Which chess engine is best?

As the blog poll predicted two weeks ago, Magnus Carlsen won the London Classic. With three rounds to go, Magnus had drawn all his games and I was doubting the wisdom of crowds, but a 2½/3 finish and a playoff win proved you right.

Last week’s question was: How do you meet 1.d4? To my surprise, the Grünfeld edged out the KID, Nimzo and Slav. Are there really more Grünfeld than KID players among our readers, or are the Grünfelders just more fanatical about clicking the vote button? Who knows.


So far this has been John writing, but now Jacob takes over with this week’s poll question: Which chess engine is best?

Our friends at Komodo are able to celebrate another great victory. There are a lot of good engines out there at the moment, with Stockfish and Johnny being the main ones, having surpassed the likes of Houdini and Rybka a long time ago (Rybka and Fritz15 are now one and the same as far as I can understand. I do not know how strong it is).

Komodo beat Stockfish by 53.5-46.5 or 8-1 if you like, in a 100-game match. You can read more here.

Does this actually answer our question? Nope! But I hope someone will do so in the comments section below…

25 thoughts on “Which chess engine is best?”

  1. The best chess engine in my opinion is The King, used in Chessmaster. Because of its configurable personalities it was great fun to play against it.

    It would be great if there was an UCI compatible version for smartphones.

  2. pure engine strength-Komodo is best, with SF a close second…all the others (commercial) have nice features, I like Fritz, but which is the best engine I don’t think there is any doubt.

  3. I lost interest in this question when the engines got stronger than any human being.
    I struggle to why I should download -and certainly not buy the latest engines…

    It may matter to the top 10 guys though.

  4. @Multeplukker
    I did notice a difference when Rybka came out: Rybka’s evaluations were much closer to my own judgment than earlier engines’s. But then Houdini followed evaluating everything with 0.00, which made it relatively useless to me. And now Stockfish seems to have similarly overblown evals as the early engines, but still better than 0.00.
    But yeah, 100 Elo more or less doesn’t really matter as far as I’m concerned.

  5. @Mark that’s very understated of you.

    Mark, together with Larry Kaufman, programs Komodo!

    @Gerhard, try playing the strongest programs. Set them to play at a ‘fixed search depth’. Start with depth of 1 ply. If 1 ply is too easy, try a game at 2 ply, etc.

  6. Why isn’t it settled, at least at this moment in time? One might root for Stockfish because it’s free and open source, or for some other engine because of the bells & whistles that might be thrown in with the purchase. That’s all well and good. If it’s a question of strength, however then it’s hard to see how Komodo 9.3’s convincing win in the latest TCEC didn’t settle the issue. It’s not as if Stockfish had a bad night’s sleep or fought with its spouse before the match.

    What’s next, a post asking whether King Arthur “really” defeated the Black Knight?

  7. For playing strength, there can be only one. But if we’re talking about analysis – which is what most of us use the engines for – surely we should be using more than one.

    I use Komodo as my primary engine but I often switch to Stockfish for a ‘second opinion.’

  8. I like to use multi-line mode (where I can see the 4 or 5 best lines in the current position). Sometimes it’s really interesting to see a surprising jump in eval between, say, apparently mundane moves #1 and #2 and apparently mundane moves #3 and onwards.

    Apparently this is something that really hurts some engines’ strength. Does anybody know by how much, and which engines are good at this?

  9. Daniel :
    Stockfish is easily superior to any other on Earth. It’s the best and nothing comes close to it.

    Quite a bold and rather unfounded claim given recent events…

  10. It’s a question of foresight. Homo Sapien is the greatest animal on earth. 100 million years ago it was but a rat-like creature. However, the foundation for greatness lay dormant within this small and vulnerable creature. A meteor is coming and Stockfish’s predilection for straightforward tactical play will carry it to victory. So what if dinosaurs ruled for millions of years? We are the kings of this planet. Likewise, Stockfish will become king and rule all those who might defy it.

  11. I’ve voted for Stockfish, but…. The question is very dificult and it all depends on the position type. Also calc speed is not quite comparable between Komo and Stock (Stock beeing kind of twice faster on open/tactical positions). Also if on same hardware with same setting if u run them for a fixed time, I bet Stock will go deeper and will provide better variations.

  12. @Thomas
    What about a Chimpansee throwing darts on a chessboard to determine the next move? I’ve heard it’s a good method to predict the movement of share prices.

  13. There is and can be no doubt: the new Komodo easily smashed all my other engines when given many many games and complex positions. Having said that it is sometimes worse in the very endgame whereas Stockfish 6 is tactically superior. I use them both in tandem although as I said Komodo has blown all my other engines away over and again. I also use Deep Junior and Houdini to come up with creative moves as well that may easily take an opponent down an unforeseen path!

  14. @Ray
    I always got into trouble with the organizers when I tried to bring my chimp to the board.
    And my opponents where scared by the darts.
    So I left the ape at home and still pretend to think over my move.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top