Random Posts

Monday, September 15, 2014

More Thoughts on Correspondence Chess

Modern CC Master's Work of Art
   I ran into a friend the other day who asked me if I was still playing chess and I told him just correspondence and the discussion, of course, got around to CC these days and engine use. Like many, he didn’t see the point of letting my engine play against the other guy’s engine. The oft heard comment that you don’t learn anything that way also came up. First, I explained I am not trying to learn anything. If you want to learn something, using engines only has a limited value.
My CC Work of Art

     For some players the point of correspondence chess is to play a game using postcards or servers against another opponent with both using only their own brains. For others it involves playing one engine against another and trying to coax a win out of the position any way you can, be it by using more powerful hardware, deeply researching openings, getting into positions engines don’t play especially well, using your own judgment, etc, etc. The problem is when the users of engines encroach on the domain of non-engine users who are actually are playing to learn something or who simply prefer to play without engines. In this case it becomes cheating because it’s against the rules. However, on places like the ICCF and LSS where there are no rules against engine use, unless you are a near beginner, you will be facing other engine users in which case failing to use an engine means you will lose. That wasn’t the case 20 years or so ago because engines weren’t that good, but not now. LSS does offer Chess960 and no engine tournaments, but I am doubtful if everybody in those events really doesn’t use an engine to some extent.
     ICCF and LSS allow engines because it's pretty much impossible to police their use and as I have stated before, it is more than just who uses the strongest engine. It is a different skill that involves knowing how to optimally use an engine in combination with positional knowledge that engines don't have. In engine assisted chess somebody always wins more games and establishes higher ratings even though they are playing against other people’s engines, so how is that? You simply can’t get a high rating on ICCF or LSS by just copying the computer's moves.
     On ICCF and LSS very few, except perhaps the very lowest rated, play simply for fun. People are out to win because wins gain you rating points and the possibility of reaching norms. Of course these points and norms are meaningless in real chess (OTB), but they do have meaning for players on whatever site they happen to be playing on whether engines are allowed or not. Besides all that, winning is better.
     Top players generally use more than one engine. I use mostly Stockfish, Critter and Houdini 2, but the best move is still not always obvious. For example, in a recent engine tournament I ran, Stockfish finished first but drew with Houdini and Critter which tied for 2nd and 3rd. However, SF also drew with Komodo 5 which finished in 6th place. Meanwhile Houdini lost to 4th place Gull 3 and drew with 7th place Deep Rybka 4. Critter suffered its only loss to Houdini. So, there are no foregone conclusions that even with a top rated engine, it will, left to itself, always win. Then too, sometimes you get quite different evaluations from different engines on the same position. All of which complicate things when trying to find the best move.
     The really top CC players do a very deep opening analysis of top lines. For example the Bogo Indian and the Catalan enjoy great success in the upper levels of CC play these days, but personally, I don’t like playing the same opening over and over. In fact I recently played the Urusov Gambit (1.e4 e5 2.Bc4 Nf6 3.d4 exd4 4.Nf3) in a couple of games because I thought there was some improvements to be found in published analysis. The result was a win and a loss, so for many of us playing the Bogo and Catalan aren’t necessary; you can still have some fun playing other stuff. Unless you are looking for ideas, opening manuals are useless. Forget about lines Fischer used in 70s. Another key factor in modern correspondence chess is to find positions that are initially evaluated favorably by chess engines, but can be shown to be faulty after deep analysis. This is where something called CAP comes in. CAP is short for the Computer Analysis Project, an effort started by Dann Corbit. Tens of millions of positions have been evaluated and some engine opening books come with a large database of such pre-computed positions. These CAP evaluations may be different from the evaluations computed by the engine during game analysis.
     If you’re a CC player, save your money on opening books unless all you want are general guidelines and ideas. You will likely find “holes” in the opening books by engine analysis and searching through your database of 2 million (or more) games. One would think that by simply copying a GMs moves in the opening it would guarantee getting a good game, right? Wrong! They sometimes play inferior moves and sometimes their analysis is wrong.
     You have to decide which openings/defenses you are going to use. As White, play the old routine 1.d4 with a Torre Attack where I’m pretty familiar with the positional themes of all the various responses or risk 1.e4 and play against the Najdorf Sicilian with all its tons of theory which I am not familiar with. And, if my opponent plays 1…e5, then what? Play some obscure line and get caught in an inferior position or stick to mainline openings and get caught in some novelty. Even in a familiar line you run into somebody who has done their homework and found an improvement. These are the same problems the Grandmasters face.
     CC players are different animals than OTB players. Chess books have been written by OTB players so a lot of stuff about openings, tactics, especially tactics, don’t apply in correspondence chess because tactics usually aren’t an issue…the engine will make sure you avoid them. What’s important are opening research, strategic ideas and endings and that's enough to turn a lot of people off engine assisted chess right there! 
     In the end correspondence chess, whether played with or without an engine, is not a game that some people are particularly interested in playing, but the main thing is that either way you choose to play it, it can be challenging. I just wish people who choose to use engines would stay off sites where their use is prohibited. If they would just do that, the controversy over engine use would pretty much be eliminated.

No comments:

Post a Comment