Thursday, June 11, 2015

The Draw Death in Correspondence Chess

    Capablanca proposed a chess variant in the 1920's because he believed chess would be exhausted in the near future and games between masters would end in draws. On his 8x10 board each side had additional pieces: Chancellor, Archbishop and two additional pawns. Then along came Fischer with his belief that openings had been analyzed to the point that games were decided by opening preparation and GM's could no longer be creative. He complained, “...chess is completely dead. It is all just memorization and prearrangement. It’s a terrible game now. Very uncreative." His solution was to begin each game with a random placement of the pieces. The idea was that by doing so opening theory was abolished and the players would be on their own. Fischer Random Chess, aka Chess960, was born. It hasn't really caught on either. The truth is that whether you add a couple of pieces or shuffle them around, the best players will still be the best players. 
    With the advent of chess engines a lot of people believed chess would be solved, but sometime in the 1980's that belief was pretty much abandoned and correspondence players at the top didn't have much faith in engines because they could still beat them. The main value of engines for them was in the area of opening research and locating opponent's games. Before that opening research was tedious...you had to have a big library of books and magazines and notebooks. Eventually though, as engines got better, they began to sneak in, mostly to blunder check so you didn't make a disastrous error. But that was soon to change.
     In the pre-computer world correspondence championships white won 37 percent, black 24 percent and 39 percent were drawn. When engines started being used things changed. White's average winning percentage dropped to 27 percent, black's to 12 percent and a whopping 61 percent of the games were drawn. And things continued to get worse. In the 2008 in the ICCF World Championship draws exceeded 80 percent and now it's over 90 percent! I checked the results in a number of high level tournaments on LSS and found the draw rate to be 85 percent. 
     Back in 2007 Ivar Bern stated he was not worried by the future of correspondence chess. His argument was that even though CC was getting close to perfection, the “small scientist inside” of him made computer use appealing. But huge advances in engines have changed things since 2007. What we have today in correspondence chess at the top levels are multi-core, multi-thread and huge memories on dedicated computers. The accuracy of the play is such that draws are almost always the rule. Make that boring draws. For that matter, even the wins are boring. At the highest level, correspondence chess has become an arms race and, as one writer stated, engines are not programmed to win, they’re programmed not to lose, hence the extraordinary number of draws. 
     On LSS I've played over games by some of the lowest rated players and judging by the number of gross blunders a lot of them are not using engines, but in many cases, judging by the near-perfect play, many of their opponents are. So, how do you explain an engine user at such a low level? The answer is that when you play on any site where MOST players use engines, and so most games end in draws, it's almost impossible to gain rating points and advance to the next class. Also, as you move up the food chain and get closer to the top, you find more and more players using dedicated computers. The result is you are pretty much stuck at whatever level you started at! 
     I checked the results of a number of top level tournaments on LSS and found only 15 percent were decisive and 85 percent were draws. My own LSS record, with somewhere around 250 games, stands at 24 percent won, 15 percent lost and 61 percent draws...not much fun! In the early days it was fun trying to outwit engines, but the time came when I just couldn't do it, so it became a matter to “testing” various engines and even trying out some discredited and rarely played openings. No engines can compete with Stockfish and Komodo and even when you play stuff like the Urusov and Latvian gambits and 60-some percent of the games are still drawn, where's the fun in that?! 
     What's left in the correspondence chess world? Nothing that I can see.

1 comment:

  1. there is another very interesting in https://www.albertochueca.com and in addition he offers a very good chess guide...

    ReplyDelete