‘You never know when you write a book – often, you don’t hear from the folks who have read it, and then…’ He shrugged. ‘Then you think: “Well, I guess nobody’s read it after all.”’ Alexander McCall Smith, The Limpopo Academy of Private Detection

I am grateful that the quote above does not apply to me, for several readers have written to me since I published The Chess Memory Palace at the end of 2022. On the principle that ‘if one person voices a question, many are thinking it’, I have answered some FAQs below.

  1. How did you come to write the book?
    I became interested in mnemonics at university, after reading Tony Buzan’s The Memory Book. I built my first memory palaces for law and economics exams in 2013, which turned out to be the first time of many. Sometimes learning facts with memory palaces and Anki flashcards was so quick that it almost felt like cheating. So it was natural to apply the techniques to chess.

    Surprisingly, there weren’t any good published systems for chess mnemonics, so I set about making my own. The first system I made was relatively complex (using vowel sounds and a convoluted piece-priority system), but as I built more chess memory palaces over the next few years – and particularly as I started explaining it in draft articles that eventually became the book – I found ways to simplify, until I felt the system was worth sharing.

  2. Why did you design Picture Notation the way you did, rather than using PAO / picture transformations / word length / other mnemonic codes?
    There is a lot to say about this, so I have written a blog post on the design of mnemonic systems.

  3. Is there an online forum to discuss The Chess Memory Palace?
    Unfortunately there’s not a good discussion place online at this time. The best single conversation is on the Art of Memory Forum, but this is not very active.

  4. What is the largest possible chess memory palace?
    It’s not obvious what the limits of memory palaces are. I think it’s fair to say that the limits of linear memory palaces are very large. The largest memory palaces historically probably belonged to medieval monks or non-literate cultures, but modern records for memorising numbers are easier to quantify. The record for decimal digits memorised in one hour is over 4600. The official record for memorising digits of pi is 70,000 (unofficial record 100,000). Assuming these were all memorised using PAO chunks of 9 digits per composite image, this suggests a record of more than 500 composite images in an hour(!), and around ~7.8k composite images to memorise pi.

    I expect chess composite images are easier to memorise, because they have 3 elements (first picture word, second picture word, location) which is fewer than PAO’s 4 elements (person, action, object, location). On the other hand, chess memory palaces have a branching structure, which may be harder to memorise (and design) than the linear structure of traditional memory palaces. There is unfortunately no research on how much difference this makes, as almost all memory literature assumes you are dealing with data of a linear structure.

  5. Can you use one set of memory palace locations for two openings?
    If you try to remember two composite images in one location, it is possible that they will become mixed up in your memory. (This is discussed in the book on page 61, and Note 6 to Chapter 3 on page 173.) Even if you can reliably store four picture words in one location and keep them in order (two from repertoire A then two from repertoire B), there is a bigger problem, that the branching structure of the two repertoires will be different. I use a new setting for every memory palace.

  6. Should you choose a setting that matches the opening name, e.g. memorise the French Defence in a French setting?
    Using a French-themed setting for the French Defence sounds fun (which you shouldn’t underestimate, you will remember things better when you enjoy it), but I don’t think would make much practical difference. If you have a large number of palaces, then you might need a method to remember which palace is for which opening, but I haven’t found this too hard in practice. In one case I imagined grand prix cars racing past the start of a palace, to signify it was the Grand Prix Attack.

  7. Should you combine mnemonics with conventional memory?
    Most likely it is best to combine mnemonics with conventional chess memory – perhaps by memorising longer/rarer/sharper lines with mnemonics, and other lines with conventional memory. I am not sure of the exact best ratio; this will be discovered through experience. (And of course, it is necessary to understand your moves and the middlegame plans, so you are not lost when you leave your memorised lines.) It’s important to know at which position your chess memory palaces start, so that you can access them when the right position arises on the board.

  8. Can you use picture notation to memorise static chess positions?
    Picture notation codes chess moves, rather than static chess board positions. If you want to memorise a chess position, you will need a different mnemonic system. US memory champion Nelson Dellis made a video with two methods to try.

  9. Do you have images of the picture words to share?
    I don’t have any sketches except the ones in the book. I just imagine the scenes in my head. Like an actual memory, you don’t need a photo-like picture: you just need to imagine enough to make the scenes memorable. If you like, you could use one of the new AI image generator tools for inspiration.