Antichess technique applied to International Masters. Man against machine and against chess professionals.

[Father]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

[lkaufman]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

The older bot, "LeelaKnightOdds", does allow the user to play White with a Black knight removed, by a special procedure. This handicap is known as "Knight and Move" odds. Normally in odds chess, it is considered that the odds-receiver has just "captured" (removed) the knight, so logically it should be the odds-giver's move, so he gets White unless a larger handicap is needed. You are correct that it does make a significant difference, the performance of the bot drops by about a class (200 elo) at "Knight and Move" compared to normal knight odds. Basically if White just opens 1.e4, Black is in serious trouble in the opening with either knight missing, he cannot avoid getting a bad position on top of being a knight down. Knight and move was never a popular handicap, as Rook odds is not that much larger and is a more interesting handicap. You are of course welcome to try the "LeelaRookOdds" bot if you want to have some chance to win games rather than only the hope for a draw. It won't be easy, but I guess you can occasionally win with rook odds in blitz. By the way, I expect that the "LeelaKnightOdds-Dev" bot will be upgraded to a new net this weekend; perhaps then even the occasional draw will be too difficult, though that might not be the case.

[Father]

[pgn][[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "02:28:31"]
[WhiteElo "2000"]
[BlackElo "2030"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. e4 d5 2. exd5 Qxd5 3. Nc3 Qd8 4. d4 e6 5. Qg4 Nf6 6. Qg3 g6 7. Nf3 Bd6 8. Qh4 Nh5 9. Qe4 Nf6 10. Qh4 Nh5 11. Qe4 Nf6 12. Qh4 { The game is a draw. } 1/2-1/2/pgn]

[pgn] Event "Casual bullet game"]
[Site "https://lichess.org/0h58zPM3"]
[Date "2024.09.27"]
[White "LeelaRookOdds"]
[Black "Catecan"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "01:34:19"]
[WhiteElo "2000"]
[BlackElo "2030"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. Nf3 d5 2. d3 e6 3. c3 f5 4. Nbd2 Nf6 5. e4 dxe4 6. dxe4 Nxe4 7. Bc4 Nxd2 8. Bxd2 Nc6 9. Ng5 g6 10. Qe2 Bg7 11. O-O O-O 12. Bf4 Qe7 13. Re1 Re8 14. Qe3 Kh8 15. Qg3 e5 16. Nf7+ Kg8 17. Ng5+ Kh8 18. Nf7+ Kg8 19. Ng5+ Kh8 { The game is a draw. } 1/2-1/2[/png][/pgn][/pgn]

[Father]

[pgn] Event "Casual bullet game"]
[Site "https://lichess.org/GcJxUT3j"]
[Date "2024.09.27"]
[White "LeelaRookOdds"]
[Black "Catecan"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "02:28:31"]
[WhiteElo "2000"]
[BlackElo "2030"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. e4 d5 2. exd5 Qxd5 3. Nc3 Qd8 4. d4 e6 5. Qg4 Nf6 6. Qg3 g6 7. Nf3 Bd6 8. Qh4 Nh5 9. Qe4 Nf6 10. Qh4 Nh5 11. Qe4 Nf6 12. Qh4 { The game is a draw. } 1/2-1/2[/pgn]

[Father]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

The older bot, "LeelaKnightOdds", does allow the user to play White with a Black knight removed, by a special procedure. This handicap is known as "Knight and Move" odds. Normally in odds chess, it is considered that the odds-receiver has just "captured" (removed) the knight, so logically it should be the odds-giver's move, so he gets White unless a larger handicap is needed. You are correct that it does make a significant difference, the performance of the bot drops by about a class (200 elo) at "Knight and Move" compared to normal knight odds. Basically if White just opens 1.e4, Black is in serious trouble in the opening with either knight missing, he cannot avoid getting a bad position on top of being a knight down. Knight and move was never a popular handicap, as Rook odds is not that much larger and is a more interesting handicap. You are of course welcome to try the "LeelaRookOdds" bot if you want to have some chance to win games rather than only the hope for a draw. It won't be easy, but I guess you can occasionally win with rook odds in blitz. By the way, I expect that the "LeelaKnightOdds-Dev" bot will be upgraded to a new net this weekend; perhaps then even the occasional draw will be too difficult, though that might not be the case.

