Practical Memory Palace Use

General Memory Chat
,
1

Hi all, hope you are well

I’m new to memory techniques and only recently created a memory palace.

My question is how does one use a memory palace to memorize large quantities of data? Like 1000 digits of pi for instance?

The premise of this question is that memory palaces only have a limited number of spaces and whta do you do when those run out? Do you simply repeat? Do you embed palaces within palaces?

Thanks for taking the time to read and respond

All the best
BigRed

2

Find a new palace, link to it from the old palace if you feel like it. For example, if I’m using my house and run out of space, I could switch to my old school and keep memorizing there. To remind me where to go, I would then place my old english teacher in the last locus of my house.

2 Likes
3

Three related but independent approaches:

  • Increase the density of loci (e.g. place 20 items around a room, on different items of furniture, in a clear order, rather than just placing 1–2 items)
  • Increase the richness of the data at each locus (e.g. using Person–Action–Object to encode perhaps 6 decimal digits, rather than only Object)
  • Increase the length of the memory palace, including by having multiple memory palaces and linking them together, as suggested by turducken above

In this way, a high school could have 30+ areas (classroom, tennis court, reception, etc.) × several loci per area × 6 digits encoded = 1000s of digits memorized. Using a few suigtable places can theoretically (and practically, if you put in the effort!) hold tens of thousands of decimal digits.

1 Like
4

What does increasing the richness of each locus look like practically? I don’t understand the person-action-object part

5

I am experimenting with the linking method along with the Memory Palace(MP). The linking method is pretty much making the same kind of absurd images you have to do with the MP, but you make the images connect in sequence with each other, like a chain, one coming after the other. What I do is I put the first image on the memory palace and then make the chain afterwards. They have to really interact with each other, otherwise you will forget them. I am saving a whole lot of loci so far, memorizing things that would take 10 or more loci in only one point.

But, there is some caveats: you have to be extra careful with the reviews, because if you forget one part of the chain, all the rest going forward goes with it. But you just need to do some extra reviews right after doing the images; there can be some confusion if you use the same images in the same chain or even in different chains, but as long as the other image that it is interacting is different from the other points of the link that it appears, there is no practical problem to be able to differentiate; and it is a little bit harder coming up with how to connect the ideas, but it is quite possible, and the benefits are, at least for my particular case, which is a gigantic quantity of notes to memorize, revolutionary. I wouldn’t be able to use the MP otherwise.

You can search online for more details about it, but it is really simple if you already are familiar with making the kind of images necessary for MPs. Just really make the images interact, and perhaps test it with non-crucial information first, just so you can train without so much stress. Most of the times, when you forget some link of the chain, just go read the notes about what you are memorizing and it will come right away. But then make the images stronger every link that you had a hard time remembering, otherwise there is a big chance it will go away next time.

The way I am now reviewing the images is:

  1. every ten links I do, I go back to the first of these ten links, and go throught them.
  2. When I have finished the chain I added to one locus in the MP, I go throught the whole chain, start to finish.
  3. When I finished for the day, I go throught all the loci and their respective chains that I added in that day.
  4. Aftwards, I review in this interval: 1 day, 2 days, 4 days, and its goes multiplying two fold, until a maximum interval of 60 days.

(OBS: I am, at the moment, at only a 4 day interval review stage with my chains, but after I reinforced some images that I had a problem - which you should probably detect at the third stage of this review process, since it is likely that even weak links are remembered after only a short time - my recovery was 100% in all the chains. I have used so far 17 loci , and they contain probably the contents of at least some 100-20 loci, had I used the MP without the linking method. So, with some little extra effort, one can really have a really big saving of loci. And I probably could have used even less than 17 loci, since I have not, so far, noticed any practical limit to the size of the chains you can make, but since I am memorizing information that I will need to retrieve in an exam situation, it would very impractical to make enormous chains; I need a fast recovery, and going throught a smaller quantity of links, and with chains divided in a logical manner, consistent with the subjects, improve the manipulation of the information.)

2 Likes
6

The quick and dirty for PAO: You have a list of people with associated actions and objects, each representing the same number (typically 00 through 99, but can be 0-9, 000-999). You memorize numbers by combining people, actions, and objects to encode the number on a single locus.
7 James Bond, Drinking, Martini
13 The Devil, handing over, a contract
22 Anakin, Constructing, Lightsaber

13722 would be the devil drinking a lightsaber. This example is entirely arbitrary, there are a few different opinions on how to go about constructing a PAO. Using the major method seems to be the most popular, aaaaand that brings me to my next point.

Up at the top of the screen is a Resources button. Click that, check out the PDF. At some point it’ll mention the PAO and bring you to a blog post about it, but the whole thing is worth reading through. It’s relatively small and will give you a complete primer on things like pegs, memory palace, and the major system.

7

Essentially, it means having more stuff going on in an image. So instead of, say, imagining the number 50 as Luke Skywalker and the next number as another person in another locus, you can also assign numbers to actions and objects. So for me, todays date (07/08/25) would be Green Lantern (07) throwing red lightning (08) at a brown trenchcoat (25). This way, I can use a single locus to memorize 6 digits instead of having to spread them out over 3.

If you want to memorize lots of things with the same structure (like fortune 500 company stats, contact data, etc.) you could also add ‘layers’ gradually like Nelson Dellis does here:

8

For me at least, the more familiar I am with a place the more details I can put in it.

And going on site in person helps immensely, because you can notice a lot of details giving you more loci in one place.

For example, I can place from memory 7 items in the kitchen of the Airbnb apartment I’m staying: fridge, pantry door, stove, window, counter, table, door to laundry room.

But if I am sitting in that kitchen at the moment I’m filling the memory palace, I observe more details and the hooks/loci are more numerous and stronger: left and right of the pantry door there are two embroidery things with flowers each one can be a hook/locus to memorize something on, but if I was only working from memory I most probably wouldn’t remember it was there at all.

Or on the ledge of the window there are one flower pot with pink flowers, one sculpture of a cockerel, one pot with a cactus (three loci instead of one). I imagine the first item to memorize planted in the flower pot and smelling like roses, I imagine the second item attacked by the cockerel, and I notice the branches, the spikes of the cactus, I imagine the third thing I want to memorize impaled on it. It’s a richer and stronger than just “there was a window with pretty things”.