I am trying to solve a riddle where someone has hidden information in a piece of text. The riddle is composed of 4 paragraphs of otherwise normal text (plain english, contents look fine, there are no spelling mistakes, no words with weird capitals... ). I am also almost sure that there is no steganography involved (and it is definitely no a prank).

After trying a lot of things, the only unusual thing that I have spotted are the lengths (number of words) of each of the paragraphs: 16 words, 32 words, 24 words and 8 words. And something tells me that this may not be due to chance.

This may be a really long shot, but still... Has anyone seen anything similar before? If you were to create a puzzle to hide a small sentence/word within a short text, would you use the number of words per paragraphs as a clue? If so, can you provide examples?


The hidden information is probably a text, so you need to find a way to make the word counts correlate with the 26 letters of the alphabet. The easiest way would be to use A1Z26, i.e. map 1 to A, 2 to B and so until Z to 26. The word count of 32 in the second paragraph rules this out.

It is remarkable that all paragraphs have a number of words that is divisible by eight. It could mean that you have eight binary bytes of information, which could then refer to ASCII. If that is the case, each of the words proper must have a binary value of either 0 / false or 1 / true. For example words with an odd number of letters could be considered as 1 and words with even numbers of letters could be considered as 0. The sentence "It was even more interesting and better fun" would then encode the binary sequence 01001101, which is 77 in decimal and the letter M in ASCII.

A similar technique is the Bacon cipher, which uses a binary code of five bits. This could be the case with your text: The 80 words could be 16 Bacon letters or 10 8-bit ASCII characters.

The words can be encoded in different ways. The parity of word lengths, as above, is only one example: Words with a double letter are 1, those without 0; words that start with letters from A-M are 1, those that start with N-Z are 0; and so on. There may ba a hint hidden in the text itself. If the encoding is ASCII, the first bit is usually 0. In general, not all codes yield valid letters.

Happy decoding!

  • $\begingroup$ Yeah, it is kind of weird to have all sentences with a wordcount divisible by 8 - although I am starting to think that it may be just a way of hinting that the sentences must be sorted (ASCII doesn't seem to work here). Thanks also for the reference to the Bacon cipher! You mentioned a few interesting ways of hiding 0's and 1's that I had not think of... I'll have to see if any of them are useful ;) $\endgroup$ – carrdelling Apr 27 '18 at 17:17


Has anyone seen anything similar before?
It's not uncommon to use number of words in a paragraph or a sentence to get a final answer. In fact, I have seen it being used multiple times, for instance here.

If you were to create a puzzle to hide a small sentence/word within a short text, would you use the number of words per paragraphs as a clue?
Well, this part is opinion-based but personally I might do that. In fact, I have in the past(Granted that wasn't my best creation. But, whatever).

So, long-story cut short, using number of words per paragraph as a clue for a final answer is indeed used by puzzlers and it's something a steganography-loving puzzle giver creator would definitely do.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.