New number one engine

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
Rebel
Posts: 7612
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

New number one engine

Post by Rebel »

Image

Just kidding.
90% of coding is debugging, the other 10% is writing bugs.
Jouni
Posts: 3932
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: New number one engine

Post by Jouni »

New number two engine in CEGT, no kidding:

1 Stockfish 19.0 x64 1CPU 3647 10 10 3700 60.5% 3570 77.6%
2 Integral v8 x64 1CPU 3621 10 10 3299 58.2% 3561 79.7%
Jouni
User avatar
Eelco de Groot
Posts: 4751
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: New number one engine

Post by Eelco de Groot »

Graham Banks is testing it I found by Google :)The Bark Or The Bite The AI says written in Rust but gave no website for sources that I could see.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
User avatar
Ajedrecista
Posts: 2271
Joined: Wed Jul 13, 2011 9:04 pm
Location: Madrid, Spain.

Re: New number one engine.

Post by Ajedrecista »

Hello Eelco:
Eelco de Groot wrote: ↑Sun Sep 20, 2026 6:55 pm Graham Banks is testing it I found by Google :)The Bark Or The Bite The AI says written in Rust but gave no website for sources that I could see.
Given the version numbering and the recent changes, I would look into:

https://github.com/FirePlank/Peras

For NNUE networks, look at:

https://github.com/FirePlank/Peras-networks

Regards from Spain.

Ajedrecista.
User avatar
Eelco de Groot
Posts: 4751
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: New number one engine

Post by Eelco de Groot »

I could not find it, very good! Thanks!

This is superfluous I am sure, but Mark Young now testing in his new Arena version (his GUI) the 3.5.0 of Peras (so this is a .3 plus newer version already than the 3.2.0, but is it also better. Waiting for Mark's test!), Pawnocchio 2.0.1 that Jouni mentioned, Stockfish 18 (not 19!) and the latest Dev Stockfish which is the most interesting of the Stockfishes perhaps because it has Vondele's search explosion fix. I have not tried any of these but the search explosion fix is important. (It's just my opinion but I kept seeing explosions literally everywhere when I tried Stock 19. In long searches, everywhere. But that is just me). Mark's new tournament: https://prodeo.actieforum.com/t1950-top ... gine-arena
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Ultrabean
Posts: 10
Joined: Wed Jun 03, 2026 7:38 am
Full name: Sandy Vee

Re: New number one engine

Post by Ultrabean »

I thought this was interesting.

1 CPU core and 512MB hash each

Time: 5 minutes with 5 seconds increment

Had the 2 engines play the Fried Liver Attack as white and black

Peras 3.6 vs Stockfish 19

[pgn][Event "[05] Engine Match (22/09/2026)"]
[Site "Local computer"]
[Date "2026.09.22"]
[Round "1"]
[White "Peras 3.6.0"]
[Black "Stockfish 19"]
[Result "1/2-1/2"]
[Annotator "?"]
[BlackElo "3750"]
[Classes "0"]
[ECO "C57"]
[FEN "r1bq1b1r/ppp2kpp/2n5/3np3/2B5/8/PPPP1PPP/RNBQK2R w KQ - 0 7"]
[GameID "0"]
[SetUp "1"]
[TimeControl "300+5"]
[WhiteElo "3600"]