Thank you very much Mr. Larry Kaufman, for the oportunity yo me of being playing chess against this engines and parámeters…

[Father]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

The older bot, "LeelaKnightOdds", does allow the user to play White with a Black knight removed, by a special procedure. This handicap is known as "Knight and Move" odds. Normally in odds chess, it is considered that the odds-receiver has just "captured" (removed) the knight, so logically it should be the odds-giver's move, so he gets White unless a larger handicap is needed. You are correct that it does make a significant difference, the performance of the bot drops by about a class (200 elo) at "Knight and Move" compared to normal knight odds. Basically if White just opens 1.e4, Black is in serious trouble in the opening with either knight missing, he cannot avoid getting a bad position on top of being a knight down. Knight and move was never a popular handicap, as Rook odds is not that much larger and is a more interesting handicap. You are of course welcome to try the "LeelaRookOdds" bot if you want to have some chance to win games rather than only the hope for a draw. It won't be easy, but I guess you can occasionally win with rook odds in blitz. By the way, I expect that the "LeelaKnightOdds-Dev" bot will be upgraded to a new net this weekend; perhaps then even the occasional draw will be too difficult, though that might not be the case.

Mr. Larry Kaufman. I want to thank you for introducing me to these three extraordinary types of man versus machine chess, which represent something completely new to me. I also hope that these games that I have played against the chess computer are useful for generating improvements or changes in the algorithms. Considering the fast pace of games, there are many more errors on my part in these examples. All in all, the line of combat is simplified into two simple components of the pulse: The strategy and experience of the human, against the tactical power and knowledge of the human programmers of these silicon monsters. Thanks again.

Human strategy and experience against the tactical power of computers, experience of programmers. Finally a pulse between intelligences.

...[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/4xMvm2Gw"]
[Date "2024.09.27"]
[White "Catecan"]
[Black "LeelaRookOdds"]
[Result "1-0"]
[UTCDate "2024.09.27"]
[UTCTime "05:08:38"]
[WhiteElo "2082"]
[BlackElo "2000"]
[BlackTitle "BOT"]
[Variant "From Position"]
[TimeControl "180+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "1nbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQk - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. d4 Nf6 2. e3 d6 3. c3 e5 4. dxe5 Ng4 5. Nf3 Nxe5 6. Nxe5 dxe5 7. Qxd8+ Kxd8 8. e4 Kd7 9. Bc4 f6 10. O-O Bd6 11. f3 Bc5+ 12. Kh1 Ke7 13. Re1 Be6 14. Bxe6 Kxe6 15. Be3 Bxe3 16. Rxe3 Kd7 17. Na3 Rd8 18. Rd1+ Kc8 19. Rxd8+ Kxd8 20. Rd3+ Nd7 21. Rxd7+ Kxd7 22. Kg1 a6 23. Kf2 Kd6 24. Ke3 Ke7 25. b3 Kd7 26. c4 Kd6 27. Nb1 Kc5 28. Nc3 c6 29. Kd3 Kb4 30. Kc2 Kc5 31. a3 Kd4 32. Kd2 Kc5 33. Kd3 Kd6 34. b4 Kc7 35. c5 g6 36. Ne2 g5 37. g4 b5 38. h3 Kd7 39. Ng3 Ke7 40. Nf5+ Kd7 41. Kc3 Ke6 42. Kb3 Kd7 43. a4 Ke6 44. axb5 axb5 45. Kc3 Kd7 46. Nd6 Ke7 47. Kd3 Ke6 48. Ke3 Ke7 49. Kf2 Ke6 50. Kg3 Ke7 51. h4 Ke6 52. hxg5 fxg5 53. Kf2 Ke7 54. Kg2 Ke6 55. Kg3 Ke7 56. Nf5+ Kf6 57. Ne3 Kf7 58. Ng2 Kf6 59. Ne1 Ke6 60. Nd3 Kf6 61. Kh3 Ke6 62. Kh2 Kf6 63. Nc1 Ke6 64. Nb3 Ke7 65. Na5 Kd7 66. Kg3 Kc7 67. Kf2 Kd7 68. Nb3 Ke7 69. Ke3 Ke6 70. Nd4+ Kd7 71. Nf5 Ke6 72. Nh6 Ke7 73. f4 exf4+ 74. Kf3 Kf6 75. Nf5 Ke6 76. Nd4+ Ke5 77. Ne2 h6 78. Nc3 Ke6 79. Ke2 Ke5 80. Kd3 f3 81. Nd1 Kf4 82. Nf2 Ke5 83. Ke3 Kf6 84. Kxf3 Ke5 85. Ke3 Kf6 86. Nd3 Ke6 87. e5 Kd5 88. Kf3 Ke6 89. Ke4 Ke7 90. Kf5 Kf7 91. e6+ Ke7 92. Ne5 h5 93. gxh5 Kd8 94. Ng4 Ke7 95. h6 Kd8 96. h7 Ke7 97. Ne5 Kd8 98. Ng6 Kc7 99. h8=Q Kb7 100. Qg7+ Kb8 101. Qd7 g4 102. e7 Ka8 103. e8=Q# { White wins by checkmate. } 1-0[/pgn]

