The last public ranking of chess centaurs was in 2014, and only compares centaurs to each other. As of 2018, what would be the ELO rating of the best centaurs compared to the best humans and chess engines?

This is similar to this question, but two differences:

  • It is specifically about if centaurs beat computers, which is a subset of this question
  • The top answer is about a GM with a computer who does not have any training as a centaur, while we know that good centaurs have very different skills than good humans do.
  • This post in talkchess site doesn't answer exactly your questions, but probably can read some interesting ideas here
    – emdio
    Commented Jan 19, 2022 at 21:20

2 Answers 2


Chess centaurs is not correspondence chess because centaur matches are played under the classical time controls. Unfortunately, centaurs is dead because the best strategy is just follow the engine moves.

In classical time controls, there's generally no advantage for human assistance. Humans are too slow, too clumsy etc ... we need more time.

This is 2018, centaurs matches are absolutely meaningless. They are like computer matches with dumb human intervention. Computer engines play stronger than human+computer unless the human player has sufficient time to digest the positions.

The answer to your question is the same ELO as the engine itself.

  • 1
    Do you have some evidence for what you wrote? I'm under the impression that a GM + computer should still hold a meaningful edge over just the computer.
    – Allure
    Commented Mar 26, 2018 at 22:29
  • @Allure your are wrong. Google about Garry Kasparov.
    – SmallChess
    Commented Mar 26, 2018 at 23:10
  • 2
    "Garry Kasparov" returns a huge list of results, not all of which is related to centaur chess. "Garry Kasparov centaur" doesn't find anything conclusive. Do you have more specific search terms?
    – Allure
    Commented Mar 26, 2018 at 23:14
  • @Allure Kasparov mentioned when he talked about advance chess. Sorry I don’t have time to google it further.
    – SmallChess
    Commented Mar 26, 2018 at 23:15
  • 1
    @SmallChess Have you had a chance yet? I'm not sure why your answer was upvoted since as yet it's incomplete. Commented Jan 16, 2022 at 21:26

In this 2015 article Magnus Carlsen agrees with the interviewer's claim that the best computers surpass human playing ability:

Interviewer: Both chess - and more recently the Asian game of Go - are now played better by computers than by humans, even by the very best like you. Does that take away the magic of brain games like chess?

Carlsen: Yes, to some extent it takes a bit of the mysticism away. But as for chess, we've known for a long time that computers are better, so the computer never has been an opponent. It's a tool to help me analyze and to help me improve at chess.

However, there are rare times when a human will play a move over the board that the engines being used by analysts in real time do not find. For example, in this video analysis of Kramnik v Leko 2004 WCC Game 8, the commentator's engine fails to find Leko's brilliant move 25 ... Qd3.

[FEN ""]
[Event "2004 World Championship Match Game 8"]
[White "Vladimir Kramnik"]
[Black "Peter Leko"]

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 O-O 8. c3 d5 9. exd5 Nxd5 10. Nxe5 Nxe5 11. Rxe5 c6 12. d4 Bd6 13. Re1 Qh4 14. g3 Qh3 15. Re4 g5 16. Qf1 Qh5 17. Nd2 Bf5 18. f3 Nf6 19. Re1 Rae8 20. Rxe8 Rxe8 21. a4 Qg6 22. axb5 Bd3 23. Qf2 Re2 24. Qxe2 Bxe2 25. bxa6 Qd3 26. Kf2 Bxf3 27. Nxf3 Ne4+ 28. Ke1 Nxc3 29. bxc3 Qxc3+ 30. Kf2 Qxa1 31. a7 h6 32. h4 g4

This fact suggests that perhaps a world-class human player could occasionally find moves that improve the play of one of the best computers. But it seems that for the most part, the human would go against the recommendation of the computer at great risk. A centaur match today would probably be quite boring.

  • 1
    I'm sorry, but that example does not hold up. My Stockfish sees 25...Qd3 in less than a second, Leela being quite a lot slower maybe due to my slow hardware, comes in still in less than 10 seconds.
    – koedem
    Commented Jan 16, 2022 at 21:11
  • The analysis video was from 2018, so I guess it's no surprise that now in 2022 the move is seen. As time passes, surely the potential for a human to outplay the best computers even for a single move will continue to shrink.
    – wberry
    Commented Jan 16, 2022 at 21:13
  • Hm, I tried Stockfish 9 now, which came out on Feb 1st 2018. It also saw it in less than a second...
    – koedem
    Commented Jan 16, 2022 at 21:27
  • 1
    That's curious. In the first part of the video agadmator says specifically that he tried it with Stockfish 9 and the move was "invisible" to it. I wonder what could account for that difference.
    – wberry
    Commented Jan 16, 2022 at 22:32
  • 1
    Clickbait? I don't know. My processor is from 2013 too so I don't think it's the hardware that would make the difference either.
    – koedem
    Commented Jan 16, 2022 at 22:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.