Got Draws?
Curious about the latest correspondence world championships, I took a gander at the ICCF website and came across this crosstable of the ongoing 29th championship final. 44 of the 136 games are finished, and every single one of them has been drawn.
This isn't quite as bad as it sounds. In correspondence chess a quick loss should be a near-impossibility at the world championship level, so the first games to finish ought to be drawn. 44 in a row seems a bit much, though. The 28th championship is almost finished; there are just two games remaining and Croatian correspondence GM Ing. Leonardo Ljubicic has clinched clear first. Of the 134 games that have finished, 18 were decisive - six of which came at the tailender's expense.
The draw death isn't much of an issue for OTB (over-the-board) grandmasters (if at all), and for the rest of us it's a complete non-issue. But is correspondence chess on its last legs? (And if it is, can it be fixed?)
Reader Comments (6)
If correspondence chess is on its last legs (without rule changes), then engine tournaments like TCEC should follow some 10 years later, according to Moore's Law. Too much reductionism? Wait and see.
I hope this unique, great chess blog will still be healthy alive in 2026 and later on!
[That's a long time, but who knows? I've been at it for 12 years now.]
The crosstable of the ongoing final looks curious: 12 draws for Manso Gil means "top of the standings", no draws yet (all games still ongoing) for Serban "dead last place" - probably no relationship whatsoever to the eventual final standings ... . Once there are decisive results, it will make much more sense to look at plus and minus scores. If the next game to finish were Serban-Manso Gil 1-0, Serban should be on top with 1/1(+1), and Manso Gil at the bottom with 6/13(-1).
As to the previous event: While I am not particularly interested in correspondence chess (long ago that I played two events myself), I am not nearly as pessimistic as Dennis about its future. It is high quality, arguably even better than engine chess - engine use is common, thus the challenge is to find something engines won't spot. [I discussed this with Dennis some time ago: engine use is forbidden in US correspondence chess, but not internationally - the rationale might be that it doesn't make sense to forbid something if this rule can't possibly be controlled and enforced.] Mutually high quality means high probability for a draw. As long as some games still have a decisive result, correspondence chess is alive. I clicked through a few games of the event - many (both wins and draws) look interesting. A few finished with rather early draws - in such an event, it becomes important to focus on your promising and your difficult positions, and if both players are in the same situation and have the same opinion about their game (most likely a draw even if we continue for many more moves) they agree a draw.
Analogy to other sports: Is high-level/professional running "almost dead" because the margin of victory is much less than a second over 100m, and at most a few minutes (still negligible with respect to two hours) over the marathon distance? No, small differences decide, same story for correspondence chess.
[DM: That's a very strange analogy. The issue isn't that the competitions are close, but that they seem to be homing in a 100% draw rate. Here's a better analogy: suppose that in soccer goalies were somehow "enhanced" to the point where they could block 99% of the shots on goal, and they grew more efficient every year. Wouldn't we think that something would have to be done after several seasons dominated by nil-nil scores? Either the enhancements would have to stop, the goal would have to be enlarged (or some other rules change would be needed), or fans would have to find another sport.]
In soccer, a tense 0-0 with plenty of spectacular saves by one or both goalkeepers is also worth watching.
[DM: Yes, *a* tense 0-0 game. If practically EVERY. SINGLE. GAME. is 0-0, however, things will have to change.]
And if a soccer match is defined as a series of games, with each attack a single game, soccer is a very drawish sport. Some draws are exciting (goalkeeper action required, near-misses), others (not finishing with an attempt to score) are boring, same in chess!? Even if a team wins 5-0 the "total score" is something like +5=30 ... .
In correspondence chess, the equivalent of "enlarging the goal" or rather "stop enhancements" would be to ban use of engines. This seems hardly possible - only via surprise visits to players and (if technically possible) scrutinizing their computers and mobile phones for positions that have been analysed with engines. I think correspondence chess never got much attention beyond its own community - results are mentioned somewhere, games end up in databases but are only noticed if they are (win or draw) spectacular and/or theoretically relevant - so it's up to the players themselves. Some keep playing, others quit. A chess contact (FM over the board) wrote me "I was an ambitious and fairly strong (peak Elo 2570) correspondence player, but engine clicking isn't my thing so I quit."
"Wouldn't we think that something would have to be done after several seasons dominated by nil-nil scores?"
Actually in the last 30 years or so FIDE has experimented with measures to reduce the amount of draws, to make the sports faster etc. Nothing radical, but this clearly shows the answer to your question is yes.
If ICCF wants to survive it will have to change rules as well, because fewer and fewer people will keep on playing.
Well as a strong correspondence player, I will contribute my 2 cents.
First of all, I do not know my draw rate over the last few years so I want to posit a hypothesis which I will come back to with the facts later. I have been playing since 2008, over 800 correspondence games have been played in that time. I have gone from 1300 to 2400. In that time, I have watched the strongest engine change a dozen times, the rise of opening books, the creation of idea, the importance of hardware increase exponentially, 64bit, 7 piece tablebases, engine matches increase in necessity and literally dozens of other changes.
The draw rate is certainly one and it is probably less due to engine strength than people think. The bottom line is it that it is harder than ever to get an advantage and even harder to send that advantage home. I would say in half my games or more that I end up with an extra pawn, exchange or some other tangible benefit that I can't convert due to the amazing increase in resources for defense.
All this information was to give an understanding from which I write my hypothesis. I would guess in the last few years that my draw rate has risen from 70% to 80%.
While it is definite that my draw rate has grown, it does not tell the full story. Has it grown this much? 10% is quite significant in my book. It is not enough for me to lay down the mantle. Indeed, I hope to play 1,000 games and achieve IM before I do that. However, the story is winning has gotten harder. I have become a far more aggressive player and spend thousand more hours than I used to win. I research openings now for 100s of hours before the game even starts. It is exhausting. It has lowered the number of games I am willing to play at a time and it has made achieving the IM title nearly impossible. Unlike OTB, in correspondence, you only need two norms because it is so difficult. You usually have to put up a +7 score when you may have at most 6 whites. Requiring someone to win as black is one of the hardest things to do at correspondence. I have missed roughly 10 IM norms by a half point at this point. To put +7 into perspective, +1 is usually enough to win a tournament. As someone who has seconded a certain GM in many of the previous world championship tournaments, I can't even begin to describe the level of work that has gone into researching openings for potential.
Is correspondence dead? Long answer no but with a *.
Just to follow up on the conclusion.
My hypothesis was: "I would guess in the last few years that my draw rate has risen from 70% to 80%."
In 2013, my draw rate was 70%.
In 2014, my draw rate was 82.4%
In 2015, my draw rate was 84.1%
So there has been dramatic increases. I can go back further but then many more factors come into the analysis. At least over the period of 2013 to 2015 the level of opponents was 2400 ELO and ICCF.
[DM: All the same, I think we can all "draw" the appropriate conclusions.]