ChessMood 2 months ago

Article: How Grandmasters Memorize Opening Variations

Hey Champions!

We have this topic in our Blog:


If you have any questions, comments or you just liked it, feel free to share your thoughts here. 

David Flynn

David Flynn 2 months ago

The first question that needs asking is how do we remember? This needs to be considered.

Why do we remember our own phone number over say the number of our best friend (because we often repeat ours to others, and we autodial the friends number - before mobiles I could remember friends numbers). Repetition is one aspect.

Why do we remember what we were doing when someone very famous died in tragic circumstances (Kennedy assassination or Princess Diana for example) but we can't remember what we had for dinner last Tuesday - novelty, emotional content, standing out from other events.

When we can't remember where we left our keys, but work out where they are because of what we were doing when we last had them - reasoning

When we're completely unaware of a song we once heard many times, but someone hums the tune or mentions a lyric, and suddenly it's back, including where you were and what you were doing when you heard this song 20 years ago - association. Smell is very powerful in this aspect and can bring back a flood of memories.

Our memory is associative. Put simply, a key unlocks a memory. If the key is unique, and strongly associated to the memory, you will often remember (I can remember what I ordered last time I went out for dinner Friday, as well as a few weeks ago the previous time, but not what I ate on Sunday). The more keys to a memory the more likely any one of them will help you remember it.

The problem with chess, and one for non-professionals is that positions look similar. The flow is linear (I go here, they go there) and positions are slowly transformed. Not every move can be ascribed a reason that is very memorable, and even if so, many others will look like the same reason can be applied. Often several moves look good, but only one leads to positions studied. Sometimes these other moves are deep mistakes. The d6 anti-Sicilian is a problem with both its sharp nature and branching. Professionals have the advantage that more of their brain will be wired to recognise chess positions, and connections are stronger because they spend more time with chess positions. I can't tell similar looking cats apart, but I bet vets can. I struggle with names because I don't meet often and work in such an environment that names are important, yet teachers can remember all the pupils in several classes they take.

Understanding the moves is undoubtedly the most important piece (even if you did remember the move, what is it doing?), but this is still memory and memorisation. Here it is using reasoning as a key to find it. If you forget the reasoning, you may or may not be able to work it out from scratch. No one is going to tell you that you learnt a reason on this move. Not all moves have reasons, some moves are natural, fashion, the right move because others are bad and so on. Perhaps that's not such a problem for a GM who will also use their more developed chess skill and intuition to find the correct move (or at least dismiss some of the candidate moves), but it's a big problem for club players.

What's described by playing many games with a training partner is repetition (as well as active learning) which is a key aspect of memorisation. However when you stop repeating, things fade, and for anyone who has tried the chessable approach will know this problem as well as the overburden of workload. You can't just spend all your time repeating memorising openings, and you still make mistakes. Sometimes there is a lot to be said about cramming before a game if your opponent plays a specific line. It's short term, but it does the job.

Another memory difference between professionals and amateurs is chunking, in which a famous experiment showed that strong players can reconstruct chess positions better than lower rated players. However if a random position is set up, they didn't perform better. Thus the idea of chunking (having fewer bits of information to remember by being familiar with pawn structure and king positions for example) was seen as the explanation. Add in more experience, superior knowledge a master will have less they need to remember (compensated by them learning more difficult openings).

I get the idea of the article though that just trying to repeat the lines again and again, rewatching videos, or worse playing through them once and assuming that's it before looking for the next shiny thing isn't enough.

I don't yet have an answer, but aside from the article's suggestions, I would add:

reduction (reduce the load down to a few critical positions which can be repeated or other tricks used, leaving out all that you can find over the board). Less information is better. Use flashcards to repeat.

Cross referencing - often moves are chosen to keep similar positions to other openings we know, and the ideas will also be similar)

Frequency/application (study games from those openings so you see the position often in other contexts than your opening pgn)

Patterns and the rule of 3 (we take notice of things that form patterns or repeat, and 3 times or more is an indicator of a pattern). Especially if it's slightly different (a variant on play it as Black is flip the board left for right, the positions look different, and sometimes new things are seen).

and if all else fails, mnemonic devices (sparingly).

The more things you do, the more likely one will succeed when you need it. Just don't be like the guy that tried to remember all the lines of Understanding the Chess Openings using mnemonics, because even if he pulled it off for all 200 or so variations (theoretically possible but difficult as well as to maintain and time consuming), it probably wouldn't be a lot of use practically. Sometimes it's just confidence we've done all we can and we'll play good chess at the board.

Edo Tokyo

Edo Tokyo 2 months ago

@David_Flynn About the part:

"What's described by playing many games with a training partner is repetition (as well as active learning) which is a key aspect of memorisation. However when you stop repeating, things fade, and for anyone who has tried the chessable approach will know this problem as well as the overburden of workload. "

I do not think that you are right at all with this. Playing games with a training partner and chessable are like mixing whisky with nutella, they do not belong together.

