TOGA IV with Rebel 14.2 NNUE evaluation

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

Moderator: Ras

User avatar
Rebel
Posts: 7426
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: TOGA IV with Rebel 14.2 NNUE evaluation

Post by Rebel »

Thanks Peter, good to see you posting again also.
90% of coding is debugging, the other 10% is writing bugs.
User avatar
mclane
Posts: 18946
Joined: Thu Mar 09, 2006 6:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub

Re: TOGA IV with Rebel 14.2 NNUE evaluation

Post by mclane »

Welcome back Peter
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
Frank Quisinsky
Posts: 7150
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: TOGA IV with Rebel 14.2 NNUE evaluation

Post by Frank Quisinsky »

Good evening Mr. Ed Schröder,

do you know what you do here with Rebel?

Have a look in the move stats from version 14.1.
https://forum.computerschach.de/cgi-bin ... ?tid=12958

Just genial what Rebel 14.1 do in attacking chess.
Elo goes high with many pieces on board, just very nice for opening analyzes.

Now Rebel 14.2 is still running "run-3".
Wasp will get a competition.

Yesterday I started "run-3".
The Short-Game stat will be a fight between Rebel 14.2 and Wasp 5.30 for place 1.
Maybe Andscacs 0.1 NNSF dev will be in the near.

You made me happy.
One program more I like!

My compliment for the work you do here!!
Please more versions ...

:-)

Go Wasp go ... at first ... sorry!
Go Rebel go ... at second ... should be clear!

For the "short-games-stat".

:-)

Best
Frank
Peter Berger
Posts: 764
Joined: Thu Mar 09, 2006 2:56 pm

Re: TOGA IV with Rebel 14.2 NNUE evaluation

Post by Peter Berger »

I’ve started another match with Toga IV 1.1 at tournament time control on two pretty identical notebooks, this time against Fritz 17.
Threads, hash set to reasonable values. As Toga doesn’t play with an opening book I set Fritz 17’s book options to „normal“.
In the first game Fritz 17 chose a Winawer line that is most probably not completely sound and not new either. So Toga reached an edge that it converted into a rook ending a pawn up - that was not enough for a win though. A pretty boring and obvious game, but this time the general setup clearly feels more like even forces, so it is likely that there will be more interesting games in the future.
Peter
[Event "Lang 120min+10sek"]
[Site "Berlin"]
[Date "2022.03.12"]
[Round "1"]
[White "Toga IV 1.1"]
[Black "Fritz 17"]
[Result "1/2-1/2"]
[ECO "C18"]
[PlyCount "121"]
[TimeControl "7200+10"]

