| Folgende 2 Benutzer sagen Danke zu udo für den nützlichen Beitrag: | ||
applechess (30.01.2026), Mapi (30.01.2026) | ||
|
|||||||||||
|
AW: Re: AW: ChessSystemTal2-v21 unbesiegbar?
Sorry to (maybe) make you sad, but there is no way a Mephisto II can ever beat current Stockfish version.
Maybe you know there is an UCI compatible Rebel Century program? Well, more accurately, a Winboard version one can use with Wb2UCI adapter. What matters is: it can run full speed on any modern CPU. An Ed Schröder gift to the chess community. A large match (1000 games) has been performed: the year 2000 Rebel Century (so muuuuuuch stronger than the Mephisto II) on fast hardware facing Stockfish with 1000:1 time odds. Rebel Century: 30 minutes for the game + 30 seconds increment/move Stockfish: 1.8 seconds for the game + 0.03 seconds increment/move The result: Stockfish won 896, drew 96, lost 8. +442 Elo performance for SF. source Set any time odds you want in favor of Mephisto II, guess the result... MfG, Tibono
__________________
Mein Profil beim ICCF (International Correspondence Chess Federation) https://www.iccf.com/player?id=89948&tab=3 |
| Folgende 2 Benutzer sagen Danke zu Hartmut für den nützlichen Beitrag: | ||
applechess (30.01.2026), spacious_mind (30.01.2026) | ||
|
||||||||||||
|
Re: AW: Re: AW: ChessSystemTal2-v21 unbesiegbar?
You are right the last Stockfish version that allowed you to turn it off was 15.1
|
|
|||||||||||
|
AW: Re: AW: Re: AW: ChessSystemTal2-v21 unbesiegbar?
Code:
Stockfish 16 by the Stockfish developers (see AUTHORS file) uci id name Stockfish 16 id author the Stockfish developers (see AUTHORS file) option name Debug Log File type string default option name Threads type spin default 1 min 1 max 1024 option name Hash type spin default 16 min 1 max 33554432 option name Clear Hash type button option name Ponder type check default false option name MultiPV type spin default 1 min 1 max 500 option name Skill Level type spin default 20 min 0 max 20 option name Move Overhead type spin default 10 min 0 max 5000 option name Slow Mover type spin default 100 min 10 max 1000 option name nodestime type spin default 0 min 0 max 10000 option name UCI_Chess960 type check default false option name UCI_AnalyseMode type check default false option name UCI_LimitStrength type check default false option name UCI_Elo type spin default 1320 min 1320 max 3190 option name UCI_ShowWDL type check default false option name SyzygyPath type string default Code:
Stockfish 16.1 by the Stockfish developers (see AUTHORS file) uci id name Stockfish 16.1 id author the Stockfish developers (see AUTHORS file) option name Debug Log File type string default option name Threads type spin default 1 min 1 max 1024 option name Hash type spin default 16 min 1 max 33554432 option name Clear Hash type button option name Ponder type check default false option name MultiPV type spin default 1 min 1 max 256 option name Skill Level type spin default 20 min 0 max 20 option name Move Overhead type spin default 10 min 0 max 5000 option name nodestime type spin default 0 min 0 max 10000 option name UCI_Chess960 type check default false option name UCI_LimitStrength type check default false option name UCI_Elo type spin default 1320 min 1320 max 3190 option name UCI_ShowWDL type check default false option name SyzygyPath type string default |
| Folgende 2 Benutzer sagen Danke zu Tibono für den nützlichen Beitrag: | ||
Major Chaos (30.01.2026), spacious_mind (30.01.2026) | ||
|
|||||||||||
|
AW: Re: AW: ChessSystemTal2-v21 unbesiegbar?
this test aimed at showing what the evolution of software has been across 25 years, 2000 to 2025. Hence SF 17 during this test was the latest dev version (nov. 25). Indeed using NNUE. Castrating SF of NNUE would have been irrelevant. NNUE is not a problem, it is current software technology. And shy attempt to fairness was the 1000:1 time odds in favor of Rebel Century. Unfair? Nope. Yet definitely not enough for the year 2000 Rebel to survive the match. Hence a clear illustration for the breakthrough chess software achieved - NNUE being a large step forward, I agree. By the way, reading that you expected more information, there is a link in my above post. Maybe you missed it. (look for: 'source'). MfG, Eric |
|
||||||||||||
|
Re: AW: Re: AW: Re: AW: ChessSystemTal2-v21 unbesiegbar?
