Talking modern correspondence chess

by Davide Nastasio
6/26/2018 – Does correspondence chess have any meaning in 2018, when chess engines definitely overcame the best humans already back in about 2005? Is it useful for a chess player to study a database of correspondence chess games? Or to use it as a reference? Discover more in an interview with Wolff Morrow, one of the top ten correspondence chess players in the USA!

Corr Database 2018 Corr Database 2018

Corr Database 2018 is an extensive collection of correspondence games, featuring classical correspondence games played by mail as well as email games.

More...

A preponderance of correspondence

Look, let me cut to the chase: this is the most stunning database one can come across. Many don't fully understand the revolution which happened in correspondence chess with the advent of engines. One could mistakenly think all the work is made by the engine, but this is not the case. The research into the opening is made in collaboration between the human and the engine, creating a new kind of chess player, never seen before: the Centaur.

We know that a comparison between the best human and an engine ceased to be relevant after 2005, because engines became so strong that no human could really pose any threat. In fact, many mistake the 3300 rating of the top engines for a human rating, but it is not. It is calculated between machines which, with one second per move, are able to avoid tactical mistakes five moves ahead and, within a fraction of a second, can calculate a checkmate in 12 or more moves.

Similarly, one could argue the AlphaZero programmers erred when they made their comparison solely against the top engine. The mistake was to think the best engine is de facto the strongest chess player on the planet. This is not true. The strongest chess player on the planet is the Centaur, one who has a machine dedicated to chess, an SSD (Solid State Drive) for the endgame tablebases — or an Endgame Turbo USB flash drive —, and an SSD for the opening repertoire, which has been deepened through the years with the Monte Carlo method. The findings made by the Centaur are nowhere to be found, even in commercial engines (if a commercial engine is the top one to find!). Against such a creature, AlphaZero would have had difficulties to win even one game.

I'm not a correspondence chess player, I'm a tournament player, my memory is limited, and to calculate a five-move checkmate —especially if it is not a forced linear sequence — can take a while. Practically, you could say, I'm a snail in a world of elephants and tigers.

So why was I so excited to acquire this database? Because I believe it can teach me a lot! How? Let's say I want to learn an opening, where could I find the latest research made by computers and checked by humans? The question is rhetorical, since the answer is all there in the Correspondence Database.

I never played correspondence chess, but I know it is very interesting, so I thought I should interview one of the best players in the United States. Wolff Morrow is a great chess artist, and one can admire his masterpieces on the covers of the Gambit books.  

Cover of Extreme Chess Tactics by Yochanan Afek (2017)

For those who love studies, here is a close-up of the position, a checkmate in four six moves:

A study by E. Pogosiants Shakhmaty v SSSR 1976

 

Play out the study on the diagram!

Another beautiful cover comes from the book The Chess Attacker's Handbook, a great read, and of course another cover masterpiece. I loved it so much that I kept it as a desktop background for many months.

Courtesy Wolff Morrow

trophy

Here are his correspondence stats:

  • ICCF Senior International Master
  • Current rating: ICCF 2514 (ranked 9th in the USA)
  • 19th ICCF United States Correspondence Chess Champion

At right is the trophy for the 19th United States CC championship.

DN: When did you begin to play chess?

WM: Tough question to answer. My father was a 'Fischer Boomer' and so he played the game for many years. When I was eight years old in 1982, he taught me the rules and we played our first game. He took so many of my pieces that he accidentally stalemated me! So I ended up drawing a veteran player in my very first game! We never played each other again after that, as I wasn't really interested in chess at the time. Although on and off over the years there would be board games at friends' houses or at school that we would sometimes break out, and chess would be one of them, but again, I saw chess as just another board game along the likes of checkers and backgammon.