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/Bupp13Ff"]
[Date "2024.09.27"]
[White "Catecan"]
[Black "LeelaRookOdds"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "04:50:16"]
[WhiteElo "2082"]
[BlackElo "2000"]
[BlackTitle "BOT"]
[Variant "From Position"]
[TimeControl "180+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "1nbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQk - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. d4 Nf6 2. e3 Nc6 3. f4 d6 4. Nf3 g6 5. c3 Bg4 6. Bb5 d5 7. Qa4 Qa8 8. Ne5 Bd7 9. Nxd7 Nxd7 10. Nd2 Nf6 11. Nf3 Bg7 12. Ne5 O-O 13. Bxc6 bxc6 14. Qxc6 e6 15. Qxa8 Rxa8 16. Nf3 Rb8 17. O-O Bf8 18. Re1 Rb7 19. a3 Rb3 20. g3 a6 21. Kg2 Ne4 22. Ra2 f5 23. Nd2 Rb8 24. Nxe4 fxe4 25. b4 Kf7 26. Bd2 Bd6 27. h4 Ke7 28. Rc1 Kd7 29. g4 h5 30. g5 Rb7 31. Rac2 Rb8 32. c4 dxc4 33. Rxc4 Rb6 34. Kf2 Rb8 35. Ke2 Rb6 36. Kd1 Rb7 37. R1c2 Rb6 38. Kc1 Rb8 39. Kb2 Rb6 40. Kb3 Rb8 41. Ka4 Rb6 42. Be1 Rb5 43. Rc5 Bxc5 44. dxc5 Kc6 45. Rd2 Rb8 46. Bg3 Rf8 47. Kb3 Rf5 48. Kc4 Rd5 49. Rxd5 exd5+ 50. Kb3 Kd7 51. a4 d4 52. exd4 Ke6 53. Kc4 Kf5 54. d5 Kg4 55. Be1 Kf3 56. d6 cxd6 57. cxd6 Ke2 58. Bc3 Kf3 59. d7 e3 60. d8=Q Kf2 61. Qe7 Kf3 62. Qf6 Kf2 63. Qxg6 Kf3 64. Qxh5+ Kg3 65. Qh6 Kf3 66. f5 Kf2 67. Qg6 Kf3 68. h5 a5 69. h6 axb4 70. Bxb4 e2 71. Be1 Kg4 72. h7 Kf3 73. h8=Q Kg2 74. Qh2+ Kxh2 75. Qh5+ Kg2 76. Qxe2+ Kh3 77. g6 { Draw by stalemate. } 1/2-1/2[/pgn]

