Release of 4 new Rebel's

[Andre]

Anyone knows how to compile the source for Mac? I'd love to give it a try

[pohl4711]

Maybe if stockfish makes it’s nets available for Rebel that might be a fair trade. I’m thinking of making some of my nets available for Rebel 14.1.

Both the new Rebel and Stockfish are open source, so Rebel could use the Stockfish net if it chose to (giving acknowledgement) - or am I missing something?

No, you missed nothing. Of course, all Stockfish nets are open source and free for downloading:
https://tests.stockfishchess.org/nns

[Gabor Szots]

Release of 4 new Rebel's with neural nets created by Chris Whittington playing the famous CSTAL style.

http://rebel13.nl/windows/rebel-14.html#other

https://github.com/egh-s/Rebel-14.1.02- ... stem-TAL-1

What is the difference between the 4 exes? The net?

[Rebel]

Release of 4 new Rebel's with neural nets created by Chris Whittington playing the famous CSTAL style.

http://rebel13.nl/windows/rebel-14.html#other

https://github.com/egh-s/Rebel-14.1.02- ... stem-TAL-1

What is the difference between the 4 exes? The net?

Yes.

[mclane]

here the latest CSTal games:

60/60
1 thread
ryzen 5 3600
ponder ON

[pgn]
[Event "CSTal3 NNUE 5'"]
[Site "ORION8"]
[Date "2022.02.03"]
[Round "1"]
[White "Chess System Tal 3 [5']"]
[Black "Deep Rybka 4.1 x64 [t1]"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "B21"]
[Opening "Sizilianisch"]
[Time "17:00:27"]
[Variation "Morra-Gambit, 4.Nxc3 Nc6 5.Nf3 e6"]
[WhiteElo "2200"]
[TimeControl "60/3600:60/3600:60/3600"]
[Termination "normal"]
[PlyCount "100"]
[WhiteType "program"]
[BlackType "program"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Bc5 5. Nb3 Bb6 6. Nc3 Ne7 7. Bf4 d5
{-0.31/19 201} 8. Qf3 {+0.61/23 132} O-O {-0.30/19 0} 9. O-O-O {+1.17/25
90} d4 {-0.24/21 219} 10. Qg3 {+1.17/25 0} Nbc6 {-0.24/21 132} 11. Kb1
{+1.17/25 0} Ng6 {0.00/20 202} 12. Bc1 {+1.17/25 0} Qe7 {-0.08/19 170} 13.
Na4 {+0.57/26 83} Bc7 {-0.08/18 0} 14. f4 {+0.79/26 101} b6 {-0.12/18 0}
15. c3 {+0.43/25 144} dxc3 {-0.07/19 49} 16. Nxc3 {+0.39/24 41} Bb7
{-0.07/19 20} 17. h4 {+0.94/27 83} Rfd8 {-0.07/18 0} 18. Rxd8+ {+0.86/26
91} Rxd8 {-0.07/19 0} 19. h5 {+0.80/28 84} Nf8 {-0.03/20 0} 20. h6
{+0.79/27 72} g6 {-0.03/20 0} 21. Be2 {+0.80/26 105} Nd7 {0.00/19 0} 22.
Be3 {+0.87/26 76} Nc5 {0.00/20 142} 23. Nxc5 {+0.87/26 0} bxc5 {0.00/22 71}
24. Rc1 {+0.87/25 1} Nd4 {+0.15/18 145} 25. Bc4 {+0.87/25 13} a5 {+0.17/19
98} 26. a4 {+0.53/24 86} Rd7 {+0.15/19 256} 27. Rd1 {+0.52/22 103} Bc6
{+0.17/19 92} 28. Qf2 {+0.52/23 10} Qe8 {0.00/18 231} 29. Ka2 {+0.58/23 63}
Qb8 {0.00/17 75} 30. b3 {+0.61/23 87} Qf8 {0.00/18 0} 31. f5 {+0.59/23 52}
exf5 {0.00/18 0} 32. exf5 {+1.30/23 59} Nxf5 {-0.16/20 0} 33. Rxd7
{+0.72/26 60} Bxd7 {-0.16/21 0} 34. Bg5 {+1.10/28 75} Be5 {-0.28/19 130}
35. Nd5 {+1.09/27 86} Bc6 {-0.33/20 0} 36. Qd2 {+0.84/29 81} Qd6 {-0.30/20
0} 37. g4 {+0.93/29 72} Bxd5 {-0.30/19 89} 38. Qxd5 {+1.66/31 60} Qxd5
{-0.67/22 0} 39. Bxd5 {+1.87/31 16} Ng3 {-0.67/21 16} 40. Kb1 {+1.81/32 46}
Nf1 {-0.67/20 0} 41. Bc4 {+2.05/34 63} Nh2 {-1.35/20 0} 42. Bd8 {+1.96/34
66} Nxg4 {-0.67/19 0} 43. Bxa5 {+1.96/33 64} Nxh6 {-0.60/20 0} 44. Bb6
{+2.03/32 87} Nf5 {-0.74/20 0} 45. a5 {+1.84/32 85} Nd6 {-0.76/23 0} 46.
Bd3 {+1.95/29 64} Bd4 {-0.76/24 94} 47. b4 {+4.12/30 69} Nc8 {-0.78/20 77}
48. bxc5 {+4.12/30 0} g5 {-0.78/19 141} 49. Bf5 {+7.38/30 73} Nxb6
{-3.77/20 0} 50. cxb6 {+7.86/29 92} Kf8 {-3.77/19 0 Schwarz gibt auf} 1-0