Wolff Morrow

Wolff Morrow at the board

What sparked my real interest in chess was in meeting new friends when I moved to Texas from Seattle in the mid-1990's. In 1996 one such friend named Sean was touted as being a 'good' chess player in town, and it struck me as odd because he didn't seem like the type of guy that would even play chess, much less to an arbitrary 'strength' level. We were all video game fanatics at that time, having Street Fighter II parties or playing Doom Deathmatch. Since I barely knew more than just the rules of the game of chess, I was under the naive assumption that the game was purely a test of your intelligence. I thought to myself: "If this guy plays serious chess, I should be able to give him a good game". So we started playing chess on a regular basis; at least 3 times a week. I quickly realized when he would beat me that there was a lot more to chess than simply your IQ. There were visual concepts that I hadn't appreciated, and basic fundamentals I was learning the hard way I knew nothing about. With that said, it didn't take long before I started trading wins with Sean, and this went on for a few months...

Then in 1997, I read in the news that Kasparov (then world champion of chess) was taking on IBM's 'Deep Blue' machine, which he had apparently defeated the previous year. Since I now was actively playing chess, I took a strong interest in this match and followed it from start to finish. The drama and tension was exciting to me. And I realized there had to be more I could learn from the game than just the level I attained from playing my friend all the time. That's when I went to the local recycled books store and browsed the chess section. The first book I selected and purchased was I. A. Horowitz's "Chess Openings in Theory and Practice". Reading this book was like a magical tome to me, revealing vast knowledge of the game I never knew existed. I found out that the opening defence my friend Sean liked to use to always trip me up was called the Scotch. I studied it and spent hours and hours going over the other opening chapters. Two weeks later, I met with my friend Sean and promptly crushed him off the board! I was hooked on chess from then on. I began going to the local club, collecting books, and playing chess online. I was 23 years old, so in all honesty, I realized I had missed out on the best years of my life to study chess as a child, and as such, I knew I wasn't going to be a world-shaker at the game. I'd just keep playing anyway for the fun and enjoyment of it.

I did play in a few OTB events and even managed to beat a National Master using the French Defence I had learned from watching a Chessbase DVD on the subject.


A Classical Guide to the French Defence

This DVD gives you the key to start out with the French Defence. GM Yannick Pelletier is a specialist of this opening, and believes that the most efficient way to understand its ideas, plans, and typical structures is to study classical lines.


However, my problem with OTB tournaments was the stress factor. I just simply couldn't relax and play to my best ability, and often found people I could easily beat in unrated friendly games would get the better of me in tournament situations. In my last officially rated USCF OTB tournament, I managed to win the open section with a perfect sweep and shared first place. The other player had seniority in rating over me but was gracious enough to defer the first place trophy to me since it was my first ever chess accomplishment, and the trophy was in memoriam of a fellow club player and friend of mine that had passed away the previous year. So it was a special victory to me, and seemed a neat way to 'retire' from rated OTB chess since my nerves were simply not letting me play to the best of my ability.

Morrow holding the David Wroe Memorial Open Section Trophy | Photo: Wolff Morrow archives

DN: For how long have you played correspondence chess? Was it always with engines?

WM: My first experience with correspondence chess was with the USCF's own correspondence program, and this was done via postcards. I only played one segment and did decently, except one paranoid player that I managed to eek out into a winning endgame accused me of cheating with an engine (even though I didn't even have a computer at that time). This really soured me on the experience, and I quit playing CC on the spot. I believe I still have a USCF CC rating of 17-something in their records.

I didn't pick up correspondence chess again until ten years later in 2007, in a special exhibition match that featured Top CC GM and 15th world CC champion Gert Jan Timmerman. In the match, he faced the "rest of the world". I was part of the world team, and in reality, it was about 20 or so of us that he took on (and not really the rest of the world). We used a 'hydra' style approach and would investigate branching lines of analysis on each move and report back our findings. Every now and then during the game, someone would suggest an interesting alternative to look at, and we'd spend hours and hours reviewing the new positions with engine software. I found this to be enjoyable because there was no paranoia or accusations about engine use. It was all 'on the table' so to speak, with both sides knowing full well that we've been exhaustively using engines to develop the deepest possible complexities in each and every move we played.

