TalkChess.com

I have finished a large tournament with new and 'old' versions of many recently progressing engines.
Very impressive results shown by Orion 0.7 NNUE (expected). FabChess 1.16, and Counter 3.6.

i7-8750H-Quad, Windows 10
Arena 3.5.1, TC 20s+3s, 2 cycles, 66 rounds
All engines used 4 CPU (except Francesca, Strelka, and Orion)

01. Orion 0.7 NNUE ---------- 52,5
02. Beef 0.2.2 ---------------- 47,5
03. Beef 0.2 ------------------ 47,0
04. FabChess 1.16 ----------- 46,5
05. Beef 0.3.6 ---------------- 46,5
06. Rodent IV 0.30 ----------- 41,5
07. Counter 3.6 -------------- 41,5
08. FabChess 1.15 ----------- 40,5
09. Tucano 9.00 -------------- 40,5
10. Koivisto 3.0 -------------- 40,0
11. Weiss 1.2 ----------------- 39,5
12. Weiss 1.1 ----------------- 35,0
13. FabChess 1.14.1--------- 35,0
14. Seer 1.1 ------------------ 35,0
15. Fruit Reloaded 15A ----- 35,0
16. Koivisto 2.0 -------------- 34,5
17. Counter 3.4 -------------- 34,0
18. Fruit Reloaded 3.2.1 --- 34,0
19. Counter 3.5 -------------- 33,0
20. Counter 3.3 -------------- 32,0
21. Halogen 8 ---------------- 32,0
22. Weiss 1.0 ----------------- 31,5
23. Mr Bob 0.9.0 ------------- 31,5
24. Halogen-dev-20201106 - 30,5
25. Cheese 2.1 --------------- 25,5
26. Francesca 0.28 ---------- 25,0
27. Cheese 2.2 --------------- 24,5
28. Francesca 0.29a --------- 24,0
29. Seer 1.0 ------------------ 23,5
30. Francesca 0.29 ---------- 23,5
31. Weiss 0.10 ---------------- 19,0
32. Strelka Gambit 2.0.5 --- 15,5
33. Orion 0.7 ----------------- 13,0
34. Francesca 0.25 ---------- 12,0

Cool tournament, thanks for including weiss

Thanks for testing the new version of Fab, very impressed by the results!!!

Terje wrote: ↑Thu Nov 12, 2020 5:12 pm Cool tournament, thanks for including weiss

Weiss shows a nice progress last weeks. We are waiting for new versions

fabianVDW wrote: ↑Thu Nov 12, 2020 6:04 pm Thanks for testing the new version of Fab, very impressed by the results!!!

I will include Fab 1.16 to a new tournament with much stronger opponents such as Minic (non-NNUE), Winter, Combusken, Demolito, Pedone, Lc0-CPU (and may be even stronger: Roc, Defenchess. RubiChess, Nemorino (non-NNUE), Schooner, Booot) etc. That would be a really tough testing for the new Fab and other rapidly progressing engines (such as Koivisto 4.0) at an ''almost elite party''. The tournamed will be called ''Winter is Coming''. I will report the results in approx. one week.

Angle wrote: ↑Thu Nov 12, 2020 2:57 pm (...)
Very impressive results shown by Orion 0.7 NNUE (expected).
(...)
01. Orion 0.7 NNUE ---------- 52,5
(...)
33. Orion 0.7 ----------------- 13,0

Thank you for organizing this tournament, and thank you for your interest in my little engine Orion

Nice, thanks for involving various versions of Francesca. Note that the difference between 0.29 and 0.29a is just a bugfix when pondering is on.

silentshark wrote: ↑Sat Nov 14, 2020 1:31 pm Nice, thanks for involving various versions of Francesca. Note that the difference between 0.29 and 0.29a is just a bugfix when pondering is on.

I know, but in my previous tests 0.29a was weaker than 0.28. I suggested that something was wrong with 'bugfix' and tried 0.29. Unfortunately (and surprisingly), the same tendency in comparison with 0.28

David Carteau wrote: ↑Sat Nov 14, 2020 11:13 am

Angle wrote: ↑Thu Nov 12, 2020 2:57 pm (...)
Very impressive results shown by Orion 0.7 NNUE (expected).
(...)
01. Orion 0.7 NNUE ---------- 52,5
(...)
33. Orion 0.7 ----------------- 13,0

The power of NNUE in action

Thanks for including both versions Seer in the tournament! It's very exciting to see the improvement going from 1.0 to 1.1.