[pgn][Event "Casual bullet game"]
[Site "https://lichess.org/t9LI4VOg"]
[Date "2024.09.27"]
[White "LeelaRookOdds"]
[Black "Catecan"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "04:01:44"]
[WhiteElo "2000"]
[BlackElo "2030"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. c4 f5 2. h4 Nf6 3. d4 e6 4. Nh3 d5 5. h5 c6 6. h6 g6 7. Bf4 Bd6 8. e3 Bxf4 9. exf4 Qe7 10. Nd2 a6 11. Ng5 Nbd7 12. Bd3 Ne4 13. Ndf3 Nxg5 14. Nxg5 Nf6 15. c5 Ne4 16. Nf3 O-O 17. Qb3 Re8 18. Bxe4 fxe4 19. Ne5 Rf8 20. g3 Rf7 21. f3 exf3 22. Kf2 Rf5 23. Qe3 Rxe5 24. Qxe5 Bd7 25. g4 Rf8 26. Kxf3 Qf6 27. Qxf6 Rxf6 28. g5 Rf8 29. Re1 Kf7 30. Re3 Bc8 31. Re5 Re8 32. b4 Re7 33. a4 Rc7 34. a5 Rd7 35. Kg3 Re7 36. Kf3 Ke8 37. Kg2 Kd7 38. Kf2 Rf7 39. Kg3 Rf5 40. Kf3 Rxe5 41. fxe5 Ke7 42. Ke2 Kf7 43. Kd2 Ke7 44. Kc2 Kd7 45. Kb3 Kc7 46. Kc2 Kd7 47. Kb3 Ke7 48. Kb2 Ke8 49. Kb3 Kd7 50. Kc3 Kc7 51. Kb3 Kb8 52. Ka3 Ka7 53. Ka4 Ka8 54. Ka3 Ka7 55. Kb3 Ka8 56. Ka4 Ka7 57. Ka3 Ka8 58. Kb3 Ka7 59. Ka2 Ka8 60. Kb3 Ka7 61. Ka3 Ka8 62. Ka4 Ka7 63. Ka3 { The game is a draw. } 1/2-1/2[/pgn]

[Father]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

A question to everyone who wants to collaborate with an answer or concept, which I would greatly appreciate.
I would like to know, with the right to be wrong in the answers, what the Fide rating of these three softwares that Mr. Larry Kaufman has informed me about their existence in Lichess would be. Let's take as a reference an open tournament in different time controls with the Swiss system of confrontations, whoever corresponds by points to face each other, abides by the color conditions that the software dictates. I thank you very much and appreciate your help in advance.

What position in the final table could these chess programs hypothetically reach or have in the final or final positions in open tournaments. I thank you very much and appreciate your help in advance.

The software I am referring to are:
LeelaRook odds
Leelaknightodds-dev and,
Leelaknightodds

[Father]

The software I am referring to are:
LeelaRook odds
Leelaknightodds-dev and,
Leelaknightodds

[lkaufman]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

A question to everyone who wants to collaborate with an answer or concept, which I would greatly appreciate.
I would like to know, with the right to be wrong in the answers, what the Fide rating of these three softwares that Mr. Larry Kaufman has informed me about their existence in Lichess would be. Let's take as a reference an open tournament in different time controls with the Swiss system of confrontations, whoever corresponds by points to face each other, abides by the color conditions that the software dictates. I thank you very much and appreciate your help in advance.

What position in the final table could these chess programs hypothetically reach or have in the final or final positions in open tournaments. I thank you very much and appreciate your help in advance.

The software I am referring to are:
LeelaRook odds
Leelaknightodds-dev and,
Leelaknightodds

I'm assuming you mean that the bot would give the same odds at in the name in every game. We don't have any data for games slower than 15'10" Rapid, but in general players play about a class better in classical time limits than at that Rapid time limit, so we can estimate. LeelaKnightOdds is probably around 2600 FIDE level giving knight odds in Rapid, so perhaps 2400 FIDE in classical. Leelaknightodds-dev may be 100 elo stronger, so perhaps 2700 FIDE in Rapid and 2500 classical. LeelaRook odds is about 250 below LeelaKnightOdds, so perhaps 2350 FIDE Rapid and 2150 FIDE classical. Meanwhile, the LeelaKnightOdds-dev bot was just upgraded to a newer, stronger net.

[Father]

... ... Mr. Larry Kaufman. I would believe, unless there is a better opinion to the contrary, that these two softwares would have serious difficulties managing the computer with the black pieces, with the advantage of a horse in favor of the human person. I can observe that the software and hardware compared to the time when the champion Kramnik faced in 2006 and was defeated by the Fritz 8 machine, until today, the machines have evolved favorably too much. In those days, I remember that at all levels of time control, I could some times, achieved a performance in ten-game matches of 40% to 60% in my favor. Today AI marks another route, but still, it has its limits, but in the same way, I suppose that machines, like human beings, have a limit to reach, in the records of the 100-meter dash athletic race. In my youth facing "Chess Challenger Level 7" by Fidelity Electronics in 1980. Keeping in my mind the development of computers and chess programs of our current times was a dream, it represented something like the present dreams of reaching a height close to cyderal space.