[Event "CSTal3 NNUE 5'"]
[Site "ORION8"]
[Date "2022.02.03"]
[Round "1"]
[White "Chess System Tal 3 [5']"]
[Black "DeepHiarcs14WCSC [t1]"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "B92"]
[Opening "Sizilanisch (Najdorf-Variante mit 6. Le2) 6.Be2 e5 7.Nb3 Be7 8.O-O O-O 9.Be3 Be6 10.Nd5"]
[Time "18:31:52"]
[Variation "Najdorf-Variante, Abweichungen, 6.Be3 e5 7.Nb3 Be6 8.f3 Be7 9.Qd2 Nbd7"]
[WhiteElo "2200"]
[TimeControl "60/3600:60/3600:60/3600"]
[Termination "normal"]
[PlyCount "84"]
[WhiteType "program"]
[BlackType "program"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8.
f3 {+0.31/24 116} Be7 9. Qd2 {+0.15/25 145} h5 10. Nd5 {+0.35/25 152} Nxd5
11. exd5 {+0.26/25 93} Bf5 12. Be2 {+0.30/26 108} Bh4+ 13. g3 {+0.22/27
117} Be7 14. c4 {+0.19/25 112} Nd7 {-0.07/21 132} 15. Na5 {+0.19/25 0} Qc7
16. O-O {+0.15/24 73} O-O 17. b4 {+0.10/23 79} Nf6 18. Rad1 {+0.16/23 83}
Rfd8 19. Rfe1 {+0.31/22 63} Bd7 {0.00/22 179} 20. Bf1 {+0.31/25 21} b6
{+0.06/22 169} 21. Nb3 {+0.30/24 82} a5 {-0.04/22 181} 22. b5 {+0.72/25 73}
Rab8 {-0.06/22 0} 23. Qg2 {+1.02/23 97} Bf5 {+0.09/23 78} 24. h3 {+0.96/23
55} Rdc8 {+0.23/21 92} 25. g4 {+0.87/24 89} Bh7 {-0.03/22 241} 26. Nd2
{+0.90/24 61} Rf8 {-0.21/23 215} 27. Kh1 {+1.01/24 47} hxg4 {-0.01/22 88}
28. hxg4 {+1.01/24 0} Qd8 {0.00/22 84} 29. Be2 {+1.36/24 49} Bg6 {-0.18/22
79} 30. Rg1 {+1.63/25 68} Nd7 {-0.57/23 188} 31. g5 {+1.63/25 0} Nc5
{-0.71/22 118} 32. Qh2 {+1.63/25 0} f5 {-1.00/21 183} 33. gxf6 {+3.07/26
47} Rxf6 {-1.78/23 70} 34. Rg3 {+3.07/26 0} Bf8 {-2.13/23 153} 35. Rdg1
{+4.24/25 52} Bf5 {-2.64/25 176} 36. Bxc5 {+4.24/25 0} bxc5 {-2.64/24 49}
37. Ne4 {+4.24/25 30} Kf7 {-2.64/23 0} 38. Qh5+ {+5.35/30 82} g6 {-2.14/24
56} 39. Qh7+ {+5.33/29 13} Ke8 {-2.50/23 32} 40. Nxf6+ {+5.33/29 51} Qxf6
{-2.47/24 0} 41. Rxg6 {+5.61/27 58} Bxg6 {-3.46/22 44} 42. Rxg6 {+5.61/28
23} Rb7 {-3.39/25 35 Schwarz gibt auf} 1-0
[/pgn]