Rather v16...
Code:
Stockfish 16 by the Stockfish developers (see AUTHORS file) uci id name Stockfish 16 id author the Stockfish developers (see AUTHORS file) option name Debug Log File type string default option name Threads type spin default 1 min 1 max 1024 option name Hash type spin default 16 min 1 max 33554432 option name Clear Hash type button option name Ponder type check default false option name MultiPV type spin default 1 min 1 max 500 option name Skill Level type spin default 20 min 0 max 20 option name Move Overhead type spin default 10 min 0 max 5000 option name Slow Mover type spin default 100 min 10 max 1000 option name nodestime type spin default 0 min 0 max 10000 option name UCI_Chess960 type check default false option name UCI_AnalyseMode type check default false option name UCI_LimitStrength type check default false option name UCI_Elo type spin default 1320 min 1320 max 3190 option name UCI_ShowWDL type check default false option name SyzygyPath type string default Code:
Stockfish 16.1 by the Stockfish developers (see AUTHORS file) uci id name Stockfish 16.1 id author the Stockfish developers (see AUTHORS file) option name Debug Log File type string default option name Threads type spin default 1 min 1 max 1024 option name Hash type spin default 16 min 1 max 33554432 option name Clear Hash type button option name Ponder type check default false option name MultiPV type spin default 1 min 1 max 256 option name Skill Level type spin default 20 min 0 max 20 option name Move Overhead type spin default 10 min 0 max 5000 option name nodestime type spin default 0 min 0 max 10000 option name UCI_Chess960 type check default false option name UCI_LimitStrength type check default false option name UCI_Elo type spin default 1320 min 1320 max 3190 option name UCI_ShowWDL type check default false option name SyzygyPath type string default Stockfish 15.1 and earlier versions generally allowed disabling NNUE (Efficiently Updatable Neural Networks) via the Use NNUE UCI option, returning to the "classical" evaluation function. Stockfish 16 (released June 2023) removed the handcrafted classical evaluation entirely, making NNUE mandatory and disabling the ability to switch it off. Regards |
|
|||||||||||
|
AW: ChessSystemTal2-v21 unbesiegbar?
That's indeed what ChatGPT says. Did you ever heard about AI hallucinations?
I better trust: 1) the UCI options list 2) the fact I keep using SF16 with NNUE disabled for some analysis matters, aside to SF17. And the difference on the field is clear, whether enabled or disabled. 3) check here. Someone posted an "issue" about disabling NNUE with SF16, but stood corrected by an official maintainer from the SF team. The requester was using a DEV version, not the official SF16. Dev version led to SF16.1, where the disable NNUE was removed, as I stated in my previous post. As a given, I am always right, Nick, don't you know that? ![]() mit freundlichen Grußen, Eric |
| Folgender Benutzer sagt Danke zu Tibono für den nützlichen Beitrag: | ||
spacious_mind (30.01.2026) | ||
|
||||||||||||
|
That's indeed what ChatGPT says. Did you ever heard about AI hallucinations?
I better trust: 1) the UCI options list 2) the fact I keep using SF16 with NNUE disabled for some analysis matters, aside to SF17. And the difference on the field is clear, whether enabled or disabled. 3) check here. Someone posted an "issue" about disabling NNUE with SF16, but stood corrected by an official maintainer from the SF team. The requester was using a DEV version, not the official SF16. Dev version led to SF16.1, where the disable NNUE was removed, as I stated in my previous post. As a given, I am always right, Nick, don't you know that? ![]() mit freundlichen Grußen, Eric ![]() |
| Folgender Benutzer sagt Danke zu spacious_mind für den nützlichen Beitrag: | ||
Tibono (30.01.2026) | ||
|
||||||||||||
|
Re: ChessSystemTal2-v21 unbesiegbar?
Hi Eric
Just for your info, I also did a search using Bing AI (not just Google). The information comes from Chessprogramming Wiki: Hybrid In August 2020 a new patch changed Stockfish NNUE into a hybrid engine: it uses NNUE evaluation only on quite balanced material positions, otherwise uses the classical one. It could speed up to 10% and gain 20 Elo [39]. At that point, NNUE helped to increase already around 100 Elo for Stockfish. In the same month, Stockfish changed the default mode of using evaluation functions from classic to hybrid one, the last step to completely accept NNUE. With Stockfish 16 release in 2023 the hand-crafted evaluation function was removed, a complete transition to NNUE based evaluation was made. This is where the search engines AI's get their information from. Regards |
| Folgender Benutzer sagt Danke zu spacious_mind für den nützlichen Beitrag: | ||
Tibono (30.01.2026) | ||
|
||||||||||||
|
Re: ChessSystemTal2-v21 unbesiegbar?
