01 Language and writing

The chapters Numbers and Language and Scripture blur a bit, but this is as difficult to separate as it is to avoid because of the subject matter. The emphasis here is a bit more on more or less meaningful text that has hidden things in it.

letter values

Nonetheless, we start with naked letters, which in the end are almost always supposed to result in numbers somehow. The most common and easiest way is to simply look for the alphabet values, i.e. the position the letter has in the alphabet. A=1, B=2, C=3, … The word Nina thus has the letter values 14-9-14-1. Sometimes you also have to count through the alphabet starting at 0, in which case it would be 13-8-13-0. If necessary, you can also count backwards from 0 or 1. You can do all kinds of things with this. Sometimes you need the sum of the letter values (38 in the usual counting starting at one). Sometimes you need the cross sum (the digits of a number added together), sometimes of the individual letter values, sometimes of the total sum. Sometimes the simple transverse sum (digits only added), sometimes the iterated one (calculate the transverse sum from the result of the transverse sum until the result is one digit). Sometimes you have to multiply the determined digits together or do other arithmetic operations, which hopefully result from the listing. Sometimes you only need the values of certain letters, for example, all initial letters or all capitalised, italicised, written in a certain colour.

paragraph value

If a letter is different from the others (colour, style), its position in the word can also be the key. For example, if the h in the word “letter” is in italics, the number 4 could be sought. If more than one letter stands out, a binary code could be sought. This “Testwort”, for example, if you take the bold letters as 1 and the normal ones as 0, would result in the binary number 0110 0110. If it were a binary number, it would result in decimal 102. Since this “binary word” conveniently consists of 8 letters, i.e. also of eight bits (thus one byte), ASCII would also be a great candidate. There, 01100110 / 102 is assigned the letter “f”. Which – and here a mystery circle closes – corresponds to the letter value 6.

Sometimes the solution to the text puzzle lies simply in the number of letters in one or all words. Or the number of syllables. Consonants. Vowels. Maybe you just have to count certain letters? Or ones that have certain properties. For example, strokes going up, like the l and the h. Or reaching down, like the p and the g. Maybe you need to count the number of round areas, like in the O or in the d. Such letter or even number areas that can be coloured in have already been the key to one mystery and another several times, which I was able to erase from the map. Perhaps one has to juxtapose such things with binary coding again. If one has determined a large number of numbers in such a way, it would also be conceivable that one simply has to consider the numbers found as even and odd numbers and accordingly translate them into a binary code.

You could also count the number of letters until the next space or punctuation mark. Or use only the first, or third, or even eighth letter of each word to decode it (to try this, it’s a good idea to write the individual words one below the other).

Speaking of the first letter, a certain but long-deceased John Laird McCaffrey has this wonderfully handsome epitaph in Montreal:

Free your body and soul
Unfold your powerful wings
Climb up the highest mountains
Kick your feet up in the air
You may now live forever
Or return to this earth
Unless you feel good where you are!
“Missed by your friends”

His friends don’t seem to have been particularly good ones, if you read the first letters of each line from top to bottom.

Numbers as words

However, this epitaph did not yield a coordinate, but we are still looking for digits. Or are we looking for words after all? Of course, numbers can also be expressed in letters, thus neatly stretching out encrypted texts, which inevitably become more difficult to decipher. With a little logic, I have often solved a 0-9 series very easily and bypassed the actual puzzle. Provided that the number you are looking for is the complete coordinate, which is known in the first 3-4 digits at the front (through the coordinate where the ? is listed – according to geocaching.com guidelines, the cache itself may only be 2-3 kilometres away from this point since a few years). But if it is letters as text, “one”, “two”, “three” or even worse “fifty-two” or “twenty-three”, it becomes much more difficult for the puzzle-solver.

Such written-out numbers or numeral words can also be placed in completely unsuspicious texts. I am TWICE observing the activiTy of the prACHTful VerEINscoin. Well, to make a nice mystery text out of it, it needs some fine-tuning, but hopefully my example is just enough to illustrate the principle.


Are you only finding letter salad? Perhaps this is only written backwards and/or the gaps between words are not where the eye would like them to be to read? Or maybe the first letter stands for a word of the north coordinate, the second for one of the east coordinate, the next again for north and so on. This code here: “znwuelilunnudlflünnefuznisgeacchhstuunndddzrweainszsiiggderienihzuenhdnert” results, such enttüddelt: zweiundfünfzigachtundzwanzigdreizehn und nullnullnunsechsunddreissigeinhundert.

The railfence or hunter fence cipher works similarly, except that here the text to be encoded is written diagonally downwards and then upwards again, and is used line by line at the end.

Fifty-two fifty-eight becomes this garden fence and with it this coded text: zdgnziwnfiaudniezeunzctzagrenifhwdh.

It works similarly with “ploughing”. The text to be encoded is written down normally, but a grid of X characters per line. Then this multiline text is taken line by line from top to bottom, and then, when you get to the bottom, from bottom to top. It’s like ploughing a field. You could also have called it “mowing the lawn” 😉 .

Easier to unscramble and scramble is the trick of always pressing the respective left or right, top or bottom of the printed letter on the keyboard. Fifty-two, shifted to the right, becomes ueroimfg+mguoh .

