The Zodiac Ciphers True Crime YouTubers Keep Getting Wrong

On December 11, 2020, an international team of codebreakers solved the Zodiac Killer's Z340 cipher, 51 years after it was mailed to the San Francisco Chronicle. The solution was verified, published in the peer-reviewed journal Cryptologia, and confirmed by the FBI. Yet a large fraction of true crime YouTube content still describes Z340 as unsolved, or worse, claims that anyone with a spreadsheet can crack it.

The Zodiac case has generated more cipher-related misinformation than any other true crime subject. YouTubers with no cryptanalysis background routinely present their own "solutions" to ciphers that are either already solved or structurally impossible to solve with the methods they describe. The confusion stems from a failure to distinguish between the four separate ciphers the killer sent, each with a different construction and a different status.

This post lays out what each cipher actually is, who solved it (when applicable), and the specific errors that recur in true crime content. You can experiment with the cryptanalytic methods discussed here using the Hill Climbing Cipher Solver and the Frequency Analysis tool.

The Zodiac Killer sent four ciphers to Bay Area newspapers between 1969 and 1971. They are commonly referred to by their character counts: Z408, Z340, Z13, and Z32. Each has a different structure and a different solving history.

Z408 was the first cipher, mailed in three parts to the Vallejo Times-Herald, the San Francisco Chronicle, and the San Francisco Examiner on July 31, 1969. It contains 408 symbols drawn from a set of 54 distinct characters. The cipher is a homophonic substitution cipher, meaning each plaintext letter can be represented by multiple different cipher symbols. This design defeats simple frequency analysis by flattening the letter frequency distribution. Donald and Bettye Harden, a schoolteacher and his wife in Salinas, California, solved it within days by exploiting the killer's predictable opening ("I like killing") and working outward from there. The solution was published in the Chronicle on August 9, 1969. The FBI confirmed the solution. Z408 has been solved for over five decades.

Z340 was mailed to the Chronicle on November 8, 1969. It contains 340 symbols and uses 63 distinct characters. For 51 years it resisted every attack. David Oranchak, a software developer and mathematician, maintained the website zodiackiller.com (hosting the Zodiac Ciphers forum) and spent over a decade organizing collaborative solving attempts. The breakthrough came in December 2020 when Oranchak, Australian mathematician Sam Blake, and Belgian programmer Jarl van Eycke applied a combination of homophonic substitution analysis and transposition detection. They discovered that the cipher text, after homophonic substitution, needed to be read in a diagonal pattern across three rows at a time. The solution was published in the journal Cryptologia in 2023 (volume 47, issue 1) under the title "The Solution of the Zodiac Killer's 340-Character Cipher." The FBI confirmed the solution on December 11, 2020.

Z13 (also called the "My Name Is" cipher) was mailed on April 20, 1970. It contains only 13 symbols. With so few characters, the cipher is mathematically underdetermined: there are far more possible plaintexts than there are symbols, meaning any number of solutions can be forced to fit. No solution has been verified. The FBI considers it unsolved.

Z32 (the "Map Code") was mailed on June 26, 1970, alongside a map of the San Francisco Bay Area with a crosshair centered on Mount Diablo. The cipher contains 32 characters and is supposed to indicate a location and a distance. Like Z13, it is too short to solve uniquely. No verified solution exists.

This is the most common error in true crime YouTube content. Videos uploaded as recently as 2024 still state that Z340 "has never been cracked" or "remains a mystery." Some channels acknowledge the 2020 solution in passing but then spend the rest of the video presenting alternative "solutions" that contradict the verified one.

The verified solution describes the killer's obsession with collecting slaves for the afterlife, consistent with the content of Z408. The plaintext was extracted using a specific reading order (diagonal, three rows at a time) applied to a homophonic substitution layer. The Cryptologia paper documents the exact method, the software tools used, and the statistical verification. The solution is not a matter of opinion. It is a published, peer-reviewed result confirmed by the FBI.

If a YouTuber presents a different Z340 solution, they need to explain why their method produces a different plaintext than the one published in Cryptologia. None of them do. They typically present a string of text that "sounds like" something the Zodiac might have written, without any cryptographic justification. This is not cryptanalysis. It is pattern matching on noise.

Every Zodiac cipher that has been solved uses homophonic substitution. In a standard substitution cipher, each plaintext letter maps to exactly one cipher symbol. In a homophonic cipher, each plaintext letter maps to a set of cipher symbols, and the encoder rotates among them. The letter E, the most common English letter, might map to eight different symbols. The letter Z, which is rare, might map to only one.

