Yet Another Person Cracks the 340

As you may have heard by now, a man from Tewksbury, Massachusetts named Corey Starliper is the latest in a long line of people who claim to have solved the 340.

Fortunately, this time, the claim is accompanied by the supposed solution, details regarding the procedure used and other insights into the circumstances surrounding the alleged decipherment. Pretty clearly, Starliper's solution is not valid.

In the primary article relating the facts of the solution, Starliper explains how he mapped all of the non-letter symbols to letters such that he ended up with a cipher that only included the 26 letters of the alphabet (the original cipher has 63 unique symbols). Then he applied a Caesar cipher - a simple substitution cipher that involves shifting the letters of the alphabet some fixed number of positions. After doing this, the solution broke down after the "first few lines," so he adjusted the cipher - presumably by reshifting the Caesar cipher - which he continued to do on an as-needed basis.

The secret to convincing yourself that your invalid solution to a Zodiac cipher is actually a valid solution requires the introduction of what I call "degrees of freedom." By introducing enough of these degrees of freedom, you allow yourself varying ways in which to move from the unambiguous cipher to the questionable solution. The more degrees of freedom you introduce, the easier time you have navigating the path from cipher to solution. Some common degrees of freedom that people tend to use include:

  • Anagrams, i.e. rearranging the order of letters in the solution
  • Garbled meaning, i.e. not requiring true coherence from the solution
  • Binary number manipulations. Sometimes you hear of people trying to apply binary number representations to some aspect of the Zodiac case (not necessarily just ciphers). Usually, when they are doing this, it involves some representation that introduces a degree of freedom.

Starliper's solution defintely has an element of garbled meaning. But the big degree of freedom appears to be the readjustment of the Caesar cipher on an as-needed basis. In general, any time somebody proposes a solution in which a given symbol in the 340 cipher maps to multiple letters, you should be extremely skeptical.

To see a solution that has none of these problems, one need only look as far as the 408 - no anagramming, no garbled meaning, no resetting of processes; just symbols mapped to letters. It's pretty clear that the 340 has some additional complexity when compared to the 408 based on analysis that a number of us in the Zodiac community have performed. However, fundamentally, the cipher is - to a high degree of certainty - a homophonic substitution cipher, i.e. each symbol maps to one and only one letter.

Apparently, Starliper managed to elicit little interest from the relevant law-enforcement agencies. Subsequently, the article conveys:

He even sent the code to a (sic) author with knowledge of the case, who, after looking over the solution, said that it appeared "not valid," according to Starliper.

"That really ticked me off," he said. "With a code that constantly changes a pattern ... you can’t attack it using brute force. There are people who have tried. Out of all of the solutions that I’ve seen this one has the highest readability and probability for accuracy that I’ve ever come across."

Unfortunately, this is another characteristic common among people putting forth invalid solutions: an unwillingness to accept critical objective analysis.

Of course, the motivations to put objectivity and discipline aside are strong. In fact, another quote from the article makes that point clear.

The discovery of a possible solution to the code wasn’t "disturbing," as Starliper said he had heard it described, but invigorating.

"To me, I found it exciting, that I was actually able to get into his head when nobody had for over 40 years," said Starliper. "It was a high. One of the best highs I had ever experienced was cracking something that nobody else had cracked in over 40 years."

Add to this the fact that Starliper's picture is now plastered around the internet (sometimes without qualifiers like "may have" and "claims") and it's probably safe to say that we've sown the seeds for more people to solve the 340.

There really needs to be some measure of potential legitimacy applied to these claims of decipherment. I mean, is there even one person with some basic knowledge of cryptanalysis who believes Starliper's solution is correct?!

If you're still interested, here's the so-called solution.


Michael Cole


    • Mike 25 July, 2011 at 18:19

      Very impressive Dave. I have to take my hat off to you for having the patience to work through Starliper’s clearly-bogus solution is such detail. Based on what you’ve described, the situation is even worse than I initially suspected. You don’t even need to have a basic knowledge of cryptography to see that you can Caesar shift your way to any arbitrary message using this approach.

      I encourage anybody reading this to take a look at what Dave has put together.

    • Michael Butterfield 2 August, 2011 at 21:24

      Yeah Dave, Great Work! But when are you going to quit using your “headache” excuse for dismissing that other “solve” out there. You’re only finding the easy ones…


    • Dave 23 January, 2012 at 12:47

      I’m willing to take an aspirin and do an analysis of the other “solve” out there (I assume you are referring to the Robert Ackerman “solution”). Just let me know.

