Cipher puzzles with the answers on their own page
Printable sheets using Caesar, Atbash, Affine, Vigenère or a transposition, with a difficulty setting that decides how much of the key you are given. Made in your browser from your own text or the lines bundled with the tool.
Every cipher here was broken long before computers. They make good puzzles precisely because they can be worked out with a pencil — which is the same reason they protect nothing. The playground will break them in front of you.
Processed locally in your browserNothing you enter here is uploaded, logged or sent to any server.
What each difficulty gives away
Easy prints the key. The puzzle is applying it, which for a first encounter with a cipher is the right amount of work — the mechanism is the thing being learned.
Medium gives a hint: the shift is odd, or the keyword is a colour, or the number of rails is between three and five. Enough to cut the search down without removing it.
Hard gives the ciphertext and the cipher's name, nothing else. For Caesar that means twenty-five keys to try; for Vigenère it means finding the key length first, which is a genuinely different kind of work and needs a longer message.
The sheet says which cipher was used in every case. A puzzle that withholds that is not harder, only more annoying.
Why groups of five make it harder
Word lengths leak a great deal. A one-letter word in English is A or I; a three-letter word ending the sentence is very often THE or AND; apostrophes give away S and T. Solvers use all of it before touching letter frequencies.
Breaking the ciphertext into fixed groups of five removes every one of those footholds and leaves only the letter counts. It is how ciphertext was actually transmitted — telegraph companies charged by the word and five-letter groups were the standard unit — so it is authentic as well as harder.
Where the bundled lines come from
They were written for this site rather than taken from a quotation collection. That avoids a licensing question, and more importantly it avoids the thing puzzle sites do constantly: attributing a line to somebody who never said it, because the collection they copied from was already wrong.
Your own text is used instead whenever you supply any, and is held to one check — a message under about twenty letters does not give a solver enough to work with, and the tool says so rather than printing a puzzle that cannot be solved.
Using these with a class
Caesar on easy, then Caesar on hard, is about twenty minutes and teaches the idea of a key space more effectively than a definition does — the second sheet is solved by trying all twenty-five keys, and the point makes itself.
Atbash is a good five-minute opener because there is no key at all, which surprises people. Vigenère on hard needs a long message and a group that has already met frequency analysis; anything shorter and the index of coincidence will not settle.
Printing
The answer key starts on its own page, so stopping one page early gives you the puzzles alone. Each key gives the plaintext and the full key, which is what you need to mark a half-finished attempt rather than only a completed one.
Related
The cipher playground to work through the mechanics interactively, or cryptograms for the same idea as an on-screen game.
Questions people actually ask
Are these ciphers secure?
No, and that is why they work as puzzles. Every one was broken long before computers existed — they can be solved with a pencil, which is exactly what makes them solvable and exactly what makes them useless for secrets.
What changes between the difficulty settings?
How much of the key you are given. Easy prints it, so the puzzle is applying the cipher. Medium gives a hint that narrows the search — the shift is odd, the keyword starts with F. Hard gives only the ciphertext and the cipher's name. The cipher is always named, because withholding that makes a puzzle annoying rather than harder.
Why does "groups of five" make it harder?
Because word lengths leak a great deal. A one-letter word is A or I; a short word ending a sentence is often THE or AND; apostrophes give away S and T. Solvers use all of that before touching letter frequencies. Fixed groups of five remove every foothold — and it is authentic, since that was the standard unit for telegraphed ciphertext.
Can I use my own messages?
Yes, one per line. They are held to one check: under about twenty letters a message gives a solver almost nothing to work with, so the tool says so rather than printing something unsolvable.
Where do the bundled lines come from?
They were written for this site rather than taken from a quotation collection — which avoids a licensing question, and avoids the thing puzzle sites do constantly, which is attributing a line to somebody who never said it because the collection they copied was already wrong.
Can I print the same sheet again?
Yes. Every sheet shows the seed that produced it, and typing that back in rebuilds it exactly — same messages, same keys, same answer keys.