09 February 2015

Houdini, Komodo, Stockfish

In a recent post, Overcoming the Operators, I noted how difficult it can be to play a correspondence game against an opponent who is simply relaying the best move suggested by his engine.

What techniques, if any, can be used against the operators? Over the next weeks I'll devote a few posts to look at this question.

A few years ago, in Sources of Inspiration, I flagged the blog 'Tartajubow On Chess' as a source for ideas on both correspondence play and engines. Last month, in Even Engines Need Time to Think, Tartajubow listed some telltale signs of an operator.

33 running games. That's how many my opponent had going and quite often he replied to my move within a few minutes so he was effectively playing at a blitz pace in most of his games. He eventually got into a position where even the engine needed some time to sort things out and 2-3 minutes simply were not enough.

He went on to discuss the three most popular engines in use today.

Houdini, Komodo and Stockfish differ in their search and evaluation functions. Stockfish aggressively prunes the analysis tree and while it searches deeply, its search is more narrow than Komodo or Houdini. Sometimes this is an advantage, sometimes it's not. Komodo is the most positionally accurate of the three and it is slower in its search than either Stockfish or Houdini. It's initial evaluation can be more accurate, but it's usually a good idea to let it think a little longer just to make sure.

Since about a year ago, I'm also using those three engines in practical play. Thanks to a series of blog posts from two years ago -- Engine Evaluation through a summary Practical Evaluation -- I understand how the engines evaluate and am always looking for new approaches to understand better their internal mechanisms. The 'Evaluation' series relied on work done by GM Larry Kaufman, and I was happy to find a subsequent interview where he discussed their strengths: Q&A with Larry Kaufman (qualitychess.co.uk; December 2013).

Q: What are the different properties of Komodo, Stockfish, Houdini, Rybka and Fritz and how would the person who wants to improve their chess use these differences to his advantage?

A: Komodo is best at evaluating middlegame positions accurately once the tactics are resolved. Stockfish seems to be best in the endgame and in seeing very deep tactics. Houdini is the best at blitz and at seeing tactics quickly. Rybka is just obsolete; I like to think of Komodo as its spiritual descendant, since I worked on the evaluation for both, although the rest of the engines are not similar. Fritz is just too far below these top engines to be useful.

Stockfish appears to be the engine of choice for many operators. Its apparent speed and tactical accuracy are impressive indeed and make it a particularly difficult opponent to overcome.

No comments: