Correspondence chess will be dead within 8 years

[Alan Tomalty]

The last world correspondence championship ended with 90% draws as has been previously reported. Since they play with computers they are playing at a level of around 3400. I conjecture that since the computer chess rating improvement is around 50 points per year (this has been a surprising constant since the beginning of computer chess) then within 8 years time the top programs will be playing at a 3800 level. Since chess is a draw at God's level of play, I conjecture that at that level of 3800, all games will be drawn. Thus to all intents and purposes the correspondence game will be dead. The proof of this is that if you look at the % of draws for each level of play, you will find that the drawing % goes up by 5% for each 200 point rating interval. At the present candidates tourney in Moscow, 75% of the games are draws. The difference between their level of 2800 and the top computers works out to exactly 5% for each 200 points. As you go down the rating level each 200 point level will draw 5% less. The graph curve is not exactly a straight line all the way down and in fact the weaker you are the more decisive the games become so that the 5% difference becomes closer to 10 % difference for each 200 point rating interval at the weaker levels but you get the picture. It may be that at the top levels above 3400 it becomes 4% or lower, but this would only add a few years to the future of correspondence chess. In the case of humans since they improve at around 5 rating points per year it will take them 200 years to reach 3800. So the human game of chess is not dead yet!

[Matthew Nicholson]

Re: Correspondence chess will be dead within 8 years

I thought correspondence chess only existed as a way to tell if an opening book was well researched...

[Paul Bonham]

Re: Correspondence chess will be dead within 8 years

The last world correspondence championship ended with 90% draws as has been previously reported. Since they play with computers they are playing at a level of around 3400. I conjecture that since the computer chess rating improvement is around 50 points per year (this has been a surprising constant since the beginning of computer chess) then within 8 years time the top programs will be playing at a 3800 level. Since chess is a draw at God's level of play, I conjecture that at that level of 3800, all games will be drawn. Thus to all intents and purposes the correspondence game will be dead. The proof of this is that if you look at the % of draws for each level of play, you will find that the drawing % goes up by 5% for each 200 point rating interval. At the present candidates tourney in Moscow, 75% of the games are draws. The difference between their level of 2800 and the top computers works out to exactly 5% for each 200 points. As you go down the rating level each 200 point level will draw 5% less. The graph curve is not exactly a straight line all the way down and in fact the weaker you are the more decisive the games become so that the 5% difference becomes closer to 10 % difference for each 200 point rating interval at the weaker levels but you get the picture. It may be that at the top levels above 3400 it becomes 4% or lower, but this would only add a few years to the future of correspondence chess. In the case of humans since they improve at around 5 rating points per year it will take them 200 years to reach 3800. So the human game of chess is not dead yet!

Interesting post. I'd like to see some links that point to sources for the claims made... although I agree, there would be no source for the claim that chess is a draw at God's level of play. One could suppose that the entire universe, and all the parallel universes it spawns and the entire tree of parallel universes, is merely a tool of God to find out whether there is a forced win in chess from the normal start position. But then again, there is the deeper question of the game of GO...

(I have a pet theory that God is in fact a sports nut, and is constantly watching all our modern sporting events... the fact that it took 4.5 billion years to get to this era where there is even professional bowling would be lost on God, being eternal and all... but i digress...)

Is it possible that correspondence players are simply not playing "adventurous" enough? Faced with this prospect of the death of their game, shouldn't the players start modifying their move decisions, taking more chances? The rewards could be enormous: imagine being known as the Tal of correspondence chess!

Well, if that doesn't happen and the death spiral does continue, eventually correspondence chess may be the willing vehicle for some kind of important rule change in standard chess to avoid draws. Perhaps Option Chess will become the new standard chess, since there is no engine that can play it and to write such an engine and make it play well would take a few decades imo.

[Mathieu Cloutier]

Re: Correspondence chess will be dead within 8 years

Of course it was just a matter of time before correspondence chess would hit that wall.

Still, I could see it going at 10% decisive games for a while. There's a human factor in there. But the day it hits 100% draws because computers are too strong, will be the day humans are really outclassed by the engines.

And, Alan : where's the evidence that top human chess players are progressing at a rate of 5 Elo points per year?