White gains material advantage.} 49... Re3 50. Rd1 Ke8 51. Ka2
Bd7 52. Bd4 Rh3 53. Bc2 Be6 54. Re1 Kd7 55. Kb2 Rf3 56. Re5
Rg3 57. Re3 Rg2 58. Kc3 Rg4 59. Rf3 Ke8 60. Rf2 Rg3+ 61. Kb4
Rg4 62. Rd2 Bd7 63. Ka5 Rf4 64. Be5 Rf3 65. Rd3 Rf2 66. Bd1
Bc6 67. Kb6 1-0[/pgn]
In this AlphaZero vs. Stockfish game White attacked while being down at least two pawns for ~32 moves. But there's much lesser chance that the entire attack is forced.

Specific people

Also, do you know specific people (maybe it's you) who are interested in this topic? Who could be interested in similar topics?

I know Tim Krabbe who, in general, collects chess curiosities and chess records.

GM Matthew Sadler (see youtube channel Silicon Road) is very interested in analyzing chess games with engines and analyzing chess games OF engines. He even has a playlist "great engine middlegames" (haven't checked it out yet).

Preliminary judgment

It seems like long middlegame attacks down material are usually <25 moves long (in human games it's kind of the same). But a 30+ moves long attack should theoretically be possible.

[jorose]

I am trying to understand your definition of an attack. Why does your Stoof example start with white down 2 pawns, doesn't that imply the attack started earlier? The AlphaZero game is definitely very unforced, so the only thing I really see that connects those 3 games is that the winning side is down material.

I will give an example where I would argue Winter is down material for 30 moves, but I assume the game doesn't fit your list. Winter gives up a rook for a couple of pawns and then sacrifices another pawn and allows a queen trade in order to trap white's knight. I would argue when cheese gives up the f-pawn material is officially about even. So could you confirm this game does not belong and define why?

[pgn][Event "TCEC Season 26 - Swiss 6"]
[Site "https://tcec-chess.com"]
[Date "2024.01.24"]
[Round "1.11"]
[White "Cheese 3.2.0"]
[Black "Winter 2.07e"]
[Result "0-1"]
[Annotator "archive"]

1. d4 d5 2. c4 dxc4 3. e4 c5 4. d5 Nf6 5. Nc3 b5 6. Bf4 a6 7. e5 Nfd7 8. Nh3 Nb6 9. e6 fxe6 10. Bxb8 Rxb8 11. Qh5+ g6 12. Qe5 exd5 13. Qxh8 Qd6 14. Ng5 Qf6 15. Qxf6 exf6 16. Nxh7 Be7 17. h4 Bf5 18. g4 Bxg4 19. h5 Bxh5 20. Be2 Bxe2 21. Nxe2 d4 22. Ng3 Nd5 23. O-O-O Rb6 24. Ne4 Kf7 25. Rde1 Nf4 26. Kb1 Re6 27. f3 Kg7 28. Reg1 Re5 29. Rg4 g5 30. Rgg1 d3 31. Rh2 Nd5 32. Kc1 Rf5 33. Rh3 Nf4 34. Ng3 Re5 35. Rh2 b4 36. Kd2 c3+ 37. bxc3 c4 38. Ne4 Ra5 39. a4 bxa3 40. Ra1 Nd5 41. Rah1 Rb5 42. Kc1 Rb3 43. f4 gxf4 44. Rg2+ Kf7 45. Rh5 Nxc3 46. Nxc3 Rxc3+ 47. Kb1 Rb3+ 48. Ka2 Rb5 49. Rh1 Re5 50. Rd2 Re2 51. Rh2 f3 52. Rf2 Bc5 53. Nxf6 Bxf2 54. Ne4 Rxe4 55. Kxa3 Bc5+ 56. Ka2 Re2 57. Rb2 f2 58. Rxe2 dxe2 59. Kb2 f1=Q 60. Ka2 Qf2 61. Kb2 e1=Q#[/pgn]

Finally, I wanted to post a game I think belongs on the list for sure, but again has very unforced play:
[pgn]
[Event "?"]
[Site "?"]
[Date "2024.03.31"]
[Round "573"]
[White "Patricia"]
[Black "Midnight v5"]
[Result "1-0"]
[TimeControl "10+0.1"]
[SetUp "1"]
[FEN "r1bqkbnr/pp1p1p1p/n3p1p1/2p5/2P5/2N1P3/PP1PBPPP/R1BQK1NR w KQkq - 0 1"]
[PlyCount "89"]
[GameDuration "00:00:26"]
[GameEndTime "2024-03-31T00:28:15.209 CDT"]
[GameStartTime "2024-03-31T00:27:48.932 CDT"]

1. d4 Nf6 2. Nf3 Bg7 3. e4 cxd4 4. Qxd4 Nb4 5. O-O Nc2 6. Qd6 Bf8 7. Qd1
Nxa1 8. e5 Ng8 9. b4 Bxb4 10. Nb5 Ne7 11. Qd6 Bxd6 12. Nxd6+ Kf8 13. Bh6+
Kg8 14. Bd3 Nf5 15. Bxf5 exf5 16. Rc1 f6 17. Re1 Qb6 18. exf6 Qxd6 19. Ng5
Qf8 20. Bxf8 Kxf8 21. Re7 b6 22. Rf7+ Ke8 23. Re7+ Kd8 24. Nf7+ Kc7 25.
Nxh8 Ba6 26. Rxh7 Rf8 27. Nxg6 Rxf6 28. Ne5 Rd6 29. h4 Nc2 30. h5 Na3 31.
Rxd7+ Rxd7 32. Nxd7 Bb5 33. cxb5 Nc4 34. h6 a6 35. bxa6 Nd6 36. Ne5 Nc4 37.
Nxc4 f4 38. h7 f3 39. h8=Q Kd7 40. Qd4+ Ke6 41. Qe4+ Kf6 42. a7 b5 43. a8=Q
Kg7 44. Qf5 bxc4 45. Qaf8# 1-0
[/pgn]

[LevyRook]

jorose, first of all, thank you very much for those games. Especially the second one is very cool.

But my definition of "attack" and my list don't matter that much. The main questions are:

Does anybody study the length of middlegame attacks in any systematic way? Does anybody collect a database? Could anyone be interested in the topic?

Apologies if I didn't make clear enough that those questions are the main focus.

I am trying to understand your definition of an attack. Why does your Stoof example start with white down 2 pawns, doesn't that imply the attack started earlier? The AlphaZero game is definitely very unforced, so the only thing I really see that connects those 3 games is that the winning side is down material.

The bigger the material disadvantage, the easier it is to check if an attack is forced or not (because if the compensation for the lost material runs out, you crumble immediately). Though sometimes it's still complicated, of course. That's why I focused on games with big enough (-2) material disadvantage.

Also, if you want to know my interests, here's a more detailed post by me. I think instead of a single definition it makes sense to explore a function of multiple variables ("maximal possible length", "size of the material disadvantage", "the amount of pieces on the board" and so on).