Mobile phone keyboard code (vanity code)

Maybe it’s just the good old (and far too often forgotten by me) mobile phone keyboard? Here, for example, there are a, b and c for the digit 2, while the 3 is occupied by d, e and f. The jumble of letters Kcbadnz then stands for the pretty northern coordinate: 52 22.369 . This form of encryption works in both directions. If you have a numerical code that has no 1s, but frequent repetitions of digits, it certainly does no harm to consult your mobile phone. 69997777833777999 gives the word mystery.

For advanced thumb typers, there is also the T9 text recognition variant. One uses the suggestion algorithm of the mobile phone. However, this is rather bad because it is seldom unambiguous, the internal dictionary can be extended by the user and thus bring false suggestions and not every mobile phone uses T9 (Motorola, for example, cooks its own text recognition soup). Nevertheless, this encoding is sometimes used.

Roman numeral script

A chronogram is a sentence or phrase, usually in Latin, in which the letters present, which are Roman numeral symbols (I, V, X, L, C, D, M), give a year (usually the year of construction of the building). Often, the usual Roman subtraction rule is disregarded and the numerical values are simply added up.

Using the subtraction rule is to avoid writing more than three of the same character next to each other, so that the next larger character is set and the smaller one is written before it, and must therefore be subtracted from the larger number. Sounds more complicated than it is. Normally, Roman numerals are written from large to small. 2013 is expressed in Roman numerals: MMXIII M=1000, X=10, I=1. Everything is written down nicely sorted by size, so here we only add in any case. M=1000 + M=1000 + X=10 + I=1+ I=1+I=1 = 2013.

I=1, II=2, III=3. IIII would be four times one character, and is therefore expressed in the subtraction notation IV. The one, small character before the large = subtract! IV therefore means 4, VI therefore 6.

Of course there are a lot of exceptions, besides the subtraction rule is more a guideline, not a law. And the chronogram is mostly just addition anyway, and even the order of the Roman numeral symbols is often ignored.

The last paragraph contains the Roman numeral ALICIMAAMADMIDIAILACMIICLIIIDIMCAMMIDIIDIACADDIDIILDMICALMLII, which is decimal 65193.

Simplified conversion (without A and without subtraction rules) gives: 14921 (simply calculated by the Numberology website).

Commonly, only the Roman numeral signs up to 1000 are used in Mysteries, thus: I=1, V=5, X=10, L=50, C=100, D=500, M=1000.

So much for the ancient Romans. But what do we do with a text in which one or the other German city is mentioned? Maybe it’s worth looking for its postcode or area code? Or maybe we look for airport codes according to IATA or ICAO, railway station numbers, motorways or country codes?

Even road signs are numbered 😉

Spelling errors and other distinctions

Are there many spelling mistakes? Even though not every cachemobile seems to have a Duden in it these days, an accumulation of errors can also work as a cipher. Word counting could work as a key here, for example counting up to the misspelled word in each sentence. Perhaps one has to translate the wrong and right words as 0 and 1 into a binary code? Or maybe the corrections as letter values are the key? Alternatively, the wrong letters translated into letter values (A=1, B=2,…)? Likewise, certain letters could be in italics or bold, slightly shifted up or down like on an old typewriter, smeared with a tiny “blob” of colour, or (a look at the source text shows it!) written in a very slightly different colour? Or is the solution simply written white on white between the listing text, which only serves as a distraction? Simply highlighting the listing with the mouse exposes such things.

Code words/jargon code

Code words were already popular but also easy to see through ciphers. Sometimes, however, you have to look twice to unmask them. For example, you can easily assign numerical words such as ATTENTION to 8, but the idea that “unauthorised” could stand for the number 401 (http error messages) is a bit more far-fetched. In case of doubt, a search engine or a site like Code Crackers will help to debunk strange terminology. If you would like to read a little more about such jargon codes, I can warmly recommend Klaus Schmeh and his books. They are exciting, funny and informative, the surprise eggs among such German-language literature, so to speak.

Also among non-radiophiles, there is a fairly well-known collection of code words that do not necessarily encode letters, after all, they start with them. Nevertheless and for the sake of completeness: the spelling board. Anton, Berta, Cesar, Dora, Emil… Or also NATO: Alpha, Bravo, Charlie, Delta, Echo, …

Another poem!

Is the listing text a well-known text, perhaps even a song or poem? If so, it may differ from the original in tiny details, and the differences (as letter values or place value) are the key?

Sometimes you find a number salad like 4-3-1, 7-3-3, 8-8-8 under a text of some kind. Mostly this simply means that you should use the x-th word in the x-th paragraph and the x-th letter of it. And then from that – as almost always – of course its letter value.

Does the cache owner write in strange hieroglyphics, like: “”$%§&/(()” ? Then take a closer look at your keyboard (if you still surf around with a standard keyboard), especially
at the line with the digits.

Letter salad

If it is absolute letter salad, then it could be base encodings for example Base64. This is used to encode 8-bit binary data into ASCII characters and contains lower and upper case letters, as well as the digits 0-9 and the + and / characters.

A close relative is Base85, which also contains various
special characters. If there are upper and lower case letters, digits from 0-9 and the characters + and -, it is probably UUencode.

Absolute letter salad can also only serve to confuse. Sometimes concentrated “reading” of this alphabet soup is worthwhile in order to still be able to extract meaning from its deepest interior.