7. Qf3+ Ke6 8. Nc3 Ncb4 9. O-O c6 10. d4 Qf6 11. Qd1 Ke7 12. Re1 Qg6 13. a3
Nxc3 14. bxc3 Nd5 15. Rxe5+ Kd8 16. Bd3 Qf7 17. c4 Nf4 18. c5 Bd7 19. Bxf4
Qxf4 20. Qe1 g5 21. Re4 Qf6 22. Qa5+ Kc8 23. Rb1 Qd8 24. Qa4 a5 25. Qb3 Ra7
26. h3 Bg7 27. Ba6 Qc7 28. Qf7 Bxd4 29. Rxd4 bxa6 30. Rbd1 Rd8 31. Qg7 a4 32.
Qxg5 Qa5 33. Rb4 Rb7 34. Rxb7 Kxb7 35. Rb1+ Ka8 36. Rd1 Kb7 37. Rb1+ Ka8 38.
Qe7 Rb8 39. Rxb8+ Kxb8 40. Qxd7 Qe1+ 41. Kh2 Qe5+ 42. Kg1 Qa1+ 43. Kh2 Qe5+
44. g3 Qxc5 45. Qe8+ Kc7 46. Qe3 Qxc2 47. Qe7+ Kb6 48. Qb4+ Kc7 49. Qe7+ Kb6
50. Qb4+ Kc7 51. Qf4+ Kd7 52. Qg4+ Kd6 53. Qd4+ Ke6 54. Qg4+ Kd5 55. Qd7+ Kc5
56. Qf7 Kd4 57. Qd7+ Kc5 58. Qf7 Kd4 59. Qd7+ Ke5 60. Qe8+ Kf6 61. Kg2 c5 62.
g4 c4 63. Qc6+ Ke5 64. Qc7+ Ke6 65. Qc8+ Kd5 66. Qg8+ Kd4 67. Qg7+ Kd5 68.
Qf7+ Kd6 69. Qf6+ Kc7 70. Qe7+ Kc6 71. Qe6+ Kc5 72. Qe5+ Kc6 73. Qe6+ Kc7 74.
Qe7+ Kb6 75. Qd6+ Kb5 76. Qd5+ Kb6 77. Qd6+ Kb5 78. Qd5+ Kb6 79. Qd8+ Kc6 80.
Qf6+ Kd7 81. Qd4+ Kc6 82. Kf3 Qb3+ 83. Ke4 Qb1+ 84. Kf4 Qb8+ 85. Ke4 Qb1+ 86.
Ke3 Qc1+ 87. Ke2 Qxa3 88. Qxc4+ Kb6 89. Kd2 Qxh3 90. Qb4+ Kc7 91. Qf4+ Kc8 92.
Qxa4 Qh6+ 93. Ke2 Kb8 94. Qd7 Qf4 95. f3 Qe5+ 96. Kd3 Qf4 97. Qc6 a5 98. Qb5+
Kc8 99. Qc4+ Qxc4+ 100. Kxc4 Kd7 101. Kb5 Kd6 102. f4 Kd5 103. Kxa5 Ke4 104.
f5 Ke5 105. Kb5 h5 106. gxh5 Kxf5 107. h6 Kg6 108. Kc6 Kxh6 1/2-1/2[/pgn]

Stockfish 19 vs Peras 3.6

[pgn][Event "[05] Engine Match (22/09/2026)"]
[Site "Local computer"]
[Date "2026.09.22"]
[Round "2"]
[White "Stockfish 19"]
[Black "Peras 3.6.0"]
[Result "1-0"]
[Annotator "?"]
[BlackElo "3600"]
[Classes "0"]
[ECO "C57"]
[FEN "r1bq1b1r/ppp2kpp/2n5/3np3/2B5/8/PPPP1PPP/RNBQK2R w KQ - 0 7"]
[GameID "0"]
[SetUp "1"]
[TimeControl "300+5"]
[WhiteElo "3750"]

7. Qf3+ Ke6 8. Nc3 Ncb4 9. O-O c6 10. d4 Qf6 11. Qd1 Ke7 12. Re1 h6 13. Rxe5+
Kd8 14. Ne4 Qg6 15. a3 Bf5 16. Ng3 Bxc2 17. Qf3 Nd3 18. Rf5 Kc7 19. Bd2 Re8
20. Bxd3 Bxd3 21. Qxd3 Bd6 22. Rd1 Re7 23. Qf3 Rhe8 24. h4 Kb8 25. h5 Qh7 26.
Nf1 Re4 27. Bb4 Nxb4 28. axb4 Qg8 29. d5 cxd5 30. Rdxd5 Qe6 31. b5 Be5 32. Rd1
Bxb2 33. Rf7 Re7 34. Rd8+ Kc7 35. Rff8 Kb6 36. Ne3 Rxe3 37. fxe3 Ba3 38. Rd3
Bc5 39. Rf5 Bxe3+ 40. Kh2 Qe4 41. Rd6+ Kc7 42. Qxe4 Rxe4 43. Rg6 Re7 44. Kg3
Bg5 45. Rc5+ Kd7 46. Rd5+ Kc7 47. Kg4 Rf7 48. Rc5+ Kd7 49. b6 axb6 50. Rxb6
Kd8 51. Rc4 Rc7 52. Rcb4 Be7 53. Rb1 Rc2 54. g3 Ke8 55. Rxb7 Kf8 56. Ra7 Rc5
57. Rb8+ Kf7 58. Rb4 Kf8 59. Rbb7 Re5 60. Kf4 Re1 61. Kf3 Re6 62. Rb8+ Kf7 63.
Rb4 Kg8 64. Ra8+ Kf7 65. Re4 Rf6+ 66. Rf4 Rxf4+ 67. Kxf4 g6 68. Kg4 gxh5+ 69.
Kxh5 Kf6 70. Kxh6 Bc5 71. Ra6+ Ke7 72. g4 Kd7 73. g5 Bb4 74. g6 Kc7 75. g7 Bc5
76. g8=Q Be3+ 77. Kg6 Bb6 78. Qe6 Kb8 79. Qc6 Ba7 80. Ra1 Bd4 81. Ra8# 1-0[/pgn]