Steam 설치
로그인
|
언어
简体中文(중국어 간체)
繁體中文(중국어 번체)
日本語(일본어)
ไทย(태국어)
Български(불가리아어)
Čeština(체코어)
Dansk(덴마크어)
Deutsch(독일어)
English(영어)
Español - España(스페인어 - 스페인)
Español - Latinoamérica(스페인어 - 중남미)
Ελληνικά(그리스어)
Français(프랑스어)
Italiano(이탈리아어)
Bahasa Indonesia(인도네시아어)
Magyar(헝가리어)
Nederlands(네덜란드어)
Norsk(노르웨이어)
Polski(폴란드어)
Português(포르투갈어 - 포르투갈)
Português - Brasil(포르투갈어 - 브라질)
Română(루마니아어)
Русский(러시아어)
Suomi(핀란드어)
Svenska(스웨덴어)
Türkçe(튀르키예어)
Tiếng Việt(베트남어)
Українська(우크라이나어)
번역 관련 문제 보고
In short, we don't know but I suspect it's an in-house engine. Also unknown what, if any, endgame tables it uses and likewise, its opening book remains a mystery.
If Ripstone have licensed or modified an existing engine it would require credit to be given, I believe so it all leaves us in the dark.
If you do find out anymore, please post because I'd be very interested to hear about it.
I've copied the PGN for the Frtiz 17 vs CU game below, if you want to see how it played out. If you don't have a chess GUI you can copy / paste using Chess.com or Lichess online.
[Event "?"]
[Site "?"]
1969년 12월 31일 오후 4시 00분[Round "?"] [White "Fritz 17"] [Black "Chess Ultra"] [Result "1-0"] [ECO "A07"] [PlyCount "93"] 1. Nf3 Nf6 2. g3 d5 3. Bg2 c6 4. O-O Bf5 5. d3 e6 6. Nbd2 Nbd7 7. Qe1 Nc5 8. Nh4 h6 9. b4 Na4 10. Nxf5 exf5 11. Rb1 Bd6 12. Nf3 O-O 13. Nd4 f4 14. Bxf4 Bxf4 15. gxf4 Qc7 16. Qd2 Rad8 17. e3 c5 18. bxc5 Nxc5 19. Qc3 Qc8 20. a4 a6 21. Rb6 Rfe8 22. Rfb1 h5 23. a5 h4 24. h3 Nfd7 25. Nf5 Ne6 26. Qxc8 Rxc8 27. Rxb7 Ndc5 28. R7b6 Nb3 29. cxb3 Rc5 30. Nd6 Rf8 31. b4 Rc3 32. Bxd5 Rxd3 33. Bc4 Rc3 34. Rxa6 Rb8 35. b5 g6 36. b6 Nd8 37. Ra7 Kh8 38. a6 Nc6 39. Rc7 g5 40. Rxc6 Rg8 41. Nf5 Ra8 42. b7 Rc1+ 43. Rxc1 Rf8 44. Bxf7 Rxf7 45. b8=Q+ Rf8 46. Qxf8+ Kh7 47. Qg7# 1-0
[Round "?"]
[White "Fritz 17"]
[Black "Chess Ultra"]
[Result "1-0"]
[ECO "A07"]
[PlyCount "93"]
1. Nf3 Nf6 2. g3 d5 3. Bg2 c6 4. O-O Bf5 5. d3 e6 6. Nbd2 Nbd7 7. Qe1 Nc5 8.
Nh4 h6 9. b4 Na4 10. Nxf5 exf5 11. Rb1 Bd6 12. Nf3 O-O 13. Nd4 f4 14. Bxf4 Bxf4
15. gxf4 Qc7 16. Qd2 Rad8 17. e3 c5 18. bxc5 Nxc5 19. Qc3 Qc8 20. a4 a6 21. Rb6
Rfe8 22. Rfb1 h5 23. a5 h4 24. h3 Nfd7 25. Nf5 Ne6 26. Qxc8 Rxc8 27. Rxb7 Ndc5
28. R7b6 Nb3 29. cxb3 Rc5 30. Nd6 Rf8 31. b4 Rc3 32. Bxd5 Rxd3 33. Bc4 Rc3 34.
Rxa6 Rb8 35. b5 g6 36. b6 Nd8 37. Ra7 Kh8 38. a6 Nc6 39. Rc7 g5 40. Rxc6 Rg8
41. Nf5 Ra8 42. b7 Rc1+ 43. Rxc1 Rf8 44. Bxf7 Rxf7 45. b8=Q+ Rf8 46. Qxf8+ Kh7
47. Qg7# 1-0
Chess Ultra's AI has 10 different difficulty settings. That doesn't mean it has 10 different engines, but 10 toned down versions of the same engine.