Playing serious games for training means to reacreate tournament conditions, serious chess, where you are in front of the chosen position and you play, let's say for 45 min. or 20 min. each a game. Then when you are playing seriously and thinking by yourself you learn, you discover a lot of nuances in a position. You make wrong decisions and your opponent gets the better hand, you analize why, then you play again and discover new things. After 5 or 6 games you get a very good feeeling for the position, you know naturally where the pieces belong, what is the best plan, etc. You play and enjoy chess with a training partner. I have been playing weekly training games in an opening decided beforehand for years with a training partner OTB, only after joining Chessmood and finding the right repertoire for me I began to win in a constant pattern.

With chessable, what do you  do? You are alone. Rehearsing lines that somebody wrote, trying to memorize all the lines (most of them worked out with engines, that many times people do not know how to handle them well). You do not think about the position you have in front of you, you do not feel the tension of playing the game. When you play a real game, you want to win. When you do chessable you want to remember the move that the author decided to get more points and keep up your streak...

This is my opinion about this, and I think that if you really want to progress in chess we should all stay away from Chessable and I am saying this after having a streak of 595 days doing it every day before joining Chessmood. During all this time my rating stayed the same, no matter how many courses or hours I spent in the 100 endgames or in openings. You can check my profile and see for yourself: https://www.chessable.com/profile/Edo_Tokyo/

The real progress I made was after joining Chessmood and finding the most coherent repertoire with the ideas interconected in all the openings and seeing the games played in the seminars and streams.

I bought many books on Chessable, and some that are made by a 2100 FIDE, who does not even provide his real name, in chessable he goes by the name "chessforlife", with lots of mistakes, still he talks like he owns the world and his opinion (engine checked) is definitive. In concrete in our line on the Scandinavian with Nf3, when Black goes Bg4 and we play h3, he says:  Quote "I was surprised to see this insipid move recommended in a couple of sources. It's very unimpressive". This move that Aronian recommended to Avetik, as he said in some stream, is a key move in our repertoire and he calls it insipid... The line showed is bad for Black on the long run, but he does not understand that... He is recommeding the reader (that probably does not have enough level to discern about this)  a line that is not good, while saying that the move is bad and hiding his name...

Sorry if I mixed the topics, just do not compare training games with partners with chessable... These are complete different worlds and I recommend you wholeheartely to play as many OTB training games with a training partner as you can...

David Flynn

David Flynn 2 months ago

I agree re: Chessable if all you do is repeat, particularly without any thought to what the moves actually mean, especially on some kind of autopilot. This sort of repetition cramming can be useful just before needing it but it has to be on top of understanding. Doing it without anything else won't bring good results (and when the moves end or deviate then what?), and chessable's training system is very much a one trick pony.

The point I was making is repetition is a key aspect of memory, and without repeating things will fade. Spaced repetition was well known long before chessable. It is only one aspect though.

Training games are far higher quality because you're not just repeating moves, you're actively using the thought process you would when using them. You can also try to fix the thought process if it's leading to inferior moves. If you're learning a line this way you still need to know many of the moves and ideas beforehand. You also still have to deal with similar looking positions, different actions.

Material quality is a separate issue for Chessable and while you have top GMs making courses, there are also some not as good as well as choosing openings that fit together and are appropriate is left to you. Chessmood's repertoire is a far superior offering IMO. However, I'm quite a fan of 100 endgames, the main reason being the choice of moves and the responses are narrow which isn't the case with openings - it's also based on a pretty good book with its only flaw being the last chapter.

Kevin D

Kevin D 2 months ago


I have a feeling that chessforlife was once a ChessMood member. Also when an author underestimates our lines that is actually better for us, as they will be less prepared to face them effectively over the board. 

Llorenç Boldú Zabih

Llorenç Boldú Zabih 2 months ago

The problem of repertoire is that every student is different. I remember a Kramnik quote where he says that he was convinced that the way one plays reflects his character so one should choose (perhaps with the help of a coach) the openings where he feels more confortable.

About Chessable my favourite course is GM So explaining how he outplayed Magnus in the 960 WC (I am refering to his thought process). I think 960 is a great tool for learning. At the end what you do? Looking for weak squares in yours and your opponents camp. Finding good diagonals for your bishops, nice outpost for your knights, the pawn structure (few islands but that offer some open files for your rooks) and what is of most importance, grab space. So in 960 we must think about development just like in standard chess, but beginning in move one instead of move 10 or 20.

I remember when I started with 960 like most of my unexperienced opponents, we started to play the pawns to open diagonals for our Queen and Bishops, but as we swimmed in unknown waters we moved the pawns only 1 square ahead. Then I asked myself why I move 1 square when in standard chess normally we move 2. So I started to move 2 and grab space from my opponents and began winning many games.

David Flynn

David Flynn 2 months ago

I think Fischer Random / Chess960 would make a good Chessmood Saturday tournament for that very reason.

NB. I did quite well at 5-a-side chess (it's a 5x5 version with no kside flank pieces) for that very reason even though I've not learned any openings where 6 of your men and much of the board has gone.