Random Posts

Monday, July 14, 2014

Thoughts on Study and CC Play and Engines

Whose Games to Study? Alekhine’s, Capablanca’s or those of Carlsen and Aronian?

     In the recent edition of Chess Life magazine, GM Andy Soltis in his column Chess To Enjoy, made the following observation: “Chess combinations have a way of being repeated…This is why studying great tactical battles of the past is so useful.” 
    Then he went on to add, “…games featured in books and magazines that you read are the ones played in the very recent events. Yesterday’s games-not the ones played a hundred years ago-will have the greatest impact…” 
     Soltis pointed out that there is, however, a certain disadvantage to playing over the games of today’s great players. He opined that the games of the great players of yesteryear were grounded in “classical” chess and the problem with games played by today’s top GM’s is that they have less to teach than the classics of old. The reason is, according to Soltis, that a typical game played by one of today’s elite GM’s is they begin with 15 or more moves of engine checked home analysis and what happens in those moves is not understandable by the rest of us. Soltis points out that you will learn more by seeing how mistakes were punished, but when a great player of today wins it’s usually because his opponent’s mistakes were almost imperceptible. That was not usually the case in the old days. There was an interesting discussion on this subject on Chessdotcom HERE.

Engine Analysis and Post Mortems

     I have heard some lower rated players don’t bother doing a post-mortem on the grounds that they aren’t good enough; they would rather let an engine blunder check the game to see where they went wrong. As a result they miss an opportunity to improve because chess is about ideas and talking things out with other players can’t help but broaden your horizons.
     All the engine is going to do is show you tactical mistakes and if there aren’t any tactics, an engine will show you positional moves that were mysteriously arrived at by its algorithm without explaining anything.
     Even for the best players, engine lines can be difficult to evaluate. If things were as easy as letting the engine select your move I’d be playing in the world CC championship. Unfortunately, even at lower levels of CC play, things aren’t that easy. On LSS my record is only +44 -35 =99 and that’s mostly against players in the 1900-2100 range.
     With engines you can analyze for days without coming up with the “final” answer as to what the best move really is. Also, there is the question as to which engine is best. Larry Kaufman (Komodo developer) believes Komodo is best for long-term analysis because it is more positionally programmed. Houdini is better when it comes to blitz and tactics. Stockfish? I don’t know. Critter also figures somewhere in the mix, too. The recently released Stockfish 5 definitely appears to be better than Houdini and Komodo, ranking ahead of them at a time limit of 40 moves in 40 minutes.
     Top level CC play sees most of the games being drawn, but a lot of that is because at that level, nobody wants to take any risks. Heavily analyzed openings are the norm…Najdorf Sicilian, Ruy Lopez, Semi-Slav, Catalan or the Nimzo-Indian. The King's Indian is rarely played because engines don’t evaluate the resulting positions well. I’ve had opponents play some rare gambits and lesser known openings against me on LSS and the results haven’t always been what you’d expect but then I don’t let the engines spend hours analyzing and I try to incorporate my own evaluations into the position, too. Not being a top level CC player means that sometimes things don’t work out well even against supposedly inferior openings. Also, in positions where there is no clear path to take, you need to figure things out yourself so there’s still, at least for those of us lower down on the food chain, some room for continuing to play modern day CC.
     Back in October 2012, ICCF announced a new CC World Champion; Ron Langeveld from the Netherlands. To achieve that status it helps to be a strong OTB player because it takes a very good understanding of how to get an edge, especially when the engines aren’t showing any particular difference between several moves. Also at the upper levels your opening repertoire has to be super-solid and you have to be able to ferret out good opening innovations. And don’t forget endings!
     Langeveld said in an interview that chess knowledge can be bad in CC in some cases because even the strongest CC players make too many mistakes without engines as a tool. One of the main problems strong players have is that because of their chess knowledge, they move too quickly.
     So, how much time does it take to make a move? Pertti Lehikoinen, who won the 20th CC World Championship which lasted three and a half years, said that in the beginning he spent eleven hours a day on his games and for several months he had to increase it to seventeen hours a day. In the end, he spent more than 14,700 hours on the final; that’s an average of twelve hours per day and that didn’t include the time he spent thinking about his games while doing other stuff! Lehikoinen also admitted to fainting because of fatigue at times. My last game to finish on LSS was a long one…9 months! The average was 3-4 months and I rarely spent more than half an hour on a move and never once did I faint from fatigue.

A snappy U.S. Marine-style salute to Mr. Lehikoinen for his dedication!

No comments:

Post a Comment