In the end, we could not quite find a forcefully winning endgame line, but we knew that with Jan being just one person, our best bet was to keep making the most difficult moves to solve. This approach eventually cracked Jan's defence as he missed the game-saving drawing line and went down in defeat. It was a wonderful experience to work very hard hand-in-hand with the engines and create a game that was at least at the time far beyond any human or engine could do on their own.  

 

DN: Do you have a correspondence chess player who is your favourite? In case the answer is positive, could you share a correspondence game which you found beautiful?

WM: One, in particular, I found played inspiringly deep correspondence chess is ICCF GM Nigel Robson (2617) from England. In reviewing his wins, I found them to be both inventive and incredibly deep. Not a single move was made with casual reference to an engine, and he would beat other strong ICCF players in what seemed to be a holdable position for the opponent. One of my favourites is this example:

 

It was an inventive approach by Nigel to create winning chances in an otherwise very difficult defence to crack. Those h5 lines by Black in the Sicilian Najdorf are typically quite effective in holding the draw for black in ICCF games, so for Nigel to grind this win out shows great patience and a deep treatment of each move on his turn.

DN: And an over-the-board (OTB) chess player?

WM: I have several favourites: Garry Kasparov, Hikaru Nakamura, Fabiano Caruana, Peter Svidler, Vladimir Kramnik, and any strong player that will trot out the Nimzo-Larsen opening like Richard Rapport. That is my favourite opening to play in blitz chess.

DN: Every how often do you have to change your machine?

WM: I've never changed my 'machine'. If we're talking about my computer for chess analysis, I'm using the same computer I've had since 2008. Modern correspondence chess isn't all about CPU power (although that obviously helps), but more about how you prepare each move and how you work with the engine in tandem with your own ideas. You can get lazy and buy a powerful computer, and probably get to a fairly decent rating on ICCF just parroting the engine, but that's not going to take you to the top. If you play too many games at the same time, you can very easily fall into this rut and stagnate because you don't have time to really investigate the position properly. I've been guilty of this myself, and I'm cutting back on my active games to give more attention to the ones I start.

To really excel at modern correspondence chess, you need to work with the engine, not just have it do all the work for you. This means you look at the suggested plans from the engine, play those moves out to a certain degree, look at and assess the resulting positions, go back and look for alternative moves, force the engine to look at your own ideas, force the engine to look at moves neither your or the machine considered, etc. You have to really work hard to make wins happen on ICCF.  One of my favourite analysis techniques is to start a notepad and section off candidate moves based on the type of move (i.e. only pawn moves, only knight moves, only bishop moves, etc.). Then I investigate and copy the research into the notepad. Ultimately when I finish my research, I have to decide what the best move to play is, and the best move is not always the highest evaluated by the computer engine.

DN: In OTB chess we have the idea, maybe a myth, that the opening repertoire reflects the inner values of the player. Is such idea also true for correspondence chess?

WM: Not at the stronger levels. You have to vary and try new approaches all the time, or your opponents get wise to your repertoire and prepare nasty traps you might not be aware of. I've done it myself many times in CC play, where I review my opponent's previous games and I see they are a life-long French Defence or Caro-Kann Defence player, and then I review those games and look for lines that they haven't been properly 'tested' against. It doesn't always pay off as a win, but it definitely adds to your chances. When it does pay off, it's very satisfying not just because you won the game, but because you put in the research to make it happen. It wasn't some fluke win over a bad line the opponent chose. You go in with the idea you're going to show them their repertoire is weak and you know it before even they realize what's happening. I suspect that is actually often the case in top OTB chess games too.

DN: Many players of the past were correspondence chess players. I think also today it's quite important, for improving to practice correspondence chess. What's your opinion about this?

WM: It depends on what you want to get out of modern correspondence chess. For me, I found it's a great tool to learn new openings or get a better understanding of main lines. As a result of my CC games, I'm more familiar with the Sicilian Sveshnikov than I would have ever learned as a passing interest in say a chess book or video on the subject. I've also heard and actually seen for myself that top OTB GMs will often use winning approaches taken directly from correspondence games, and then they get all the credit for coming up with this 'brilliant' new novelty to win the game. As I said, I've seen this happen on more than one occasion, and I say to myself "Uhh, no. He didn't invent that move, it was played 3 years ago on ICCF and we CC players are quite familiar with it".