This construction flattens the frequency distribution. If you run the Frequency Analysis tool on a homophonic ciphertext, you will see a much flatter distribution than standard English. This is by design. The Zodiac Killer, or whoever constructed the ciphers, understood that simple frequency analysis breaks monoalphabetic substitution. The homophonic layer forces the analyst to first identify which symbols map to the same plaintext letter before frequency analysis can proceed.

True crime content routinely ignores this. YouTubers apply single-symbol frequency analysis directly to Z340 or Z408, find that the distribution looks flat, and then declare the cipher unsolvable or present a "solution" based on visual pattern matching. The actual solving process for Z340 required identifying the homophonic groups first, then applying a transposition, then reading the result. The Chi-Squared Calculator can help measure how closely a candidate homophonic grouping matches English letter frequencies, which is part of what Oranchak's team used.

Z13 and Z32 are short ciphers (13 and 32 characters respectively). A 13-character cipher has so few constraints that hundreds of plausible plaintexts can be made to fit. This is a fundamental property of short ciphers, not a matter of insufficient effort.

True crime channels regularly present "solutions" to Z13 that produce a name (often the YouTuber's preferred suspect) and claim this as evidence. The problem is that with 13 symbols, you can assign letters to produce almost any 13-character string you want. There is no way to verify such a solution without external corroboration. The cipher does not contain enough information to uniquely determine its plaintext.

The same applies to Z32. A 32-character cipher paired with a map and a crosshair suggests the plaintext encodes a bearing and distance, but the encoding scheme is unknown. Without knowing the scheme (is it degrees? radians? compass directions?), the ciphertext is underdetermined. You can produce a solution that points to any location on the map by choosing the right encoding assumptions. That is not solving a cipher. It is reverse-engineering a desired output.

The FBI's Vault release of Zodiac case files includes the cipher materials and laboratory analysis notes. These records confirm that the FBI's cryptanalysis unit was unable to solve Z13 and Z32, and that the agency considers both open.

True crime content frequently mislabels the Zodiac ciphers. The most common misidentification is calling them "simple substitution ciphers" or "Caesar ciphers." They are neither. Z408 and Z340 are homophonic substitution ciphers with an additional transposition layer in Z340. Z13 and Z32 are too short to classify with confidence, but they are not Caesar shifts.

A Caesar cipher shifts every letter by a fixed amount. You can test this yourself with the Vigenere Auto-Solver, which handles polyalphabetic ciphers and can confirm that no single shift produces readable text from any Zodiac cipher. The construction is far more complex than a rotation.

Another misidentification is calling Z340 a "transposition cipher." It is not a pure transposition cipher. It combines homophonic substitution with a reading-order transposition. Calling it a transposition cipher omits the substitution layer, which is the part that took 51 years to crack. The transposition alone, applied to the raw symbols, produces nothing readable. Both layers had to be attacked together, which is what made the cipher so resistant.

David Oranchak documented over 650,000 failed solving attempts on his zodiackiller.com forum before the 2020 breakthrough. The number of failed attempts is itself evidence that the cipher is not a simple type. If it were a standard substitution or a standard transposition, it would have fallen to automated tools within hours.

The Z340 plaintext, once decoded, reads as a continuation of the killer's writings about collecting slaves for paradise. It does not name a suspect. It does not contain a confession with verifiable details. It does not provide a location. This disappointed many true crime consumers who expected the cipher to crack the case open.

The Cryptologia paper by Oranchak, Blake, and van Eycke is candid about this. The solution provides insight into the killer's psychology and confirms that the cipher was a genuine communication, not random gibberish. But it does not identify the killer. The FBI's confirmation statement notes the same: the solution is cryptographically valid but does not advance the criminal investigation.

This is worth stating plainly because true crime content often implies that solving a cipher will solve the case. Ciphers contain whatever their author chose to put in them. The Zodiac Killer chose to put rambling threats about the afterlife in his ciphers, not his name and address. Solving the cipher tells you what he wrote. It does not tell you who he was.

If you want to understand the solving method in detail, the Hill Climbing Cipher Solver implements the same class of optimization algorithm that Oranchak's team used to search the homophonic substitution key space. Running it on a short homophonic cipher gives you a feel for why Z340 took 51 years: the search space is enormous, and the scoring function has to distinguish between near-misses and the correct solution.