08 February 2021

Correspondence Chess in the 2020s

A couple of recent posts on the Talkchess forum raised questions about the current impact of engines on correspondence chess:-

The earlier thread was started by GM Larry Kaufman, a chess engine heavyweight who has been mentioned many time on this blog (see the search box on the right). The later thread was started by someone who admits to 'very little knowledge in the correspondence chess world'.

The bottom line is that top players all use equivalent hardware & software. They play draws against each other and are 'top players' because they have an advantage over players with lesser hardware and/or software.

Any decent club player, let's say 1500 Elo or more, having sufficient means and technical knowhow could build an engine platform comparable to those used in the TCEC/CCC (see Stockfish Wins TCEC S20; CCC 'Rapid 2021' Underway, for the latest post in a long-running series on this blog). That player would immediately become one of the best correspondence players in the world.

Why don't more people do so? Probably because they have more challenging uses for their time and money. As for the problem of too many correspondence draws, the problem has been building for years. I did an analysis more than eight years ago, Increasing Draws in ICCF Finals (December 2012), and concluded,

If the trend continues, we'll eventually see some yellow [1-0], no red [0-1], and a lot of blue [1/2].

And that's where we are today. Correspondence chess is all about mechanical results that lead to meaningless ratings and meaningless titles.

So why do I play? To learn more about chess -- to take apart a game in progress and understand what makes it tick, like an auto mechanic takes an engine apart to understand its internals. I've been doing it for decades and I have no plans to stop.

How do I take a game apart? That would be a good topic for another post.

1 comment:

Douglas Stewart said...

Someone 1500 ELO with a great setup in ICCF might draw the strongest players, but ironically their struggles might be more with trying to get wins against the 2000-2200 players. In that range you can still try to induce mistakes, use move orders to try to steer to a favorable opening line, and find the path that gives you the best possible chance to convert an advantage into a win. Some chess skill is a benefit.

I find ICCF interesting to build a broad opening repertoire (to be used probably mostly in blitz games) and hoping looking at how the engines play might bleed over into my OTB play. I've been playing less than a year though so I might change my mind as time passes.

In addition to ICCF, I also play correspondence at leagues that don't allow engines and that's plenty exciting and at no risk of being ruined by stronger computers. It's not too hard to diverge from standard opening lines in no-engine leagues that do allow access to ICCF databases.