DN: Russia in OTB chess has the highest number of GMs and titled players in general, is there a similar correlation with correspondence chess?

WM: I can't speak intelligently about the statistics of how many titled ICCF players there are for each country. But if you take say the top 100 CC players, only seven of them are Russians. So there's a pretty wide variety of interest all over the world. Since you play at home and don't need to travel, it picks up more of a wide birth of titled players.

DN: Could you explain how difficult it is to become Correspondence World Champion, in comparison with OTB World Champion?

WM: Again, it would be unfair for me to make such a comparison since I don't have the OTB experience it would take to become world champion in that arena. Having said that, I can speculate based on what I've seen. In OTB chess, you have to play well enough during the year in certain events to qualify for the Candidates Tournament, then you have to win that, and then beat the current World Champion. In ICCF, there's a longer cycle of elimination tournaments. First, there are the Preliminaries, then the Semifinals, then the Candidates Tournaments, and then finally the winners of the Candidates Tournaments play in a final tournament for the world crown. As such, it can take many years to become world CC champion, even if you knew you were going to win each qualifying event. This also means that there is more often than not a new champion rather than the same person winning it by defending their title against a particular person. To make the comparison more simple from an OTB perspective, imagine Fabiano Caruana's win in the recent Candidates Tournament; now imagine him having to win again in that very same tournament format an additional two times in a row to become the World Champion. That's about what it's like on ICCF, and it takes extreme 'luck' combined with hard work and YEARS of making the best moves to get there.

DN: The OTB chess world is plagued with cheating controversies in many different forms, be it sandbagging, or people hiding phones in the restrooms. Is there an equivalent form of cheating in correspondence chess?

WM: Well, as you know by now, I'm obviously playing in a CC arena that allows computer engine usage. That being said, online sites that prohibit computer engine usage are RIFE with cheaters. You just can't stop them all, and every top player list I've seen on a given site has people invariably using engine help in clever ways to not get entirely caught 'red-handed'. This is why I exclusively play on ICCF when it comes to CC. You don't have to worry about cheating or trying to stop it, because everyone can use whatever they have to come up with the best move. The playing field sort of becomes more fair, although there is the possibility of collusion of players to gang up on one person, like say in a country versus country event, but this is a rare situation.

DN: What is your Centaur opinion over AlphaZero? Would you be able to beat it in a fair match?

WM: That whole spectacle seemed to be a spun tale with a stacked deck. If AlphaZero were to play in an actual engine tournament like the TCEC, I seriously doubt it would even make it to the semifinals. In terms of beating it though, I couldn't say if that would actually happen. It could play something like the Petroff and easily draw and that would look like some sort of small victory on its part. However, I will say it will NOT beat me in a Centaur CC game. That's just not going to happen.

DN: Did you look at the new AI called Leela Chess?

WM: I've heard it's based on the same concept as AlphaZero, and I understand it effectively builds a database based on games it plays against itself. It's up to 2500 FIDE level now from what I understand. I really don't know much about it, but I can say again that these types of engines won't be able to catch up with the knowledge built up from CC databases, unless, of course, those games are added as some sort of percentage-based play factor.

DN: In chess, I have a database of favourite games, which I watch before a tournament for inspiration. Do you have a set of 10 favourite correspondence games?

Below: Wolff Morrow also won the Silviu Nenciulescu Memorial CC Trophy

trophy

WM: I never selected a top-10 list, but sometimes when I'm looking for inspiration, I'll review a strong ICCF player that I know will continue to rise in the ranks, and I look to see how he reacted to an opening I might be facing. The most important part of modern CC right now is how you handle the opening. And for that, you need to do a lot of research before you even play the first move. If somebody likes to play the Berlin Defence, you better know this ahead of time and be prepared to switch to an Italian game or play d4 instead. There are some openings that are quite 'dead' on ICCF in terms of having any real chance of catching Black in a bad position, so you have to be constantly aware that your opponents might just decide they want that easy draw.
That being said, the drawing rate is becoming higher and higher on ICCF due to the extreme preparation and various openings being effectively 'solved'. As such, I have to admit that modern Centaur CC won't last much longer before one can set up a repertoire and simply never lose a game based on playing safe defences as Black, and conversely playing the most effective side-stepping lines as White to prevent Black from doing the same to them.

