Famous Unsolved Codes and Ciphers in History

Some ciphers resist every attempt to crack them. Despite centuries of effort by professional cryptanalysts, amateur codebreakers, and supercomputers, a handful of famous unsolved ciphers continue to guard their secrets. Each one represents a puzzle that some of the smartest people in the world have failed to solve — and any one of them could be cracked tomorrow by a fresh pair of eyes.

Here are the most compelling unsolved codes in the history of cryptography, along with the ones that finally fell and the techniques that broke them.

1. The Voynich Manuscript (~1400s)

The Voynich Manuscript is a 240-page book written in an unknown script, illustrated with strange botanical drawings, astronomical diagrams, and figures of naked women bathing in green liquid. It's been carbon-dated to the early 15th century. No one knows who wrote it, what language it uses, or what it says.

The manuscript is named after Wilfrid Voynich, a book dealer who purchased it in 1912. It's now held at Yale University's Beinecke Rare Book Library.

What's been tried: Statistical analysis shows the text has structural properties consistent with a natural language — it has word-length distributions, character-frequency patterns, and Zipf's law behavior similar to real languages. But no proposed decryption has been accepted by the academic community. Theories range from an unknown natural language to an elaborate hoax to a cipher using a lost key.

Why it might never be solved: If the manuscript is a hoax — meaningless text designed to look like a language — then there's nothing to decode. If it's a real language that's since been lost, the text alone may not contain enough information to reconstruct it.

2. The Zodiac Killer's Z340 (Solved 2020)

The Zodiac Killer sent four encrypted messages to San Francisco newspapers between 1969 and 1970. The first, called Z408, was a 408-character homophonic substitution cipher that was cracked within a week by a schoolteacher and his wife.

The second, Z340, was a 340-character cipher that resisted all attempts for 51 years.

How it was finally solved: In December 2020, a team of three amateur cryptanalysts — David Oranchak, Jarl Van Eycke, and Sam Blake — broke Z340 using a combination of transposition and substitution analysis. The cipher used a complex scheme: the 340 characters were read in a diagonal pattern (not left to right), and the substitution included multiple symbols per letter (homophonic) plus deliberate errors.

The decoded message was rambling and taunting, consistent with the Zodiac's other communications, but did not reveal the killer's identity.

3. The Zodiac Z13 (Still Unsolved)

The Zodiac's shortest cipher is just 13 characters long. It accompanied a letter claiming to contain the killer's name. With only 13 characters, there isn't enough data for statistical analysis — thousands of possible solutions fit the pattern.

Why it may be unsolvable: Thirteen characters of a homophonic substitution cipher provide almost no leverage for cryptanalysis. Without additional context or a lucky guess, the Z13 may remain permanently ambiguous.

The Zodiac's Z32 (32 characters embedded in a map-based clue) is also unsolved, though it's more likely to be solvable given its greater length. For a deeper dive, see our article on the Zodiac Killer's ciphers.

4. The Dorabella Cipher (1897)

In July 1897, the English composer Edward Elgar sent a letter to his friend Dora Penny containing 87 characters written in an alphabet of wiggly lines — each symbol looks like a squiggle oriented in one of eight directions, with one, two, or three curves.

Elgar never explained the cipher, and Dora Penny said she never decoded it. It's been analyzed as a possible substitution cipher but no solution has been universally accepted.

What makes it unusual: The cipher has only 24 unique symbols for 87 characters, which is consistent with a monoalphabetic substitution in English. But frequency analysis hasn't produced a clean solution — possibly because the plaintext is a personal message too short and quirky for standard frequency patterns to apply, or because the cipher uses additional tricks beyond simple substitution.

5. Kryptos (CIA Headquarters, 1990)

Kryptos is a bronze sculpture installed in the courtyard of CIA headquarters in Langley, Virginia. Created by artist Jim Sanborn, it bears four panels of encrypted text. Three have been solved; the fourth (known as K4) has not.

Sections solved:

• K1 used a Vigenere cipher with the keyword "PALIMPSEST"
• K2 used a Vigenere cipher with "ABSCISSA"
• K3 used a columnar transposition

K4 is just 97 characters long and has resisted every public attempt since 1990. Sanborn has released two clues: characters 64-69 spell "BERLIN" and characters 70-74 spell "CLOCK." Even the CIA's own analysts haven't cracked it publicly.

Why it matters: Kryptos sits in the courtyard of the world's premier intelligence agency, visible to thousands of intelligence professionals daily. Its unsolved section is a deliberate provocation — an unbroken cipher staring at people whose job is to break codes.

