Why release Revolution 3.40-081125?
Our automated testing pipeline (“Codex”) has formally green-lit Revolution 3.40-081125 after an extended SPRT A/B match against the current baseline (Revolution 3.10-300925). The new build shows a consistent, statistically supported gain at a practical rapid time control, with no evidence of instability or regression.
Over 1,500 games at 10+0.1, 2 threads, 32MB hash, using the UHO_Lichess_4852_v1.epd suite, the DEV build achieved:
- Elo delta: +5.33 ± 9.11 (nElo: +10.29 ± 17.58) in favour of the new DEV
- Score: 50.77 % (394 wins / 371 losses / 735 draws)
- LOS (Likelihood of Superiority): 87.43 % vs baseline
- Draw ratio: 50.27 %
- PairsRatio: 1.13 (effectively pair-clean, good colour symmetry)
- Pentanomial (Ptnml(0-2)): [5, 170, 377, 193, 5], WL/DD ratio: 1.03
- SPRT: LLR 0.28 (9.5 %) in the acceptance band [0.00, 2.50]
In plain terms: Revolution 3.40-081125 is reliably stronger than the previous stable release under realistic tournament conditions, and the match conditions (pairing, colour balance, openings) were sufficiently clean to trust the result. Given Codex’s OK, the SPRT outcome, and the absence of pathological behaviour in the accompanying PGN and logs, we consider this build a safe and justified upgrade to publish as the next official release.
Hello, have a look to 55. .. -Qe4?? what a Blunder from Revolution 3.40-081125 with 19.3 sec rest on the clock ! any idea?
feel free to ask me about my settings….
Best regards, Thomas
Event “rated bullet game”]
[Site “https://lichess.org/hjGbchYR”]
[Date “2025.11.11”]
[White “t”]
[Black “]
[Result “1-0”]
[GameId “hjGbchYR”]
[UTCDate “2025.11.11”]
[UTCTime “20:39:42”]
[WhiteElo “3149”]
[BlackElo “3139”]
[WhiteRatingDiff “+5”]
[BlackRatingDiff “-5”]
[WhiteTitle “BOT”]
[BlackTitle “BOT”]
[Variant “Standard”]
[TimeControl “30+0”]
[ECO “E08”]
[Opening “Catalan Opening: Closed”]
[Termination “Normal”]
[Annotator “lichess.org”]
1. Nf3 d5 2. d4 Nf6 3. c4 e6 4. g3 Bb4+ 5. Bd2 Be7 6. Bg2 O-O 7. O-O c6 8. Qc2 Nbd7 9. Bf4 b6 10. Nbd2 Bb7 11. e4 { E08 Catalan Opening: Closed } dxc4 12. Nxc4 c5 13. Rfe1 cxd4 14. Nd6 d3 15. Qxd3 Bxd6 16. Bxd6 Bxe4 17. Qd4 e5 18. Qa4 Re8 19. Rad1 Bxf3 20. Bxf3 e4 21. Bxe4 Nxe4 22. Rxe4 Nf6 23. Rxe8+ Qxe8 24. Qxe8+ Rxe8 25. Bb4 a5 26. Bc3 Kf8 27. Bxf6 gxf6 28. Rd6 b5 29. Kf1 b4 30. Rd5 Re5 31. Rxe5 fxe5 32. Ke2 Ke7 33. Kd3 Ke6 34. Kc4 e4 35. Kd4 f5 36. h3 a4 37. b3 axb3 38. axb3 h6 39. h4 h5 40. Kc4 Ke5 41. Kxb4 Kd4 42. Ka4 Kd3 43. b4 Ke2 44. b5 Kxf2 45. b6 e3 46. b7 e2 47. b8=Q e1=Q 48. Qf4+ Kg2 49. Qxf5 Kxg3 50. Qxh5 Qe4+ 51. Kb5 Qxh4 52. Qe8 Qg5+ 53. Kc6 Qf6+ 54. Kd7 Qd4+ 55. Kc8 Qe4 56. Qxe4 Kh2 57. Kd7 Kg1 58. Ke6 Kf1 59. Kf5 Kf2 60. Kg4 Kg1 61. Kg3 Kf1 62. Qe8 Kg1 63. Qe1# { White wins by checkmate. } 1-0
On lichess, you can play through Brotli or Banksia. In those environments, you have to configure the chess engine correctly. It’s not like on playchess. It also depends on whether or not you use a VPN, etc. If you tell me which environment you play in, I might be able to help you. But I’ve tried this version on playchess and fastchess without any blunders.
thank you in advance,
Jorge
Thank you for the quick response, I play at Lichess with BanksiaGui-0.58-win64 over the public network via Wi-Fi.
This Game was playing with 30 Threats, 64 MB Hash and 3-4-5-pieses-Syzygy…I had Move Overhead on 100, is that the mistake? Ponder on, 2 .bin books loaded and ShowWDL on, otherwise I didn’t change anything in the settings
Best regards, Thomas
Look at the game in Fritz 20’s guide; the engine is no longer playing, it’s in the six-piece endgame. The error must be in the configuration, but it’s not the engine.
https://ibb.co/Sw0N8sdg