The State of correspondence chess


Back to forum

Daniel Parmet    (2020-04-28)
The State of correspondence chess

I have played correspondence chess now for 13 years. During that time, I have played 983 correspondence games. These days I mostly play at ICCF and some of these issues may be ICCF specific... but since ICCF has no forum and I want to get a sense of the health of correspondence chess in general... I posit my thoughts here.

First of all, I think the number of correspondence players and the number of correspondence games are decreasing across the board on all correspondence websites due to the things I want to talk about.

Second, I primarily shifted my playing to ICCF years ago for two reasons: 1) The higher level of competition available; 2) The norms available. Although I was concerned with their fees which are usually minor but, in many cases, certain organizers do construct outlandish tournaments that you need to be wary of (looking at you Venezuela).

On the first point, I think ICCF is a little more open to high caliber players competing up until a point (they really try to prevent you from playing a 2450+ player until you are 2450+ yourself). And the rating protections get tougher and tougher the further you go but they make it easy to play 2300 players. While most websites outside of ICCF, usually have one annual Cup / WCH or Thematics, these other websites usually make it impossible to play anyone more than a few hundred points above you no matter your rating outside of these few events.

On the second point, I think ICCF norms are somewhat of an illusion. They’ve always been hard and much harder to achieve than OTB norms which received a watering down of requirements of decades ago. In fact, ICCF norms are so much harder than FIDE norms that one actually needs to achieve two norms to receive the prerequisite title in ICCF vs the standard three norms required by FIDE. In the US, for example, there are 116 ICCF Titled players in history (13 GMs, 25 SIM, 78 IMs) vs 828 FIDE Titled players in present (101 GMs 166 IM 561 FMs) []. Now however, there is a proposal, for the ICCF GM Title only, proposed by Dennis Doren, ICCF Rules Commissioner who really does a lot for correspondence chess, and SIM Uwe Staroske, ICCF Qualifications and Ratings Commissioner, to remove the requirement to have to play GMs to get the GM Title [leaving IM and SIM untouched] []. This proposal states, “A search of the ICCF data indicates that 21 players obtained at least 2 GM norms across 24 games but failed to get the GM title because of the requirement of "5 GM" opponents. (Only 5 of those players are currently active).” Leaving aside the fact that this proposal violates the very definition of the GM Title, one must beat the club in order to join it, the proposal further outlines the real problems without addressing them, “The GM Title has already become far harder to earn than it used to be, due to the rating suppression caused by the increase in draws.” Wow, let’s unpack that one line because it is a doozy!

Really, this one line, that is easily overlooked, is two huge problems that correspondence is facing: 1) death by one thousand draw paper cuts and 2) rating deflation. I will argue later that there is a third huge problem but let’s start with the ones acknowledged by ICCF itself. Every correspondence player knows the draw rate is going up. As engines and hardware get stronger, players are able to save positions that in the past would have been lost and we are finding ever easier ways to head straight towards 0.00 as Black. I would love to see a detailed analysis that describes how much harder it has become to win as Black against a decent correspondence player (let’s say someone 2300+). In the last five years, I have beaten three 2300+ players as Black without counting mouseslips (one in 2015, one in 2016 and one any day now in 2020) despite playing extremely aggressive openings like the KID (for the record that’s three Black wins out 103 Black draws or 2.91% Win rate). That may be part of the draw problem, but I have witnessed my own draw rate skyrocket 2014: 82.4% 2015: 86.7% 2016: 90.2% 2017: 90.6% 2018: 91% 2019 is still in progress. Often for these norms, you need to score +2, +3, +4 or +5 despite the fact that +1 usually wins the event… and with the draw rate North of 90% in a 12-13 game event that means you are likely to win 1 game on average… but in many events the entire cross table often sees one to three entire wins (look at a recently completed tournament here where I scored my first IM norm that required +0 and I scored +1). My win was one of five wins in the entire tournament 100/105 = 95.2% draw rate! []. People love to tell me that’s fine because we are talking about such a weak event as Category 8 [2449 was the rating average]. Fine, I do not accept your argument but let’s look at the World Championship then shall we? Let’s look at the most recently concluded World Championship 30 which finished on 10/2/2019, Category 13 [2562 was the rating average]. This event was won by the new World Champion SIM Kochemasov, Andrey Leonidovich 2540 []. Congrats to the new World Champion on his two wins! The event had 8 decisive games out 136 or a draw rate of 91.2% (not far off my own). But wait did I say SIM? I did. In fact, congratulations to the World Champion on scoring his final GM norm as well! This World Championship saw 5 SIMs compete in a field with 12 GMs. While 3 of the SIMs finished 1st 2nd and 3rd, only our new World Champion scored a GM norm. The problem is with all the draws that norms are not just becoming hard, but maintaining or increasing one’s rating is becoming hard. And one’s rating is how one receives any decent invites to have a chance at a norm in the first place.