Often times people speculate about whether chess can be solved. I can say with fair certainty, with how it's going on modern CC at the top levels, it's already very close to that point. It's my opinion that in 25 years a modern CC player, armed with a filtered database of CC games and the strongest engine, will effectively play 'solved' chess should they decide to approach the game from a "safety first" standpoint.

A big thank you to Wolff Morrow for letting me borrow his wisdom and deep knowledge of the exciting world of modern correspondence chess!

Final thoughts

What are the games from the correspondence database really teaching us? I think the short answer is: Strategy! Pure Strategy. Why? Because once you get rid of combinations, tactics, and missed checkmates, what happens is that in order to win the game you must overcome the opponent in a creative effort, where the human spirit is put at test, and where the machine's brute force is stretched to the limit, in order to achieve true beauty.  Stay tuned for more news from the world of Correspondence Chess, and this latest Chessbase Correspondence Database, which includes a lot of surprising games.


Corr Database 2018

Corr Database 2018 is an extensive collection of correspondence games, featuring classical correspondence games played by mail as well as email games.


Links


Davide is a novel chess aficionado who has made chess his spiritual tool of improvement and self-discovery. One of his favorite quotes is from the great Paul Keres: "Nobody is born a master. The way to mastery leads to the desired goal only after long years of learning, of struggle, of rejoicing, and of disappointment..."

Discuss

Rules for reader comments

 
 

Not registered yet? Register

Karbuncle Karbuncle 7/6/2018 07:19
@ John Deere, I previous had beaten a National Master in an OTB rated game using the French Defense as black. If I didn't have 'any chess ability', that certainly would never have happened. It doesn't matter anyway though, as you're just demonstrating the typical counter-arguments I get. "Well lets look at his OTB games from several years ago and dismiss any comparisons he makes to OTB chess skill". Heck, if you want to go that far back to make a point, you might as well look at my "OTB chess ability" in this game that got featured for the queen sacrifice at the end (keep in mind I was a mere 1400 when I played this game):

http://www.chessgames.com/perl/chessgame?gid=1492238
jjmolina jjmolina 7/4/2018 09:12
Chess ability (knowledge) is largely irrelevant when it comes to centaur play. One of the best players, Anson Williams, was a 1600 player himself (back when he played at all).
John Deere John Deere 7/4/2018 12:32
To be clear, this "open" tournament Mr. Morrow won with a "clean sweep" was a THREE-round, 13-player tournament with most players under 1600. He may be good at doing a brute force examination of many different computer lines, but to pretend he brings any chess ability to the table is silly. But hey, data science is surely a more valuable skill than chess.
jjmolina jjmolina 7/1/2018 12:14
Wins aren't guaranteed no matter what you do, so draws are the only thing you need to concern yourself with. You talk about the top, but when was the last time anyone won a game at that level? I'm talking best four players.
Karbuncle Karbuncle 7/1/2018 09:49
@JJmolina, it depends on your definition of 'enough'. It's enough to get a decent rating, but not enough to reach the top and be a champion. You don't get there drawing every game, because you'll never make the rating cutoff for invites. You have to show a level beyond that by beating those type of players.
jjmolina jjmolina 6/30/2018 06:54
Parroting moves is enough, as long as your PC is properly configured (book, engine, TBs). Enough to draw.
Karbuncle Karbuncle 6/30/2018 02:23
@Nay Tun, I'm over 2500 with a computer that is 10 years old. I've had this argument before, but the simple fact of the matter is it's a different skill set to be the best in each arena. A 1300 player might actually be very good at centaur chess compared to an OTB GM on the same computer hardware. It just depends on experience and technique. As I mentioned before, I have a winning record against titled FIDE players that took up ICCF, and not only were they all using computers, but computers more powerful than my own. There's just a lot of naive assumptions out there about centaur CC, where the inexperienced will try to compare OTB understanding and rating levels. You just can't do that because it's apples to oranges. Consider my own case: I quit OTB chess many years ago in the 1700s Elo range. Yet I constantly get dismissed over that rating as somehow proof 'any idiot' can use a computer to excel in modern CC. Now I do play blitz on playchess, and my rating is over 2200 on there. That's because I'm not bothered by nerves, and I've learned quite a lot since I last played an OTB rated game. That being said, it's still a different skill set than how you work with an engine. Parroting moves won't get you to the top. You have to know how to BEAT opponents that parrot engines moves to get to the top.
jjmolina jjmolina 6/29/2018 01:12
Just look at the results from last year's tournament, none of the big machines were able to beat any of the top centaurs, and they used markedly inferior HW.This isn't a claim, is a fact.
Nay Tun Nay Tun 6/29/2018 12:28
Most correspondence players are 1200-1300 OTB players. How good correspondence players means how good their engines are!! A 5-years-old kid running on 200 cores Stockfish will crush any centaur running 100 cores or 50 cores!