[mclane]

[pgn][Event "CSTal3 NNUE 5'"]
[Site "ORION8"]
[Date "2022.02.03"]
[Round "1"]
[White "Chess System Tal 3 [5']"]
[Black "DeepShredder13UCIx64 [t1]"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "B12"]
[Opening "Caro-Kann"]
[Time "19:53:00"]
[Variation "Vorstoß-Variante, 4.Nf3 e6"]
[WhiteElo "2200"]
[TimeControl "60/3600:60/3600:60/3600"]
[Termination "normal"]
[PlyCount "96"]
[WhiteType "program"]
[BlackType "program"]

1. e4 c6 2. d4 h6 3. Nf3 d5 4. e5 Bf5 5. Nc3 e6 6. Bd3 Ne7 7. O-O Nd7
{-0.16/27 75} 8. Ne2 {+0.11/23 103} Bxd3 {-0.17/30 75} 9. Qxd3 {+0.14/23
52} c5 {-0.13/28 60} 10. c4 {+0.13/24 148} Rc8 {-0.11/29 71} 11. Rd1
{+0.09/25 130} dxc4 {-0.08/31 24} 12. Qxc4 {+0.09/22 84} Nb6 {-0.13/30 10}
13. Qd3 {+0.17/25 140} c4 {-0.11/31 26} 14. Qe4 {+0.15/25 90} Ned5 {0.00/33
124} 15. h4 {+0.22/25 100} Na4 {+0.22/31 104} 16. Rf1 {+0.11/24 109} Qd7
{+0.30/29 87} 17. h5 {+0.51/24 67} b5 {+0.10/29 158} 18. Qg4 {+0.51/24 0}
a5 {0.00/31 78} 19. Re1 {+0.60/25 102} b4 {+0.11/31 76} 20. b3 {+0.60/25 0}
Nab6 {+0.33/30 58} 21. bxc4 {+0.67/26 71} Nxc4 {-0.06/29 164} 22. Bd2
{+0.67/26 0} a4 {+0.18/31 51} 23. Rec1 {+0.55/25 74} Qb7 {+0.14/31 4} 24.
Rab1 {+0.34/25 94} b3 {+0.06/32 93} 25. axb3 {+0.57/25 72} axb3 {0.00/35 0}
26. Nf4 {+0.19/25 84} b2 {0.00/35 0} 27. Rxc4 {+1.40/27 55} Rxc4 {0.00/38
12} 28. Nxd5 {+1.42/29 67} Qxd5 {0.00/39 14} 29. Rxb2 {+1.34/26 2} Rc8
{0.00/40 149} 30. Rb6 {+1.34/26 0} Rd8 {0.00/37 72} 31. Be3 {+1.40/26 51}
Qa5 {-0.32/32 190} 32. Rb7 {+1.48/27 64} Qd5 {-0.35/35 213} 33. Rc7
{+1.48/27 0} Rh7 {-0.24/34 84} 34. Nh4 {+1.94/26 54} Rd7 {-0.46/31 101} 35.
Rc2 {+1.94/26 0} Rh8 {-0.41/33 156} 36. Kh2 {+1.94/26 0} Rg8 {-0.65/27 70}
37. Bxh6 {+2.77/27 56} Qb7 {-0.72/30 5} 38. Bg5 {+2.73/25 78} Be7 {-1.14/33
176} 39. Bxe7 {+2.73/25 0} Kxe7 {-1.51/31 133} 40. Rc3 {+2.73/25 0} Kf8
{-1.30/29 156} 41. Rg3 {+2.73/25 0} Qb1 {-1.50/31 65} 42. Rf3 {+2.73/27 30}
Qa2 {-1.55/31 5} 43. Qg5 {+2.92/25 130} Qa7 {-1.78/33 75} 44. Rf4 {+3.64/27
90} Qc7 {-1.22/30 75} 45. g3 {+3.57/30 79} Qb7 {-1.29/31 44} 46. Rg4
{+5.02/27 71} Qb2 {-3.12/27 65} 47. Qf4 {+4.98/28 6} Ke7 {-4.03/31 93} 48.
h6 {+6.88/29 86} Rxd4 {-4.03/31 1 Schwarz gibt auf} 1-0[/pgn]

