Computerized cryptanalysis has obviously made formerly "secure" hand ciphers like Playfair, Four Square, and the Hill Cipher obsolete because they can be defeated in seconds. But is there a hand cipher where a modern computer attack does not have an advantage over a cryptanalysis done by hand? There are certainly tasks that are more quickly done by humans, rather than computers. I'm wondering if there is a hand cipher that takes advantage of this.

  • 2
    $\begingroup$ Duplicate of crypto.stackexchange.com/q/1653/351 (see also crypto.stackexchange.com/q/844/351). $\endgroup$
    – D.W.
    May 19, 2013 at 3:10
  • $\begingroup$ @D.W. Do you know that there is a close-link where you can formally mark a question as a duplicate? (Just below the tag line.) $\endgroup$ May 20, 2013 at 10:40
  • $\begingroup$ Thanks, @PaŭloEbermann. Yes, I'm aware of it. (For reasons that probably aren't relevant to anyone else, it doesn't work for me: on my primary platform, something on my browser's configuration makes the "close" link not work. Maybe an ad blocker or something, I've never taken the time to fully trouble shoot it. My apologies for cluttering things up with comments as a result.) $\endgroup$
    – D.W.
    May 20, 2013 at 16:19
  • $\begingroup$ @D.W. Almost certainly an ad blocker. I have to disable my "privoxy" to get the StackExchange "close" etc. links to work. $\endgroup$
    – Nemo
    Oct 23, 2013 at 18:53

2 Answers 2


One time pad is definitely both easy to do and has perfect secrecy, but key management is a pain and can compromise security. Basically a Vigenère cipher with a key as long as the the message should be secure, because different keys can create ALL possible messages with equal probabilities.

Again, it's a one time pad, so no KPA, CPA, or CCA security. Actually I don't think any pen & paper cipher system has such security.

  • $\begingroup$ Isn't a Vigenere cipher with a key as long as the plaintext basically a one-time pad? $\endgroup$
    – rath
    May 20, 2013 at 9:33
  • $\begingroup$ @rath: Not necessarily! The selection process for the key is what matters. If you use English (or whatever language) text, a standard Vigenere cipher technique, then the biased frequencies in the key let a cryptanalyst potentially break your ciphertext. But if your key's letters are from a uniformly random distribution, then it is pretty much the same thing. $\endgroup$
    – Reid
    May 20, 2013 at 13:26
  • $\begingroup$ @Luke Sheppard: Just curious , what would be your motivation for wanting a hand-generated cipher in the first place? The only scenario I can think of would be in prisons, where gangs usually communicate to their fellow gang-bangers by some kind of handwritten code. $\endgroup$ Jun 3, 2013 at 22:58
  • 1
    $\begingroup$ @WilliamHird A scenario would be a journalist investigating abuses in a totalitarian state. Journalist generates a series of one-time pads (literal pads) and keeps a photocopy/scan of them in a secure bank vault in their democratic home country. While taking interview notes and other observations in the totalitarian state, each note is encrypted with a one-time pad which is then burned. Once the journalist gets home, they can decrypt their notes. You'd probably also want physical hiding techniques for the pads and the encrypted notes. $\endgroup$ Oct 18, 2013 at 0:29

The solitaire cipher is pretty good. The number of ways to shuffle the card pack is 2.3 e71. This is a huge number of possibilities that even a computer couldn't check through...

  • $\begingroup$ It would be nice if somebody showed how you can create a nice scientific format using TeX here by editing the answer. $\endgroup$
    – Maarten Bodewes
    Aug 15, 2015 at 11:18

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