The method

My method is not something that I invented entirely on my own, but I have combined various things and refined certain aspects of it. I will try to explain it here to some level of detail.

Memorizing digits

First of all, I don’t memorize digits, I memorize images or micro stories, which I translate into digits when asked for the digit at a specific position. Digits are quite abstract things for most people, also for me, and are not so suitable for memorizing as such. Therefore I have a system by which I can translate digits to words, and words back to digits. The system is a bit similar to the keys on the phone, where 2 is represented by a, b, or c, and where 3 is represented by d, e, or f, and so on.

Both my system, and the phone system, use the idea of a one-to-many relationship when going from digit to letter. So if we look at the phone again, and say we have the digits 346, then we can translate that to the word dim but we can also translate it to the word fin. On the other hand, when going back from word to digits, a certain word can only be translated to one specific sequence of digits, like fin for example means unambiguously 346.

This system has the advantage that although a certain sequence of digits might occur many times in the total sequence one wants to memorize, one can translate all these occurrences to different unique words which will reduce the risk for confusing them and mixing everything up. Another advantage is that one can select words that fit your personality or the specific place on the imaginary walk along which these words are put, which brings us to another central aspect of the whole system – the imaginary walk.

The imaginary walk

Memorizing words (or the image or micro story the words describe) instead of digits is combined with a system to keep everything in order, much like a computer memory has an addressing system to find a specific memory cell. This is done by establishing an imaginary walk. Even though I call it an imaginary walk it might still originally come from a real walk, like a walk in your home town for example. It is imaginary in the sense that you keep it in your mind and imagine that you walk it.

Along the imaginary walk you establish places or stations where you later on put your words (images, micro stories), and it is best to divide the walk into chunks of 10 such stations, and it is good if those chunks are naturally defined areas in your walk, if possible. Like, one chunk can be a park, wherein you have 10 stations.

I use my home town, Stockholm, for my imaginary walk. It is already established in my mind since I have lived here most of my life. At each station you store 10 digits, so one chunk will contain 100 digits. Thus it becomes easy for you to find the right station when you are requested to tell what is the digit at a certain position of the sequence you have memorized.

The translation system

The system I use for translating between digits and words (or letters) requires 2 words for each group of 5 digits, and consequently 4 words for 10 digits (seen as 2 groups of 5 digits).

When requested to tell the digit of a specific position in the sequence that I have memorized, it takes normally around 20 seconds for me to retrieve the digit. So, if someone asks me what is the digit at position 1423 then I will be able to say 6 within twenty, or sometimes just ten, seconds.

My present system is optimized in various ways. It gives me a good ratio in the one-to-many aspect I mentioned above, so that I can choose among many possible words when I “encode” my digits into words. It also divides the digits in groups of 10, which makes it natural for us and easy to count.

Previously, in one of my first attempts, one word represented 3 digits, and I had only one word at each station. Thus I had to divide a given position by 3 in order to find the right station, whereas now I just divide by 10 which is instantaneous and easy. And now I have much denser packing of digits into my walk, 10 digits per station, instead of 3. So where I could previously store 30 digits, I can now store 100.

I use nouns, adjectives, and verbs. So at one station, say outside a certain shop on a street, I might have the words “quick”,  “chameleon”, “funny”, “lawyer”, which make the micro story “a quick chameleon runs after a funny lawyer”. The word “runs” here is not used, even though it could have been used (as a verb) in a story. I have been trying to make rules for how the words are to be used, but sometimes I don’t follow those rules strictly, and it seems to work anyway. I am not going to get into those details here though.

So, in any case, the words “quick”,  “chameleon”, “funny”, “lawyer” translate back into 10 digits. And it is done like this. When translating a word to digits, find the first consonant, then find the first vowel after that and then find the next consonant after that. So, in “quick”, we thus have the letters Q, U, C. These letters translate to digits in a simple system where you look at the shape or similar aspect of the letter, like Q, for example, looks similar to the digit 0. So the first digit is 0. The U has two vertical lines, so it represents 2. And C looks similar to 6. So “quick” represents the digits 026.

Here is a table showing the translation I use from consonants to digits, the only exceptions from the rule that the letter should somehow remind you of the digit is R (5) and X (9). (Further down I will show a similar table for the vowels.)