  1. Kroy, Wen D. 28 July, 2011 at 15:59 Reply

    Righteous. Starliper is clearly a douche. I sometimes wonder if everything qualifies as news. Is fact checking a bygone custom of the elderly? I feel better knowing that at least a few people can see farce coming up Broadway. People were murdered. You have to wonder what kind of parasite could use them as currency to pad his facebook. Now I have a picture of him and I know exactly what sort of parasite.
    Thanks Mike & Dave

    • Mike 28 July, 2011 at 19:03

      I believe Starliper had good intentions and that he honestly *believes* that his solution is correct. But that matters little; the solution is clearly not correct.

      I fault the Tewksbury Patch for running the story. They should have gotten some independent authority to confirm that there is at least a reasonable probability that the solution is correct. If they had tried to do that, I’m pretty certain they would have failed. Starliper, at that point, probably would have put his solution on the web someplace where it would have disappeared into the same digital obscurity that all of the other bogus 340 solutions occupy.

  2. lmhm 13 February, 2013 at 07:21 Reply

    The 340zodiac letter say,

    And yes, he used masonic signs, some
    of them modify, arabic numbers, self made
    signs and caesar code, in 0, 3, 6, 9, 12, and
    something else that complete the decoding
    chart for the z340.

    y, arabic numbers, hes own self

  3. Sean 13 February, 2013 at 09:53 Reply

    I know you put a lot of thought into this over the years Mike, has your opinion chnaged in any way with regard to what’s at play here and/or whether there is a message/partial message?

    • Zodiac Revisited 14 February, 2013 at 09:48

      Hi Sean,

      I think it’s a legitimate cipher with a real message. Further, I believe it’s primarily a homophonic substitution cipher (HSC), similar to the 408. However, I don’t think it is only a HSC – if it were, I believe somebody, e.g. the zkdecrypto folks, doranchak, or perhaps even yours truly, would have solved it by now. Rather, I think there are one or more non-trivial complexities in addition to the fundamental HSC structure.

      If this cipher is ever to be solved, it’s likely going to require a combination of skill and luck. The “skill” involves standard cryptanalysis knowledge and computer programming ability, which a number of people certainly possess. The more elusive “luck” will involve correctly identifying enough of the additional complexities to sufficiently weaken the cryptographic strength of the cipher.

      The good news is that time and technology are on our side… I suspect at some point the cipher will be solved. But whether that happens tomorrow or in 200 years, I don’t know. I’m just hoping it happens at some point during my lifetime.

      That’s my $.02… What do you think?

  4. Sean 14 February, 2013 at 12:29 Reply

    Hey Mike,

    We are on the same page with this one. Most likely very similarily encrypted to the 408, with a “twist” or two. Those twists probably earthed in amateurism as opposed to sophisticated knowledge of code. My best guesses would go to things like inclusion of nulls, lines out of sequence, perhaps even the omission of a letter in the solution etc. I don’t spend as much time as I used to on it but every now an then (with a special thanks to doranchak) it’s great to be able to head over to the Zodiac webtoys
    and put different ideas to the test. I really do believe (or is that hope) that all will fall someday to someone with a very simple idea.

  5. Shea 13 May, 2013 at 13:08 Reply

    Well written piece. Just wondering about this line, however:
    “In general, any time somebody proposes a solution in which a given symbol in the 340 cipher maps to multiple letters, you should be extremely skeptical.”

    Have you reached this conclusion based on your study of the Zodiac killer? If so, what has led you to make this statement? I think my college cryptography class spent maybe two classes on HSCs and other substitution ciphers before moving into more complex cryptography, most of which included symbols mapping to multiple letters. I’m just curious as to why you’ve chosen to rule out those possibilities. If you’ve addressed this elsewhere on the site, a link would be fine.

    • Zodiac Revisited 13 May, 2013 at 23:25

      Thanks for the question, Shea.

      Yes, that comment was specifically based on my experience with the case and the killer’s ciphers.

      Certainly, there is an entire class of polyalphabetic ciphers which involve ciphertext symbols mapping to multiple plaintext symbols. Undoubtedly, you must have studied these in your class. And also certainly, as long as the three unsolved Zodiac-killer ciphers remain unsolved, we cannot disprove that the killer used a polyalphabetic cipher. Of course, we also cannot disprove that the killer used some arbitrary mapping of ciphertext symbols to plaintext symbols, one which follows no “rules” at all.

