November 2006 Archives

Cracking the Zodiac Killer's 340 Cipher, Part 6

| No Comments | No TrackBacks

Not terribly long ago I posted an update on my work to try and crack the Zodiac Killer's 340 cipher. No, I haven't broken it yet, but I do think I am making more progress than ever before.

I hit upon the idea of maximizing my use of the computer's CPU cycles in solving the problem. Computers are best at simple math, so I reduced the problem of identifying the right message key to one of simple mathematics. This means I've removed any attempt to identify English words in the result, instead focusing on whether the key being tested would yield a result that contains the characters A-Z in approximately the same frequency they'd appear in the killer's known writings.

As a result, the laptop on which I'm writing this is simultaneously trying over 400,000 keys per second. The 3 desktop systems in my den are cranking through 2,000,000 keys per second total. At the rate we're going, I'll have tried all the potential combinations of the 10 most common symbols in the 340 cipher by the end of the week. The programs will automatically start on the 11th most popular, etc.

Any of the more promising results can be fed back into the program and used to try to identify more of the remaining symbols.

So far, in spite of the program trying literally billions of potential keys, I've come up with only a handful of possible keys. All but one of these I've been able to tentatively rule out after plugging them into the cipher, since they produce results that bear little resemblance to English. The remaining key has yielded words like "SETTLE", "YET", and "DIES". It may be on to something, I don't know yet.