Da ich auf diesem hohen Spielniveau viele Unentschieden beobachte, wollte ich herausfinden, ob dies normal ist. Da es in diesem Zusammenhang um das Schachprogramm Tal 2 geht, habe ich es als Beispiel gewählt und die 60 top Gegner aus der 40/15-Elo-Rangliste von CCRL herangezogen.
Seine Wertung bei CCRL für 40/15 beträgt 3583 ELO. Siehe Tabelle unten: Code:
~~~~~~~~~CHESS PROGRAM~~~~~~~~~ ~ELO~ ~~~SCORE~~~ ~Won~ ~Lost~ ~Draw~ ~Total~ ~PERF.~ – **Stockfish 17.1 64-bit 4CPU 3645 12.5*−*17.5 0 5 25 30 41.70% – **PlentyChess 7.0.0 64-bit 4CPU 3643 11*−*19 0 8 22 30 36.70% – **Stockfish 16 64-bit 4CPU 3640 12*−*17 0 5 24 29 41.40% – **Stockfish 16.1 64-bit 4CPU 3638 26*−*36 0 10 52 62 41.90% – **Torch v4 64-bit 4CPU 3638 12.5*−*17.5 0 5 25 30 41.70% – **ShashChess GZ 64-bit 4CPU 3636 12.5*−*19.5 0 7 25 32 39.10% – **Torch v3 64-bit 4CPU 3635 12.5*−*17.5 0 5 25 30 41.70% – **Obsidian 16.0 64-bit 4CPU 3635 12*−*18 0 6 24 30 40.00% – **Torch v2 64-bit 4CPU 3632 16.5*−*19.5 0 3 33 36 45.80% – **Torch v1 64-bit 4CPU 3630 17.5*−*26.5 0 9 35 44 39.80% – **Stockfish 15.1 64-bit 4CPU 3629 13.5*−*18.5 0 5 27 32 42.20% – **Dragon by Komodo 3.3 64-bit 4CPU 3626 25.5*−*36.5 0 11 51 62 41.10% – **Obsidian 15.0 64-bit 4CPU 3625 13.5*−*16.5 1 4 25 30 45.00% – **Dragon by Komodo 3.2 64-bit 4CPU 3624 13*−*19 0 6 26 32 40.60% – **Obsidian 14.0 64-bit 4CPU 3621 13*−*17 0 4 26 30 43.30% – **PlentyChess 5.0.0 64-bit 4CPU 3621 13*−*17 1 1 24 26 43.30% – **PlentyChess 4.0.1 64-bit 4CPU 3618 13*−*17 0 4 26 30 43.30% – **Berserk 13 64-bit 4CPU 3617 32*−*33 0 1 64 65 49.20% – **Integral v7 64-bit 4CPU 3617 14*−*16 0 2 28 30 46.70% – **Alexandria 8.0.0 64-bit 4CPU 3617 13*−*17 0 4 26 30 43.30% – **Caissa 1.21 64-bit 4CPU 3615 13.5*−*16.5 0 3 27 30 45.00% – **Horsie 1.1 64-bit 4CPU 3615 12.5*−*17.5 0 5 25 30 41.70% – **Stormphrax 7.0.0 64-bit 4CPU 3613 13*−*17 0 4 26 30 43.30% – **PlentyChess 3.0.1 64-bit 4CPU 3611 13.5*−*16.5 1 4 25 30 45.00% – **Ginkgo 5.40 64-bit 4CPU 3609 13*−*17 0 4 26 30 43.30% – **Obsidian 12.0 64-bit 4CPU 3609 15.5*−*20.5 0 5 31 36 43.10% – **Alexandria 7.1.0 64-bit 4CPU 3609 13.5*−*16.5 0 3 27 30 45.00% – **Berserk 12 64-bit 4CPU 3608 27*−*32 2 7 50 59 45.80% – **Fritz 20 64-bit 4CPU 3608 14*−*16 0 2 28 30 46.70% – **Halogen 15 64-bit 4CPU 3607 15*−*15 0 0 30 30 50.00% – **Viridithas 17.0.0 64-bit 4CPU 3606 15*−*15 1 1 28 30 50.00% – **Starzix 6.0 64-bit 4CPU 3606 14*−*16 0 2 28 30 46.70% – **Horsie 1.0 64-bit 4CPU 