First up: I am not asking you to decypher anything. This question is about attacking a Playfair cipher; more specifically – how to create potential grids when doing so.

I'm currently working on a message in playfair cipher text. The issue is there is no grid, and only potential crib is attack. I was told there is a key used, but we have to discover it.

What I‘ve tried (and been able to do)

After using frequency analysis, I managed to find that QK appears to be the letters th since it is the most common diagraph. I found that C and V are most common letters in the ciphertext after the letters QK.

iagraphs that follow QK

VE - CQ - AH - RQ - VM - BC - BP - VU - SG - SH - CF - UV - UY - SZ - SQ - VQ - FH - AE - RZ

QK - VE = 1
QK - CQ = 4
QK - AH = 1
QK - RQ = 2
QK - VM = 1
QK - BC = 1
QK - BP = 2
QK - VU = 1
QK - SG = 1
QK - SH = 2
QK - CF = 1
QK - UV = 1
QK - UY = 1
QK - SZ = 1
QK - SQ = 1
QK - VQ = 1
QK - FH = 1
QK - AE = 1
QK - RZ = 1

Most common after QK: CQ



I'm pretty sure uv/vu is er/re. But I am asking myself if one of these could be e


I am having trouble figuring out how to get the right idea to help create a potential grid that would lead to the deciphering the text. How can/should I create and validate potential grids when attacking a Playfair cipher, based on what I’ve got so far?

(In case of doubt: I am looking for a “how to”. This is not a “decypher this” question.)


1 Answer 1


Both your suspicions are correct, BTW. (I solved it, and know the key-word).

You have a crib, namely ATTACK, with a typical pattern: two digrams that are each other's reverse, which means that (if we assume that the crib starts at an even position) the cipher text should also have the pattern "abba". I see a CQQC pattern and you note that CQ is frequent after TH, so that fits nicely with CQ = AT, to give THAT.

The text starts with QK = TH, so the next letter could be E (as A would probably be followed by T), so V is one of the letters that E can encipher to, so it must be in the same row (probably) or the same column as E (same column means also that they are next to each other, so same row is more likely). This also fits with your UV/VU supposition. It then follows that E,V,U are all in the same row (or column, less likely). And this row is then in the keyword (by how keyword squares are made: E is far away in the alphabet from U and V).

So from this we already get quite a few ideas on how letters are positioned relative to each other, and at the end of the square we expect some almost alphabetic sequences, like WXYZ (minus letters that go into the keyword), etc.

Fill in some letters (write a program) and guess more letters. If my ATTA guess was OK, you also get CK that becomes KT, which gives us info: overlapping crypto/plain means same row or same column. Etc. etc.

Have fun puzzling it out.

  • $\begingroup$ Your ideas were very helpful thanks! I was able to break it after filling out the grid the keyword was zombievirus spelled ZOMBIEVRUS $\endgroup$
    – MD_90
    Feb 27, 2016 at 0:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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