Harold Kravcik’s 340 solution

Harold Kravcik created a stir a few years back when he produced a solution to the 340-character cipher. Some people believed it to be the correct solution, so it was submitted to the FBI, and Harold required others to sign non-disclosure agreements to view his solution. This led to a lot of hype that the decades-old mystery of the 340 cipher had finally been solved. But eventually, confidence in his solution was lost. The drama resulting from all this caused Harold, and others who believed in the solution, to receive ridicule and derision from folks in the Zodiac community who are exhausted by the parade of discredited cipher solutions that have emerged since the Zodiac killer committed his awful crimes.

But what’s wrong with Harold’s solution?

His solution can be found here: http://340cipher.com. I’ve included his solution below to try to make it easier to understand.

First, let’s look at the substitution key he came up with. Below is a table showing the cipher symbols and his substitutions. You can see that he uses a slightly different key in the second half of the cipher.

Symbol:
First half: W N N O E H K S E S I L R O O D
Second half: W N L E H K S E S O G R O O L
Symbol:
First half: L G R A A L L I O N T G O I V F
Second half: L Z R A A A L O N D G O I L H
Symbol:
First half: E I L I F I T L S I N F N O O I
Second half: E I Z I N I W K S I K F N O O I
Symbol:
First half: E E E G I D S W I E R L G I D
Second half: E E R G I D W I E R L G I D

Below is what his solution looks like. The original cipher text is shown in black. Harold’s solution is shown in green. Then, his interpretations of his solution are displayed in blue.


Folly, I left two alive.
Folly, I left two. I live li


fe knowing Noel. Hide-


n-gag, I knive Noels wife.


See Leigh? Freak kill


er. I kill, I do. It knows.


Eagle? I wish. It will go
glee. I wish it will go


work in a ** gov. field.


Solved wife for killer.


I like killing. God, do I


sin? Free killings!


KKKs ol’ lodge gone right?


Magic hoodie. I knew


girls workin’ doing L.A. sk-


yline, good reg. folks.


Goon hoard, golf kiosk.


Lingered. I need girl.


**** Sold a kind ez (easier)


er ending wishes lee.


I’d win all new shin-dig.


light news. I i’m Zodiac.

Here are some of the problems that I see with his solution:

  • The solution permits anagramming. There is no order to the anagramming scheme – it is used when convenient to produce “interesting” words.
  • Worse, the solution permits partial word matches. Again, when convenient, it is used to produce interesting words. In several places, Harold combines anagramming AND partial word matches, increasing the degrees of freedom.
  • We know that Z’s writing contains many misspellings, so we expect to have to permit some degree of partial word matches in potential solutions. But I think this is done way too often in Harold’s solution.
  • Subjective interpretation of words is biased towards producing “interesting” phrases. We cannot know if all the other unexplored variations can be excluded. Example: Line 8 says “ovleds wike foi kilr”. Harold interprets this as “solved wife for killer”. But the same anagramming and partial word matching technique can produce “Devils woke for killer”. Or “Solved if I woke killer”. Or “woe if kids overkill”. Or “love kids, few for killer”. Or “loved swim foe; i kill ‘er!” How do you know which one is right? This is the curse of pareidolia.
  • Resulting phrases form a halting manner of speech, similar to Robert Graysmith’s solution. This is a quality of many subjective anagram-based solutions, due to how much easier it is to subjectively force shorter phrases out of gibberish than it is to force longer ones.
  • Harold appears to adjust the key for the 2nd half of the cipher text, but even so, the 2nd half is filled with gibberish despite the presence of a few interesting words. The solution includes many more examples of selective partial word matches and anagramming to overcome the gibberish.

Because so many degrees of freedom are permitted in the approach, many other such solutions are possible, and are equally fruitless. The main reason for this is that if you use so much selective interpretation of gibberish, you’ve vastly increased the size of the set of acceptable solutions. A competent programmer could write a hillclimber algorithm that optimizes the appearances of scrambled and unscrambled key words, and the appearances of partial word matches. The program could even target specific and arbitrary subject areas, so you could probably make the plaintext appear to resemble Dr. Seuss’s biography.

I think Harold has a lot of confidence in his approach because he claims that no one else has produced a strictly homophonic substitution that has as much readable plaintext as his solution. Graysmith’s solution comes to mind, since it uses a similar approach. But I’m sure other similar solutions can be made. I’m not sure how many people are willing to do that work, though.

Still, there is an interesting mystery lingering in this: Did Harold get some of the phrases right? Is there a correct partial solution somewhere in his solution? And, if so, why does the rest of the solution go wrong? Is there some other step that will clear the fog?

Harold has been corresponding with me via email, and he’s been a good sport during this analysis process, despite the drama that broke out when he first showed people his solution. Harold, keep plugging away at the cipher — maybe one day you’ll discover something interesting.

UPDATE (August 26): Harold has been refining his solution in an attempt to improve it. Visit his updated site to view his new solution.