Utterlely useless claims!!
jjmolina jjmolina 6/28/2018 09:46
Correspondence chess can't be equated to centaur chess, too many people don't use engines, or they use them in a way that would result in many loses, at faster time controls.

Only the best players fall under the "centaur" category, many of them frequently play Freestyle.
melante melante 6/28/2018 08:56
@Karbuncle: well said! :)
genem genem 6/28/2018 04:20
'Centaur' Chess is an appropriate name for this modified form of 'Correspondence' chess. The combination of human + computer plays the best chess that humanity knows how to generate, better than either human or computer alone. This fact alone is enough to make Centaur chess important and interesting.

I wonder whether Correspondence games databases make it easy to distinguish between Correspondence games vs Centaur games - the distinction is important.
turok turok 6/28/2018 12:57
no interest in CC chess since engines became relevant-
Karbuncle Karbuncle 6/27/2018 04:57
@ tip4success, you still can comes up with novelties, but it doesn't come on move 15 any more. They happen in the middlegame now. And I disagree about losing relevance. I've seen plenty of top GM games where they used a novelty that directly came from modern correspondence chess, and would go on to win the game. Modern correspondence chess is just a different beast now. You use 'digital' creativity instead of purely biological. It appeals to some, and not so much to others. I just hoped to shed some light on the appeal it does have for someone that finds centaur chess fascinating, as well as the concept of just how good can a human-computer combo be. But most importantly, it allows you to match wits with someone that might be an OTB GM thinking they can step into the arena and dominate. For example, I've played many FIDE IMs and GMs that took up ICCF play, and I've not only never lost to them, but I actually have a winning record against. I take pride in that because no doubt every one of them had a more powerful computer than me. But it was my approach and strategy that proved either equal or stronger. That's exactly why I like centaur chess on this level. Psychological factors, eidetic memory, and genius-level calculation no longer determine who is the strongest, and rather instead is replaced by who uses better strategy and creativity. As I said, my computer is 10 years old. if it were about "pressing a button" as Nigel Short once opined, I'd never have gotten as far as I have.
tip4success tip4success 6/27/2018 01:53
CC chess has lost its relevance long ago, a bit of a sad thing, but you can't stop progress. Too bad I used to love cc chess. I remember coming up with a novelty in the 1970s. Good luck with that nowadays.
garyklien garyklien 6/27/2018 04:54
neat looking forwar dto hear more and seeing some gamews
Rama Rama 6/27/2018 04:00
re: "Against such a creature, AlphaZero would have had difficulties to win even one game."
Well that's pretty speculative, however at least you would have an opposing side to vouch for the outcome of the match.
SunriseK SunriseK 6/27/2018 03:06
The given study is not a mate in 4 but a mate in 6! Only in line when Black accepts the sacrifice, it becomes a mate in 4.
1