      Fortunately, I think the probability is low on both counts.

      Polyalphabetic ciphers get their strength from the algorithmic mapping of the ciphertext symbols to plaintext symbols. Homophonic Substitution Ciphers (HSCs as you call them) get their strength from the size of the ciphertext alphabet – and the specifics of their mapping from ciphertext to plaintext.

      Generally, polyalphabetic ciphers use a smaller alphabet, such as A-Z. If you want to increase the cryptographic strength of the cipher, you make the mapping more complex.

      For an HSC, if you want to make it stronger, you increase the number of symbols in the ciphertext alphabet (and possibly change the way you map ciphertext to plaintext).

      So, yes, you could have a polyalphabetic cipher which uses an alphabet of 63 symbols and maps each of the 63 symbols to different letters based on its algorithm. However, the likelihood of that being the case just feels low. It’s a bit like trying to create a secure lock and deciding to use both a combination and a key. You’re really better off choosing one or the other and then strengthening your chosen solution based on its specific characteristics.

      The other tidbit of experience that went into that statement has to do with people attempting to solve the cipher – Corey Starliper is a perfect example. These people don’t know a polyalphabetic cipher from a polygon – or a parrot named Poly, for that matter. Nor do they know the basics of how a HSC works. They just start substituting symbols with letters; when their solution starts to look like gibberish, they change symbol mappings until they find a solution that overcomes their minimal level of skepticism. The result? An incorrect solution with ciphertext symbols mapped to multiple plaintext letters. Sadly, it happens rather often.

  6. HereAndNow 19 February, 2014 at 08:37 Reply

    Seems to me that there’s a whole lot of vitriole being flung at this (wannabe?) code breaker and I’m curious as to why, exactly, such is the case.

    Reasonable folks can differ reasonably without ad homonym attacks and inflammatory descriptions such as “hoax” and “bogus” (and the like). The fact is that the Zodiac case has generated an abundance of opinion and, since the Zodiac’s identity remains unknown, most of it is clearly not correct. If one were to get all exercised over the sum of it all there would be little time for anything else.

    The bottom-line here? Cool your jets, folks and be more respectful. Finding fault is not the issue and such is actually a *good* thing. It’s the whole hostile thing that is the problem and makes the environment counter-productive to discovery. This is an interesting forum but it won’t be if this hostility continues. As it says in the heading here “no personal attacks …”. Words to the wise.

    • Arthur Crtis 17 November, 2014 at 15:28

      I strongly disagree with “HereAndNow.” Corey Starliper is a fraud and he should be condemned as such. The words “hoax” and “bogus” readily apply to what he did and discussions about his effort will abound with derogatory words. He’s in the same class as Robert Graysmith and people like that are very detrimental to anyone trying to learn the facts about the Zodiac case.

  7. Stacey E. 13 April, 2015 at 04:37 Reply

    his determination that he “solved” it is pure arrogance. He probably is really convinced he did it, because then he gets to think he’s smarter than everyone else who’s tried. I sat and tried to solve the cipher left behind by the guy who died on a beach in Australia in the 70’s, and talk about a huge pain in the ass. I read several years ago, someone’s interpretation of these and it was so disjointed, I couldn’t believe they’d think anyone else would believe them. The main problem with “solving” these things is that he may have used a different set of codes for each sentence, or worse, each word.
    I have to wonder about the relationship that newest guy had with his dad that he’d actually write a book pointing the finger at him, after he died of course, so he couldn’t defend himself.
    So many people stupidly think this zodiac goofball had other people licking stamps for him. It’s really obnoxious that people with little insight insist on injecting their nonsense into trying to solve crimes.
    Another website has a “recent” letter supposedly from this guy and the writing isn’t even correct. Just because some jackass thinks it’s clever to pretend to be a notorious killer doesn’t mean you should give it any attention. The man had (has) a very distinctive handwriting, and this is written in giant, clear letters. Come on, already.
    (Are goofball and jackass personal attacks? If so, I’m sorry. I hate to think I insulted a serial killer and a doofus)

  8. Theodore Alexopoulos 21 December, 2017 at 19:39 Reply

    This is fun. I commented on another post, too. I don’t know why everybody assumes it will be a substitution and 1-1 matching algorithm, even if the previous cipher was. I could create easily a crypto algo that could denote distances instead of substitution with rolling feature and conditional, one to one and mixed etc etc. Why not?

Leave a reply