Memorizing Chess Openings and Classic Games

I would like to go about memorizing a good number of “tournament” openings and a handful of classic games. Does anyone have any experience memorizing chess play? Would you be willing to share some tips, techs or experience?

Thank you!

I’ve never used mnemonics to memorise chess openings/games. But it’s a fairly simple idea isn’t it? The pieces are the hard part to code.

A chess board is 8x8, e1 to h8, the squares being named. Turn each square into a number according to their position in the alphabet. A = 1, B = 2, C = 3, D = 4, E = 5, F = 6, G = 7, H = 8

So e4 would be 54. You use your image for 54 to represent that square. Let’s take a King’s Indian Attack opening:

  1. e4 e6
  2. d3 d5
  3. Nd2 Nf6
  4. g3 c5
  5. Bg2 Nc6
  6. Ngf3 Be7
  7. O-O O-O

As you can see we can easily code for the squares into numbers. This is easier for you since you use the Major system. You could memorise this via the story method, if you don’t want to use loci.

To code for the pieces, I’d have the square to move to as the main image, taken from my library of pre-made images for 00-99, then add in a feature for the pieces. I’d do:

Pawn: no feature. Lone object
Knight: Horse
Rook: Bird
Bishop: A priest
King: Throne
Queen: Crown

For castling, I’d just imagine a castle. For Queenside castle, I’d imagine…hmm…not sure what I’d imagine for this. You might want to sit down for a minute to think of a good differentiator between Queenside castle, O - O - O, and Kingside castle, O - O.

So for this game, in the major system using loci:

(Link method combined with loci)
locus 1: Image for 54 (e4), linked to image for 55 (e5)
locus 2: Image for 43 (d3), linked to image for 45 (d5)
locus 3: A horse biting down on your image for 42 (Nd2), linked to horse biting down on image for 66 (Nf6)
locus 4: Image for 73 (g3), linked to image for 35 (c5)
locus 5: A priest praying upon/holding/doing whatever sticks for you, to your image for 72 (g2), linked to…

you get the idea. :slight_smile:

I don’t see why you couldn’t memorise openings by rote. It would save a lot of trouble, I think.

Of course, if you want to memorise multiple chess games this would swallow a lot of loci, hence why I suggested using the story method. The story method just involves you making a story in your head to remember things. Might want to go read the thread “Loci vs Linking” to clarify the distinction I’ve drawn between the link method and the story method.

I use a PAO system, so that’d allow me to do do 3 moves in one locus if I used the same method as above; things would get cluttered on the locus, but I think I’d be able to do it!


i memorize four short games (<10 total moves like 5 moves each) by rote could recall day couldn’t remember a thing like 2 or 3 moves max and they were all mixed up.even though i could play the games perfect on my computer hours before.this is probably not very helpful.i guess i should have completed my images for the Dominic system.

Hype - Very very awesome. Thank you! All great suggestions. I hadn’t thought of converting the A,B,C to 1,2,3 and that does let me use my major system… awesome suggestion.

My first thought was journey because dwelling on a chess game I’ll want to be able to move back and forward through a game with ease. I imagine that would be harder with a story. All in all - that was an awesome post you referenced me to and I learned a lot from that one. Thanks!

To agree with our friend Martin, if rote memorization worked for chess openings, I would know a lot of games :wink:

Thanks again Hype.

Oh, I didn’t mean rote for entire games. I just meant rote for the opening sequences, i.e. Sicilian, accelerated dragon, King’s Indian Defence, Scotch game, French defence…you know!

I have contacted several mnemonists – Dominic O’Brien, Anthony Metivier & Chess GM Raymond Keene - and none could give me a comprehensible answer,
but saying it could be easy to apply standard techniques to it but nothing precise beyond that…

Here’s an email exchange I had with Anthony Metivier discussing him wanting to eventually develop one such technique:

I was thinking that if we can embed small micro element like a deck of card on a desk somewhere along a journey, or a library with shelves or …
to have additional memory palaces – a comment you’ve made about memory palaces connecting on another level in “renovating MPs”.

Why couldn’t we put a MP (or several) imbedded in a “main” memory palace: a bit like the pack of cards or shelves are separe micro-journey along the main MP…?

I was thinking of something like a computer (or several) with virtual googled somewhere in my MP journey, that I could envision putting google on and been taken on a virtual reality tour of another MP and then when that MP is completed, I could take glasses of and resume my journey along my loci in my main MP…

This would be particularly useful for opening learning in chess due to the many many variations that branches out, but then you have to go back to root of moves choices to go to opening mainline… You could even have colored sets of googles and have 1-2-3 google on same computer and you choose to go through MPs or not according to move you play, since if you have 3 moves choice on move 9 and play one of them, you don’t need to know the other possibilities that would have taken place if you digressed in another move. Such sub-MP could have sub-MPs of their own to include further variations…

The concern is that – I am building an opening repertoire at the moment – and when I look at the amount of variations visually, it would take a ton of MPs to encompass all…
So I was thinking of having major system and SEM3 or peg system (alpha and numeric) – a bit like Chessbase software classify variations - like systems for sub-lines and variations from main line which would be on main MP…

A bit like:

Variation 1 (red)
Sub-variation A (red)
Sub-variation AA
Sub-variation AB
Sub-variation AC
Sub-variation B (blue)
Sub-variation BA
Sub-variation BB
Sub-variation BC
Variation 2 (blue)
Variation 3 (yellow)

I just through that transposition could use chromatic circle of colors and be of that color mixed with the color it is transposing into of same variation but different line (likes-v AB transposing to s-v BB could be purple “red and blue”). Of course this would cause problem if more than 3 variation, but it is rarely the case in sub-variations, unless you explain bad move and replies to them which could all be of another color to indicate bad moves and reply lines to good lines…

I am aware that the multiplicity of chess variations for even just one opening cause problem with the need for a ton of memory palaces, making it convoluted and not realistic if you want to include stem games of the opening as well…

I came across something on Ben Pridmore’s website that is called Mechtnon System:

I think that using the ECO theory tables in that kind of house system (one house per opening) could do; or something like Buzan’s SEM3.