[Chessqueen]

Release of 4 new Rebel's with neural nets created by Chris Whittington playing the famous CSTAL style.

http://rebel13.nl/windows/rebel-14.html#other

https://github.com/egh-s/Rebel-14.1.02- ... stem-TAL-1

How many Elo did Rebel gained due to NNUE ?

[BrendanJNorman]

So I played a game against the "1-minBlitzer" version, which I assumed to be the weakest, and most appropriate to not destroy me in 20 moves, and the game was definitely interesting.

I did as I always do, that is, play over the board with no view of PV or eval (so as to simulate a human opponent), and CSTalRebel was not given an opening book (also the same as a human opponent).

CSTal was set to play 3+2 time control, while I had unlimited time (I used about 20 minutes or so, including time used relaying moves to the physical board)

It seems to me that I completely outplayed him in the early stage of the game, then got into a little bit of trouble on the kingside (playing ...Bxh4 instead of the more direct ....g5), but then after "losing" the exchange I managed to set up a fortress (not an "anti-computer" fortress, this was a normal game here! lol).

If anyone would like to see a video of me analyzing this game, plz let me know.

[pgn][Event "banksia game"]
[Site "?"]
[Date "2022.02.04"]
[Round "?"]
[White "CSTal 1"]
[Black "hm::Norman"]
[Result "*"]
[ECO "B12"]
[PlyCount "207"]
[EventDate "2022.??.??"]