3605 14*−*16 0 2 28 30 46.70% – **PlentyChess 1.0.0 64-bit 4CPU 3604 16.5*−*19.5 1 4 31 36 45.80% – **Integral v5 64-bit 4CPU 3604 14*−*16 0 2 28 30 46.70% – **Alexandria 7.0.0 64-bit 4CPU 3603 15*−*15 1 1 28 30 50.00% – **Uralochka 3.42a 64-bit 4CPU 3602 14*−*16 0 2 28 30 46.70% – **RubiChess 20240112 64-bit 4CPU 3601 16*−*19 0 3 32 35 45.70% – **Caissa 1.17 64-bit 4CPU 3600 18*−*24 0 6 36 42 42.90% – **Clover 8.0 64-bit 4CPU 3600 14.5*−*15.5 1 2 27 30 48.30% – **Caissa 1.18 64-bit 4CPU 3600 18*−*18 0 0 36 36 50.00% – **PlentyChess 2.0.0 64-bit 4CPU 3600 14.5*−*15.5 1 2 27 30 48.30% – **Ethereal 14.25 64-bit 4CPU 3599 24*−*27 0 3 48 51 47.10% – **Obsidian 11.0 64-bit 4CPU 3597 17.5*−*18.5 1 2 33 36 48.60% – **Obsidian 10.0 64-bit 4CPU 3597 15*−*17 0 2 30 32 46.90% – **Halogen 14 64-bit 4CPU 3596 14*−*14 1 1 26 28 50.00% – **Caissa 1.16 64-bit 4CPU 3595 14.5*−*17.5 2 5 25 32 45.30% – **Caissa 1.15 64-bit 4CPU 3592 16*−*16 0 0 32 32 50.00% – **Titan 1.1.0 64-bit 4CPU 3589 14.5*−*15.5 0 1 29 30 48.30% – **RubiChess 20230918 64-bit 4CPU 3588 16*−*16 2 2 28 32 50.00% – **Uralochka 3.41a 64-bit 4CPU 3585 16*−*16 1 1 30 32 50.00% – **Seer 2.8.0 64-bit 4CPU 3585 17*−*19 1 3 32 36 47.20% – **Alexandria 6.1.0 64-bit 4CPU 3584 18*−*18 1 1 34 36 50.00% – **Berserk 11.1 64-bit 4CPU 3584 16*−*16 2 2 28 32 50.00% – **Caissa 1.14.1 64-bit 4CPU 3583 16*−*16 0 0 32 32 50.00% – **Lizard 10.4 64-bit 4CPU 3582 14.5*−*15.5 0 1 29 30 48.30% – **Clover 6.2 64-bit 4CPU 3581 15.5*−*14.5 2 1 27 30 51.70% – **Obsidian 9.0 64-bit 4CPU 3581 15.5*−*16.5 0 1 31 32 48.40% – **Stormphrax 6.0.0 64-bit 4CPU 3581 9*−*9 0 0 18 18 50.00% – **Rebel 16.3 64-bit 4CPU 3579 18*−*18 0 0 36 36 50.00% TOTAL AV. 23 195 1813 2031 45.77% PERFORMANCE 3608 1.13% 9.60% 89.3% WON GAMES TOTAL % 10.7% Wenn CSTal2 beispielsweise Ihr Lieblingsprogramm ist, wie viele von Ihnen würden dann 100 Spiele spielen wollen, nur um im Durchschnitt einen einzigen CSTal2-Sieg zu erzielen, oder 23 Siege nach 2031 Spielen? ![]() Selbst mit Stockfish 17.1 werden Sie gegen CSTal2 nur etwa in jedem sechsten Spiel einen Sieg sehen. Tatsächlich sieht man, wenn man alle Spiele zusammennimmt, im Durchschnitt nur einen Sieg alle 10 Spiele, wenn man Glück hat und es vom Gegner abhängt. Ich bin so froh, dass es noch dedizierte Schachbrettcomputer gibt! ![]() Der arme CSTal2, nur ein Sieg pro 100 Spielen, was für ein Pech! ![]() Gruss Nick |
![]() |
| Themen-Optionen | |
| Ansicht | |
|
|