6. The Beale Ciphers (1885)

In 1885, a pamphlet was published in Virginia describing a cache of gold, silver, and jewels buried by a man named Thomas Jefferson Beale in the 1820s. The treasure's location was allegedly encoded in three cipher texts.

Cipher 2 was solved using the Declaration of Independence as a key — each number pointed to the first letter of the corresponding word. The decoded text described the treasure's contents (gold, silver, jewels) and its approximate location in Bedford County, Virginia.

Ciphers 1 and 3 — supposedly encoding the exact location and the names of the treasure owners — have never been solved. No key text has been found to work.

The big question: Are the Beale ciphers genuine, or an elaborate 19th-century hoax? Skeptics point out that the pamphlet's publisher had financial motives (he sold copies) and that no treasure has ever been found. Believers note that Cipher 2 was legitimately encoded and that hoaxing all three ciphers with realistic statistical properties would have been remarkably sophisticated for the 1880s.

7. Chaocipher (Invented ~1918, Solved 2010)

John F. Byrne invented the Chaocipher around 1918 and spent decades trying to interest the U.S. government in it. He claimed it was unbreakable but never revealed how it worked. After his death, his papers were eventually donated to the National Cryptologic Museum.

In 2010, Moshe Rubin reconstructed the algorithm from Byrne's papers and published it. The Chaocipher turned out to use two rotating, mutually permuting alphabets — a genuinely clever design that produced strong encryption for its era.

Why it's significant: Chaocipher was unsolved for nearly a century not because the cipher was unbreakable but because nobody knew the algorithm. Once the mechanism was understood, the cipher could be analyzed normally. This highlights an important principle: a cipher's security should come from the key, not from secrecy of the method.

What Makes Ciphers Impossible to Crack?

Several factors can make a cipher permanently unsolvable:

Insufficient ciphertext. Short messages don't contain enough statistical patterns for analysis. The Zodiac Z13, at only 13 characters, is the extreme example.

Unknown encryption method. If you don't know whether you're dealing with substitution, transposition, a book cipher, or something else entirely, you're searching a vast landscape of possibilities. The Voynich Manuscript suffers from this.

Lost keys with no reconstruction path. A Vigenere cipher with a long, random key is essentially a one-time pad — mathematically unbreakable without the key. If the Beale ciphers use book cipher keys and the book is lost, no amount of computing power will help.

Possible hoaxes. If the ciphertext is random gibberish, every "solution" is a false pattern imposed by the analyst.

The Citizen Cryptanalysis Community

Today, thousands of amateur cryptanalysts collaborate online to attack unsolved ciphers. The Zodiac Z340 was cracked not by the FBI or NSA but by three hobbyists coordinating across three countries, using custom software and donated computing time.

Communities on Reddit, specialized forums, and our own cipher tools bring fresh perspectives that professional cryptanalysts sometimes lack. Amateurs aren't bound by institutional assumptions, and they bring diverse backgrounds — linguistics, pattern recognition, programming, historical knowledge — to problems that have stumped specialists.

If any of these famous unsolved ciphers intrigue you, start by exploring the cipher types they're based on. Practice with our substitution cipher solver, learn frequency analysis, and study how solved ciphers were broken. The next breakthrough could come from anyone.

Frequently Asked Questions

Has anyone claimed the Beale treasure?

No. Despite nearly 150 years of searching, no treasure matching the Beale description has been found in Bedford County, Virginia. Most professional historians consider the Beale ciphers a likely hoax.

Is the Voynich Manuscript a hoax?

The debate is ongoing. Statistical analyses show language-like properties that would be difficult to fake in the 15th century, but no decryption has been verified. The current consensus leans toward it being a genuine text in an unknown language or writing system — but a sophisticated hoax can't be ruled out.

Why did the Zodiac Z340 take 51 years to solve?

Z340 used a combination of homophonic substitution and transposition that nobody had anticipated. The characters were read in a diagonal pattern rather than left to right, which made all conventional attacks fail. It took custom software specifically designed to test unconventional reading orders to find the solution.

Can artificial intelligence solve these ciphers?

AI has been applied to several unsolved ciphers with limited success. Machine learning can detect patterns and suggest substitution mappings, but it struggles with ciphers that use unknown or unusual methods. AI is a powerful tool for testing hypotheses but hasn't yet produced a verified solution for any major unsolved cipher.