{5120MB, LAPTOP-NCDN8BTK} 1. d4 {[%eval 14,25] [%emt 0:03:39]} e6 {[%emt 0:00:
10]} 2. Nc3 {[%eval 28,25] [%emt 0:04:16]} d5 {[%emt 0:00:08]} 3. e4 {[%eval
21,26] [%emt 0:03:41]} Bb4 {[%emt 0:00:08] (Sf6)} 4. e5 {[%eval 15,24] [%emt 0:
06:20]} c5 {[%emt 0:00:11]} 5. a3 {[%eval 21,26] [%emt 0:03:15]} Ba5 {[%emt 0:
00:09] (Lxc3+)} 6. Qg4 {[%eval 32,24] [%emt 0:05:43]} Ne7 {[%emt 0:00:14]} 7.
dxc5 {[%eval 34,26] [%emt 0:02:54]} Bxc3+ {[%emt 0:00:09]} 8. bxc3 {[%eval 29,
27] [%emt 0:00:52]} Qa5 {[%emt 0:00:08] (0-0)} 9. Bd2 {[%eval 32,28] [%emt 0:
03:52]} Ng6 {[%emt 0:00:10] (0-0)} 10. Nf3 {[%eval 48,25] [%emt 0:04:19]} Nd7 {
[%emt 0:00:08] (Dc7)} 11. c4 {[%eval 85,26] [%emt 0:02:45]} Qa4 {[%emt 0:00:08]
} 12. h4 {[%eval 92,25] [%emt 0:02:21]} O-O {[%emt 0:13:10]} 13. Qd4 {[%eval
109,31] [%emt 0:00:01]} Ne7 {[%emt 0:04:40]} 14. cxd5 {[%eval 104,29] [%emt 0:
00:01]} Qxd4 {[%emt 0:04:44]} 15. Nxd4 {[%eval 92,25] [%emt 0:00:01]} Nxd5 {
[%emt 0:05:42] (exd5)} 16. h5 {[%eval 102,23] [%emt 0:02:20]} h6 {[%emt 0:03:
58] (Sxe5)} 17. Rh3 {[%eval 118,26] [%emt 0:03:14]} Nxe5 {[%emt 0:01:46]} 18.
Rg3 {[%eval 108,24] [%emt 0:00:07]} Kh8 {[%emt 0:04:26] (Sf6)} 19. Be2 {
[%eval 95,24] [%emt 0:04:24]} Nf6 {[%emt 0:00:42] (a6)} 20. Bc3 {[%eval 116,25]
[%emt 0:02:28]} Ne4 {[%emt 0:04:21]} 21. Re3 {[%eval 113,30] [%emt 0:00:01]}
Nxc3 {[%emt 0:03:34]} 22. Rxc3 {[%eval 114,28] [%emt 0:00:01]} Bd7 {[%emt 0:07:
23]} 23. Rd1 {[%eval 132,32] [%emt 0:00:01]} Ba4 {[%emt 0:03:59]} 24. Rb1 {
[%eval 128,31] [%emt 0:00:01]} Nc6 {[%emt 0:02:58]} 25. Nb5 {[%eval 122,29]
[%emt 0:00:01]} Bxb5 {[%emt 0:03:49]} 26. Bxb5 {[%eval 146,28] [%emt 0:00:01]}
Kg8 {[%emt 0:04:11] (Tab8)} 27. Bxc6 {[%eval 170,28] [%emt 0:01:54]} bxc6 {
[%emt 0:00:33]} 28. Rb7 {[%eval 196,28] [%emt 0:01:58]} Rfb8 {[%emt 0:00:41]}
29. Rc7 {[%eval 188,26] [%emt 0:01:05]} Rc8 {[%emt 0:02:16]} 30. Rd7 {[%eval
201,29] [%emt 0:00:01]} Rd8 {[%emt 0:03:47]} 31. Rcd3 {[%eval 190,31] [%emt 0:
00:01]} Rxd7 {[%emt 0:04:11]} 32. Rxd7 {[%eval 168,29] [%emt 0:00:01]} Kf8 {
[%emt 0:01:26]} 33. f3 {[%eval 167,26] [%emt 0:00:23]} Rb8 {[%emt 0:02:22] (a5)
} 34. Rc7 {[%eval 166,27] [%emt 0:01:47]} Rb5 {[%emt 0:00:50]} 35. Rxc6 {
[%eval 140,26] [%emt 0:01:21]} Ra5 {[%emt 0:00:34]} 36. Kd2 {[%eval 124,23]
[%emt 0:01:25]} Rxa3 {[%emt 0:02:58]} 37. Rc8+ {[%eval 113,25] [%emt 0:00:01]}
Ke7 {[%emt 0:00:08]} 38. Rb8 {[%eval 112,28] [%emt 0:01:53]} Ra5 {[%emt 0:01:
54]} 39. Rb7+ {[%eval 105,25] [%emt 0:00:01]} Ke8 {[%emt 0:02:36]} 40. Rc7 {
[%eval 107,26] [%emt 0:00:01]} g6 {[%emt 0:01:06] (Ta3)} 41. hxg6 {[%eval 104,