A question to everyone who wants to collaborate with an answer or concept, which I would greatly appreciate.
I would like to know, with the right to be wrong in the answers, what the Fide rating of these three softwares that Mr. Larry Kaufman has informed me about their existence in Lichess would be. Let's take as a reference an open tournament in different time controls with the Swiss system of confrontations, whoever corresponds by points to face each other, abides by the color conditions that the software dictates. I thank you very much and appreciate your help in advance.

What position in the final table could these chess programs hypothetically reach or have in the final or final positions in open tournaments. I thank you very much and appreciate your help in advance.

The software I am referring to are:
LeelaRook odds
Leelaknightodds-dev and,
Leelaknightodds

I'm assuming you mean that the bot would give the same odds at in the name in every game. We don't have any data for games slower than 15'10" Rapid, but in general players play about a class better in classical time limits than at that Rapid time limit, so we can estimate. LeelaKnightOdds is probably around 2600 FIDE level giving knight odds in Rapid, so perhaps 2400 FIDE in classical. Leelaknightodds-dev may be 100 elo stronger, so perhaps 2700 FIDE in Rapid and 2500 classical. LeelaRook odds is about 250 below LeelaKnightOdds, so perhaps 2350 FIDE Rapid and 2150 FIDE classical. Meanwhile, the LeelaKnightOdds-dev bot was just upgraded to a newer, stronger net.

Thank you very much Mr. Larry Kaufman. Allow me to present my last two examples of man versus machine with odds. For me, being able to have these confrontations has been a fantasy and great happiness to be able to participate in epic battles with the probability of sustainable and accessible results, where I can find the right balance to face chess machines, silicon monsters. The perfect title of these computers is that principle that they teach us in chess: "Material against time;" "Time against material."
[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/qXgubWWu"]
[Date "2024.09.27"]
[White "Catecan"]
[Black "LeelaRookOdds"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "20:10:57"]
[WhiteElo "2082"]
[BlackElo "2000"]
[BlackTitle "BOT"]
[Variant "From Position"]
[TimeControl "180+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "1nbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQk - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. d4 Nf6 2. e3 d5 3. f4 c5 4. c3 Nc6 5. Nf3 h5 6. Bd3 h4 7. O-O h3 8. g3 Bg4 9. Be2 e6 10. Ne5 Bf5 11. Nxc6 bxc6 12. Bd3 Bd6 13. Bxf5 exf5 14. Qf3 Ng4 15. Re1 Qe7 16. Qe2 g5 17. Nd2 Kd7 18. Nf3 f6 19. a3 Re8 20. Qf1 Qe4 21. Bd2 c4 22. Rad1 Kc7 23. Bc1 a5 24. Bd2 a4 25. Bc1 Rh8 26. Re2 Rh5 27. Rde1 Kb6 28. Nd2 Qd3 29. Nf3 Qe4 30. Nd2 Qd3 31. Nf3 Qe4 { The game is a draw. } 1/2-1/2[/pgn]

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/Deb5DVVR"]
[Date "2024.09.27"]
[White "LeelaRookOdds"]
[Black "Catecan"]
[Result "1/2-1/2"]
[UTCDate "2024.09.27"]
[UTCTime "20:08:03"]
[WhiteElo "2000"]
[BlackElo "2082"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "180+0"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/1NBQKBNR w Kkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. d4 d5 2. Bg5 Nf6 3. e3 e6 4. Bd3 c5 5. c3 c4 6. Bc2 Be7 7. Bf4 Bd6 8. Nd2 Bxf4 9. exf4 Nc6 10. h4 Bd7 11. Ngf3 Qb6 12. Qc1 O-O-O 13. b3 cxb3 14. Nxb3 Na5 15. Nc5 Nc4 16. Ng5 Rhf8 17. O-O Qb2 18. Qxb2 Nxb2 19. Rb1 Nc4 20. Rxb7 a5 21. Ra7 Kb8 22. Rb7+ Kc8 23. Ra7 Kb8 24. Rb7+ Kc8 { The game is a draw. } 1/2-1/2[/pgn]

Thanks to everyone again. CY.