The draws are a death by one thousand cuts as I recently played one of the ICCF’s proposal’s outlined “21 players that could have obtained a GM norm.” My rating is 2389 and his rating is 2504 (although SIM, he is recognized by all his peers as a GM caliber player). As Black, I obtained an easy draw without ever being in any trouble at all. The player had a rather angry initial discussion with me post mortem about how he felt it was wrong that a 2504 should have to play a player as weak as 2389 where the draw would kill his rating. He felt that his rating was being destroyed by these draws with weaker players and that ICCF should protect him from us. He felt I have it easier as a lower rated player because I can gain rating from these draws. Let’s look at his argument that one is causing the other and it is only happening to those 2500+. At the time that draw occurred, I gained exactly 1.17915 rating points from it (and he lost the same); however, this was the first draw in over 40 games in which I *gained* rating points (this statement is no longer true as a few higher rated players have since given me draws but at the time of the game’s conclusion this was the case). Yes, that’s right, ICCF already does such a good job of protecting higher rated players that it actively hands out advice to new players to be very particular about what invites and events they play because the draws could kill their initial rating. I too have experienced a net negative loss of rating points from draws and still seen my rating going up only due to the fact that wins are easier and ever so slightly more common to come by at my level. However, it means I am not exempt from the draw problem. It is patently false that this problem is limited to those 2500+ as in my last 43 draws, I lost rating in 42 of them and gained rating from 1 of them. Therefore, it appears draws are causing rating deflation and this is the real problem in both norms and correspondence in general. With the exception of matches, perhaps there is a way to have draws not count against one’s rating since there are so many of them? It kind of blends the Chess rating concept with that of Bridge where one cannot lose rating points once earned. What we can see is that the player’s argument that draws are causing rating deflation is probably true. One problem is at least partly causing the other one.

There is a third more devious problem worse than the two outlined above in my opinion. While rating deflation, draws, less players and norms are real issues… they are dwarfed by the change in behavior caused by these issues. I know it is a bit overdramatic to talk about such issues in a time of COVID, but there has been a great increase in the number of players playing Dead Man Defense (often shortened by correspondence players to DMD+ and DMD=). It is important to note that the death rate in COVID for those in the elderly category is markedly higher and the correspondence community in general is also markedly higher. I have heard estimates of the average age of correspondence player being 70-75 range though I haven’t seen any data. Back to DMD, what is DMD and why is it such awful behavior? The players are hoping you die before you win so they can claim either a win on time or if it goes to adjudication then at least claim a draw. The other hope is that you might mouse slip by being forced to play more moves which while that would never happen over the board does surprisingly account for a large portion of wins in ICCF correspondence high-level play. One of the main problems this issue causes is that if someone takes an early draw against a player who then goes on to die, the entire rest of the field gets a free half point and you are punished for playing your game quicker than your peers. Often, players over the board resign once mate is unstoppable or a simple endgame is reached in which the result is known to players of all levels. In correspondence, often even sooner than these players will resign or offer draws, knowing that perpetual check is unavoidable should we play another 10 moves past the piece sac against a bare king? How about when the engine reads +25 +30 or +40? So, for the most, correspondence players draw or resign much earlier than one might over the board due to engine and tablebase assistance. On that note, depending on the tournament, players can outright claim wins and draws either on the 6-piece tablebase (always allowed) or the sometimes allowed on an event by event basis the 7-piece tablebase. It is considered out right rude to make a player play all the way to the 6-piece tablebase to claim. I recently claimed one win in a six piece tablebase up an entire piece where my jolly opponent wanted to discuss the game in a post mortem (rarely done in correspondence in general anyways). I declined to even respond to him even though I was already having a very lively and fun post mortem with a Venezuelan on our extremely interesting draw. A worse example is the 92 move game I played with opposite colored bishops where I had two extra pawns. I offered a draw as white and the higher rated player to my lower rated opponent who declined it, forcing me to play to a 7-piece tablebase claim to end the game. This kind of behavior used to be quite rare. In the past, I would say it happened in 1 out of every 100 games… these days it seems to happen in every other game (1/2!). I have seven different opponents right now that are DMD+ against me where the engine reads +148 (or in some cases even sees mate! The 2504 player that complained about my rating earlier also complained someone was DMD+ him… I remarked that I have no less than 7 players DMD+ me and if they would resign? My rating would be about 2450 right which sort of eliminates his claim about our “giant” rating difference). The issue is that due to rating deflation these players need to artificially keep their rating high as long as they can because that’s how they will get their next invite. With the new terrible time control that is not yet Official (although there is a proposal to make it Official:, players only need to make a move once every 50 days to pointlessly extend the game. I have a DMD= draw currently going on 16 months now where the player is just moving Kg1 Kf1 Kg1 every 50 days. This time control exasperates the DMD problem. When I contacted ICCF Officials to point out the severity of this problem, I was told that I should report it to the TD on a case by case basis only if it is DMD+ as they will not look at DMD= at all. However, it is usually the TDs that are the biggest offenders (6 of the 7 players described above were TDs). In fact, it is usually the same general casts of characters which allows for an easy black list to be created that bars these players from play until they can fix their atrocious behavior. This behavior needs to be punished. These players need to be reprimanded. In the end, lack of norms, rating deflation and the draw death will not make me quit correspondence chess. It is DMD+/DMD= that will make me quit. This experience is my personal experience with high level correspondence over thirteen years and I would love to hear from other correspondence players concerning these problems.

Herbert Kruse    (2020-04-30 21:04:25)
The State of correspondence chess

ICCF is affraid of real opps, i tried it, but fight against burocrats is senseless