1. d4 c6 2. e4 d5 3. e5 Bf5 4. Ne2 e6 5. Ng3 Bg6 6. h4 h6 7. c3 ({Relevant:} 7.
h5 Bh7 8. Bd3 Bxd3 9. Qxd3 c5 10. c4 cxd4 11. cxd5 Qxd5 12. O-O Nc6 13. Bf4
Nxe5 14. Bxe5 Qxe5 15. Nd2 Nf6 16. Rfe1 Qd5 17. Nc4 Be7 18. Re5 Qd7 19. Rae1
O-O 20. Nf5 Bb4 21. Nxh6+ gxh6 22. R1e4 Nxe4 23. Rxe4 f5 24. Rxd4 Qg7 25. Rd7
Rf7 26. Rxf7 Qxf7 27. Ne5 Qf6 28. f4 Bc5+ 29. Kf1 Qh4 30. Qb3 {0-1 (30)
Matinian,N (2475)-Petrosyan,M (2607) Lichess.org INT 2020}) 7... Nd7 8. Nd2 Ne7
9. Nf3 c5 10. Bd3 Bxd3 11. Qxd3 {[#]} cxd4 $146 ({Predecessor:} 11... Nc6 12.
Bd2 Be7 13. Nh5 Bf8 14. g3 Qc7 15. Nf4 O-O-O 16. Qc2 Kb8 17. Nd3 Ka8 18. O-O
Be7 19. b4 c4 20. Nc5 Rdg8 21. a4 g5 22. h5 g4 23. Nh2 Bg5 24. a5 Bxd2 25. Qxd2
Rg5 26. b5 Ncxe5 27. Nxe6 fxe6 28. dxe5 Qxe5 29. Rfe1 Qf5 30. a6 b6 31. Qf4 Ne5
32. Re3 Rxh5 33. Rxe5 Qxe5 34. Qf7 Qb8 35. Qxh5 e5 36. Nxg4 Re8 37. Re1 d4 38.
cxd4 c3 39. Qg6 Qc8 40. Qc6+ Qxc6 41. bxc6 {Weis,H-Sauter,C Eppingen 2004 1-0})
12. cxd4 Nc6 13. O-O g6 14. Bf4 Qb6 15. Rab1 O-O-O 16. a3 Kb8 17. b4 Be7 18.
Qe3 Bxh4 19. Nxh4 g5 20. Rfd1 Rdg8 21. Kh1 Qd8 22. Nf3 gxf4 23. Qxf4 Qe7 24.
Nh5 Rg6 25. g3 Rhg8 26. Qe3 Ka8 27. Qc3 Qd8 28. Kg2 Nb6 29. Kh3 Nc4 30. b5 Ne7
31. Rg1 Nf5 32. g4 Ne7 33. Nh4 Rg5 34. f4 R5g7 35. Nxg7 Rxg7 36. Rg3 Ng6 37.
Nxg6 fxg6 38. Rf1 h5 39. g5 Rf7 40. Qb4 Rf5 41. Kh4 Qf8 42. Qxf8+ Rxf8 43. Rc3
Kb8 44. Rd3 Kc7 45. Rf2 Kb6 46. Rb3 Rf5 47. Kh3 Rf8 48. Kg2 Na5 49. Rc3 Nc4 50.
Kg3 Kxb5 51. Rb3+ Kc6 52. a4 b6 53. Re2 a6 54. Rb1 Rf5 55. Rf1 Kc7 56. Rfe1 Kc6
57. Rh1 Kc7 58. Rf1 Kc6 59. Rb1 Kc7 60. Rb3 Kc6 61. Rf2 Kc7 62. Kh4 Kc6 63. Rb1
Kc7 64. Rbf1 Kc6 65. Rc1 Kc7 66. Re1 Kc6 67. Ref1 Kc7 68. Rd1 Kc6 69. Re1 Kc7
70. Rc1 Kc6 71. Kg3 Kc7 72. Rcc2 Kc6 73. Ra2 Kc7 74. Kh4 Kc6 75. Rf1 Kc7 76.
Rf3 Kc6 77. Kh3 Kc7 78. Kg3 Kc6 79. Kh4 Kc7 80. Re2 Kc6 81. Kg3 Kc7 82. Rd3 Kc6
83. Rc2 Kc7 84. Rh2 Kc6 85. Re2 Kc7 86. Rf2 Kc6 87. Rg2 Kc7 88. Ra2 Kc6 89. Rc3
Kc7 90. Rd2 Kc6 91. Rdc2 Rf7 92. Ra3 Rf8 93. Ra1 Rf7 94. Re1 Rf8 95. Rg2 Rf7
96. Ra2 Rf8 97. Rf2 Rf5 98. Ra1 Kc7 99. Re2 Kc6 100. Rc1 Rf8 101. Ra2 Rf7 102.
Rf2 Rf5 103. a5 b5 104. Rb1 *

[/pgn]

Here's my setup for those interested. Excuse the mess, kids were crazy today.

[mclane]

There is a nice review by Boban Stanojević about ed and chris engine (net) on eds homepage.



http://rebel13.nl/windows/rebel-14.html#other

[Chessqueen]

Release of 4 new Rebel's with neural nets created by Chris Whittington playing the famous CSTAL style.

http://rebel13.nl/windows/rebel-14.html#other

https://github.com/egh-s/Rebel-14.1.02- ... stem-TAL-1

What about changing Stockfish-code, so that Stockfish can use these interesting new nets? This will not lead to a stronger Stockfish, of course, but perhaps to a very interesting playing (and somewhat weaker) Stockfish - which would be still much, much stronger than Rebel 14.1 (and uses multithreading etc.).

I would love this and testing it for lower draw-rates and shorter games than the default-Stockfish!

Regards - Stefan (SPCC)

And how do you want to find out that stockfish tal plays unsound ?! By testing it against Komodo?

The engine creates art, like a painter does, on fruit level.
Imo on the level of stockfish, this is difficult to measure.


So far chess system tal 3 NNUE plays only one core and only on fruit level.

It is still on a high level and much higher then chess system tal 2 was doing, because search has improved a lot from 1999 to 2004.

No it would be better to send it to TCEC and have them test it here with this group, and it might get lucky to end up in the bottom of the stack or probably in last place https://tcec-chess.com/