  D Q 0
  J L 1
  F V 2
  K N 3
  M W 4
  R S 5
  G C 6
  T Z 7
  B H 8
  P X 9

Let’s continue with the word “chameleon”. Here we have the letters C, A, and M. The C again looks like a 6, the A looks like a 4, and the M has 4 lines all in all, so it is a 4. So 644.

Now we have totally six digits, but we said before that two words represent five digits. So here comes a trick. Look first at only the consonants, they give us the digit sequence 06 64. Then look at the two vowels, they give us the digits 2 and 4. The trick here is that you add 2 and 4 and you get 6, and that is the fifth digit. So the total sequence of five digits is thus 06 64 6.  I usually envision them as the dots on the five-side of a die, counting the positions as

  1   2  
  3   4  

But when counting positions, in this whole system, I always, without exception, count from 0 to 9 instead of 1 to 10. Things fall into place the right way if you do that. So try to get used to that. Thus the above counting will instead look like this, and here I also include the second word pair so that we get all the way from 0 to 9

  0   1       5   6  
    4           9    
  2   3       7   8  

So, let’s now go on with the words “funny” and “lawyer”. They give us the sequence 23 14 6. So all in all “a quick chameleon runs after a funny lawyer” gives us the ten digits 06646 23146. And I envision it like this

  Q U C       F U N  
  0 2 6       2 2 3  
    6           6    
  6 4 4       1 4 4  
  C A M       L A W  

Now, what if the sum gets greater than 9, let’s say you add 7 and 8, and you get 15? Then simply consider that to be a 5 instead, i.e. subtract 10 or look only at the unit position of the number.

The advantage of adding the vowels in this way is to create a good ratio in the one-to-many relationship, i.e. it gives you more freedom in picking words for your images or micro stories. It also has the advantage of making it possible to let the vowels represent all digits from 0 to 9. Not even the Swedish alphabet has 10 vowels, we have 9. And English has only 6, if you count y as a vowel. So adding them as explained above solves that problem.

Here are the vowels as I use them in the English language. Note, that here, in the English system, I have added a special rule, namely “Add 5 if both letters are U or E” or you could say “Add 5 if both digits are 2 or 3”. This rule is represented by the blue color in the matrix below. The purpose of this rule is to make the representation of all resultant digits more even. You can see in the left part of the table here in how many ways each of the ten digits can be represented. Each one can be represented in 2 ways, except the 0 which can be represented in 3 ways. Without the rule we would have a more unbalanced matrix.

  English, General
0 3     “Add 5 if both letters are U or E”  
1 2     “Add 5 if both digits are 2 or 3”  
2 2       O I U E A Y  
3 2       0 1 2 3 4 5  
4 2   O 0 0 1 2 3 4 5  
5 2   I 1 1 2 3 4 5 6  
6 2   U 2 2 3 9 0 6 7  
7 2   E 3 3 4 0 1 7 8  
8 2   A 4 4 5 6 7 8 9  
9 2   Y 5 5 6 7 8 9 0  

(Note that the gray part of the matrix is a mirror image of the opposite part, so one only needs about half of the matrix so to say. Also note that the only letter that does not really remind of its digit is Y.)

In the Swedish alphabet we do not have to use a special rule to balance the matrix, we get a good distribution without it. Here is the Swedish matrix (which can also be used for Norwegian and Danish).

0 5       O I U E A Y Ä Ö Å  
1 4       0 1 2 3 4 5 6 7 8  
2 5   O 0 0 1 2 3 4 5 6 7 8  
3 4   I 1 1 2 3 4 5 6 7 8 9  
4 5   U 2 2 3 4 5 6 7 8 9 0  
5 4   E 3 3 4 5 6 7 8 9 0 1  
6 5   A 4 4 5 6 7 8 9 0 1 2  
7 4   Y 5 5 6 7 8 9 0 1 2 3  
8 5   Ä 6 6 7 8 9 0 1 2 3 4  
9 4   Ö 7 7 8 9 0 1 2 3 4 5  
  45   Å 8 8 9 0 1 2 3 4 5 6  

For Swedish speaking people, the idea with Ä representing 6 is that the sound of Ä is similar to E in sex (Swedish for six), Ö could be seen as similar to sju because of the word sjö, Å is åtta because of the å in åtta.

Using computer

For selecting appropriate words when I create my images or micro stories I use a computer program that I made, which is equipped with a Swedish dictionary. The program presents suggestions for words to me, to help me find good words.

The process to memorize 100 digits takes about 30 minutes, if the process of selecting the words (with the help of my program) is included.