These are easily multipliable to have lots of stuff into them, particularly the Mechtnon System.

They are not memory palaces, but I think that is what makes them perfect to memorise so intricate variations as chess opening ones…

I would just recommend normally studying, analyzing the openings(why is this move made?), and playing them, perhaps in some blitz chess on or some place like that. After playing them for a couple weeks/months, you know pretty much all the possibilities for the first 10-15 moves. I used to play chess a lot, and my father was one of the best players in Poland, so yeah.

I would stress analysis. Why is the move made? Why not a different one? What advantages does it give you? What does it open? Which squares are now protected or unprotected? Could also use some chess program to play against, one that analyzes the benefits of each move mathematically.


I have thought about this too. But not planning to get back into chess for a while. What you should do is use Correspondence Chess notation. No letters numbers only. No mention of pieces, just the squares. See

The bottom of the white board would be 11 21 31 41 51 61 71 81

e4 e5 would be 5254 5755 I think you find these numbers already in your memory systems.

I agree with you. I have been playing chess for many many years and I’ve never needed a memorisation strategy for memorising them. Just analysing the situation, why the move, what can we done and what happens if I don’t do that move. Also, I always think in crazy sacrifices, that could possibly be made. I use to play aggressive, and this also may give me another trigger to memorise some plays.

dzikijohn, I don’t agree with you in this one. Sorry. I think that the common sense is the best way to memorise chess, it is a logic game. I mean, I would find easier to memorise things like, bishop attacks king, or knight goes out to threaten the pawn guarding the king… Rather than counting the square. I think that for the squares, if you practice you can get until the point that you can tell which is the square (notation) in a very few time, so I don’t really see the point. Whereas with you method, you have to encode 1 by 1, chess most of the times has combinations, that you can remember easily even if you don’t know in which absolute square is occurring…

But It is a nice topic! Go ahead and comment what you think of this! :slight_smile:

I also played chess from times to times and I am interested on the subject. I agree with Batman, I believe a deep understanding approach is more useful and reliable for long term outcomes than raw memorizing.

That being said, I still believe method of Loci can still play an interesting role in learning chess opening. Why not a very light encoding in order to detect recall errors and hint in case of blackout ? I worked on a very simple code for this purpose with only 1 move=1 digit. I must precise that I did not took the time to make use of this, then I have no feed back to give, but the idea is worth to be suggested here I think.

  • 0 = king & castle
  • 1 = rook
  • 2 = a to d pawn
  • 3 = e to h pawn
  • 4 = knight moving in a to d area
  • 5= knight moving in e to h area
  • 6 = bishop moving in a to d area
  • 7 = bishop moving in e to h area
  • 8 = queen moving in a to d area
  • 9 = queen moving in e to h area

A more optimal code could be imagined (for example reallocate 5 & 6 when bishops are taken out), but one advantage of this script is that it can be easily generated with a script which reads a pgn file.
EDIT : I just decided to code it in python based on a small library which parse pgn files. This is available there

marceltella…I am talking about memorizing opening variations. Not how to play the game. I don’t think you have played chess much, have you.

Hello dzikijohn! Sorry, I was not trying to attack you in any way, I was just giving my opinion. I have played chess my whole life.

Hi All,
Before algebraic or figurine algebraic, the old system (Descriptive Notation) was very visual. P-K4, P-K4, Kt-KB3, etc. When I started to learn chess, this was the way you did it. Thank God for the Germans and their wide use of algebraic notation, which allows one to read any game published in any old journal.

I think you need to study stem games first, and you also need to focus one one opening. The Ruy Lopez, for example, is vast! So maybe you want to study the Marshall variation or the Keres or whatever. Stick to one.
Alexander Kotov has another type of memory (analysis) method: the Tree of Analysis, which he describes in Think like a grandmaster (

Modern chess had its birth (arguably) in 1953 at the Zurich candidate masters match. If you study and memorize those games, ones that come after and variants will come easier.

Focus on that book.
Focus on one opening for white and one opening for black.
Learn it in descriptive as well as algebraic notation
Sketch it out
Visualize the various images of the games at key moments
Have fun!
Repeat, repeat, repeat.

sorry marcel…didn’t mean anything…just seemed that you were talking about trying to apply mnemo to tactics…which is useless I agree…but, to opening which people do memorize it would speed up their learning…of course, people spend too much time in beginning on memorizing openings…when they should be studying tactics and endings…so if they could do it faster then that would be good…

hey! Yes, sure, I understand your point of view… I just read in the book of Joshua Foer something that I was referring to and that I have always knew but I have never been able to express or explain. In the book, he says that people who play chess, often look at the pieces, not as individual pieces but as structures of tension, where are strong and weak points. And if you think of it, it is much more easy because you are codifying several pieces at once most of the times. I may read that chapter again, it is very interesting. It is kind of what I was talking in two posts above.

It is well known that if you show a chess grandmaster a position from a real game for a few seconds he can memorize the board. But, if you just place the same number of pieces randomly on the board they can’t memorize them.

Sure, it would be really interesting to know how they actually codify the pieces, this could give some ideas to apply to other fields…