27] [%emt 0:01:33]} fxg6 {[%emt 0:00:08]} 42. Rg7 {[%eval 97,29] [%emt 0:04:45]
} g5 {[%emt 0:01:14]} 43. Rg6 {[%eval 94,29] [%emt 0:02:36]} Ke7 {[%emt 0:00:
14]} 44. Rxh6 {[%eval 94,33] [%emt 0:01:30]} Rxc5 {[%emt 0:00:08]} 45. Rh7+ {
[%eval 94,34] [%emt 0:02:00]} Kf6 {[%emt 0:00:08]} 46. Rxa7 {[%eval 80,24]
[%emt 0:01:29]} Rd5+ {[%emt 0:00:15]} 47. Ke1 {[%eval 84,24] [%emt 0:01:45]}
Rc5 {[%emt 0:00:57] (Te5+)} 48. Kd1 {[%eval 79,26] [%emt 0:01:16]} g4 {[%emt 0:
00:09] (Td5+)} 49. Ra3 {[%eval 79,33] [%emt 0:01:17]} gxf3 {[%emt 0:01:03]
(Tc4)} 50. Rxf3+ {[%eval 79,34] [%emt 0:01:14]} Ke5 {[%emt 0:00:09] (Ke7)} 51.
Kd2 {[%eval 78,26] [%emt 0:01:12]} Rc7 {[%emt 0:00:45] (Tc4)} 52. Rc3 {[%eval
62,27] [%emt 0:02:55]} Rg7 {[%emt 0:01:11]} 53. g3 {[%eval 61,23] [%emt 0:00:
01]} Ke4 {[%emt 0:00:46]} 54. Re3+ {[%eval 61,20] [%emt 0:00:20]} Kd5 {[%emt 0:
00:29] (Kf5)} 55. Kd3 {[%eval 45,24] [%emt 0:01:39]} Rg8 {[%emt 0:00:08] (Tg4)}
56. Ke2 {[%eval 32,26] [%emt 0:01:37]} Rc8 {[%emt 0:00:08] (e5)} 57. Kd2 {
[%eval 16,24] [%emt 0:01:00]} Rg8 {[%emt 0:00:08] (e5)} 58. Rd3+ {[%eval 16,24]
[%emt 0:01:01]} Ke4 {[%emt 0:01:03]} 59. Ke2 {[%eval 16,22] [%emt 0:00:01]} Rc8
{[%emt 0:00:22]} 60. Kd1 {[%eval 4,25] [%emt 0:01:17]} e5 {[%emt 0:00:10] (Th8)
} 61. Kd2 {[%eval 2,28] [%emt 0:01:27]} 1/2-1/2
User avatar
Rebel
Posts: 7426
Joined: Thu Aug 18, 2011 12:04 pm
Full name: Ed Schröder

Re: TOGA IV with Rebel 14.2 NNUE evaluation

Post by Rebel »

Frank Quisinsky wrote: Fri Mar 11, 2022 10:16 pm Good evening Mr. Ed Schröder,

do you know what you do here with Rebel?

Have a look in the move stats from version 14.1.
https://forum.computerschach.de/cgi-bin ... ?tid=12958

Just genial what Rebel 14.1 do in attacking chess.
Elo goes high with many pieces on board, just very nice for opening analyzes.

Now Rebel 14.2 is still running "run-3".
Wasp will get a competition.

Yesterday I started "run-3".
The Short-Game stat will be a fight between Rebel 14.2 and Wasp 5.30 for place 1.
Maybe Andscacs 0.1 NNSF dev will be in the near.

You made me happy.
One program more I like!

My compliment for the work you do here!!
Please more versions ...

:-)

Go Wasp go ... at first ... sorry!
Go Rebel go ... at second ... should be clear!

For the "short-games-stat".

:-)

Best
Frank
Thanks for all your kind words. NNUE is magic. It has produced the playing style I always wanted. Aggressive, sometimes a bit reckless and it doesn't matter, I am a happy man.
90% of coding is debugging, the other 10% is writing bugs.