Here is a quote from Simply Chess (another chess game on steam):
"Play against the computer – World class "Stockfish" AI with 100 difficulty levels."
You can indeed make a low rated AI using Stockfish.
This is the standard model employed by every other chess program I can think of, or have used. The difficulty is simply one of many adjustable parameters for a chess engine, such as Stockfish. There is absolutely no need for a separate engine. I have Stockfish (and several others) installed in the Fritz GUI. You tell it where the main binary is located, set the initial parameters and it builds a UCI (universal chess engine) file and that's it. It doesn't spam modified copies of the binary exe. Here's the contents of my Stockfish UCI file:
Name=Stockfish 12
Author=the Stockfish developers (see AUTHORS file)
Filename=C:\Program Files (x86)\Common Files/ChessBase/Engines.uci/Stockfish 12\stockfish_20090216_x64_bmi2.exe
Priority=below normal
[OPTIONS]
UCI_Chess960=true
Threads=6
You can see I've told it to use 6 cores by default but I don't need a 6-core specific engine (or version). It can handle whatever you want, provided the cores are there.
If engines can play at lower strength how can we conclude it's not Stockfish?
The rating of a regular human grandmaster is around 2500, so the grandmaster AI corresponds to that.
What is the point of playing against an uncrippled Stockfish? It has a rating of 3500, no human can beat it.
If you bought a Ferrari, without access to a race track, would you demand it be speed limited?
As to the point of playing against a 3500 opponent, how about as a training tool and for analysis? There are many benefits to playing against or employing a strong engine as a kibitzer. At my level, I can't hope to beat CU on Grandmaster but there will be folks out there who can so why not give them the option for more of a challenge? It doesn't hurt lower rated players.
Having an 'uncrippled' engine means that everyone, regardless of skill level can play against an opponent that suits them. or, as you suggest, the devs deliberately chose an arbitrary cap which to me is odd, in comparison to other chess programs.
Look, you seem utterly convinced and determined to persuade us that CU is Stockfish. I'm doubtful, that's all, so we'll just have to disagree on this. I don't particularly care either way. However, proceeding to school us noobs with your big revelation that engines can play with variable strength (wow, really?) doesn't exactly steer us towards friendly discussion.
There is no mention of third party developers or engine in the game credits which is what really leads me to the simpler conclusion that the engine is in-house. If I'm wrong, so be it, and I'll be interested to know more, as I've already said in an earlier post.
I just checked my local files for another chess game I have on steam: Check vs Mate. It also has no source codes. But it's store page does say it is using a version of the Fritz engine! So a license of Fritz does not require the sharing of the engine in local files.
Here we see that it uses Fritz 10 ... a much older version even though the game is much newer than fritz 10's release date. So maybe Chessbase only licenses old old versions of Fritz so other products can't compete with them: https://en.wikipedia.org/wiki/Battle_vs._Chess
Like CU, CvM has 10 difficulty levels :thinking:
https://www.unrealengine.com/en-US/developer-interviews/ripstone-discusses-the-impact-of-ue4-on-the-development-of-chess-ultra?sessionInvalidated=true
3 devs (1 of whom was an intern) with apparently 0 chess-software experience and a tight deadline :thinking: the plot thickens
Here's the interesting thing though; the handicap mode of Fritz 10 had an adjustable Elo rating of 1375 to 2337 (though the engine could play around 2780 full strength). Thinking 2337 with CU GM around 2400 ... hmmm.
So Fritz 10 could be a potential candidate. I'm no expert on intellectual property / copyright but I would have thought Ripstone would be required to credit Chessbase / Fritz? The Fritz engine has different developers after version 13 and I'm not even sure who the licensing rights now belong to for older versions.
Keep up the detective work and keep us posted!
I think that's it's gotta be it!
To be honest, it's the whole licensing issue that has me puzzled, or lack thereof in the credits. I have no idea what kind of deal Ripstone might have cut with a third party developer and I suspect we'll never know for sure.