In one copy of a 1942 edition of German historian Gert Buchheit’s biography of Rainer Maria Rilke, someone has glued a typewritten and hectographed alphanumeric text. The text fills 33 pages with 18,760 characters in groups of four. Analysis shows that it’s less ordered than English or German but more ordered than random text. To date, no one has been able to make sense of it.
Here’s the cryptogram itself, and here’s an analysis.
From Klaus Schmeh’s Encrypted Book List.
06/21/2026 UPDATE: Reader Logan Swiecki-Taylor writes:
I have a note on the aforementioned post.
The cryptography community (including the author of that very blog, Klaus Schmeh, security expert Tobias Schrödel, and statistician Floe Foxon) essentially ‘solved’ the mystery by figuring out exactly what it is: It is literally ‘Keyboard Mashing.’
The text was generated by someone mindlessly sliding their fingers across a German QWERTZ typewriter.
- Groups like
qwer, tzui, and cvbn are just straight lines across the keys.
- Groups like
cxsw and mju7 are diagonal swipes.
- Statistical analysis published in the journal Cryptologia in 2022 mathematically proved that the physical distances between the keys pressed are too short to be random or to be any known substitution cipher. It is intentional ‘lazy typing.’
The pages in the book were printed using a ‘hectograph’ (an early duplication method used for making a few dozen copies of a document). The prevailing consensus among historians is that this was practice material for radio operators learning Morse Code.
Military radio operators needed to practice transmitting and receiving meaningless letter sequences because actual encrypted military communications (like Enigma messages) look like random 4-letter or 5-letter blocks. Instead of going through the tedious mathematical effort of properly encrypting a real message just for a training exercise, an instructor simply rolled their fingers across a typewriter to generate pages of fake ‘ciphertext’ and printed copies for the class to practice their Morse transmissions.
Timeline of crypto community decipher
January 2018: The Initial Clue (Crowdsourcing)
The mystery was first brought to public attention in early 2018 when the book’s owner, Dr. Karsten Hansky, shared the pages with German crypto-historian Klaus Schmeh. Schmeh posted it on his popular cryptography blog on January 10, 2018, listing it as an ‘unsolved World War II cryptogram.’
On that exact same day, an astute blog commenter going by the name ‘Gerd’ pointed out that military Morse code training typically used random gibberish so that students couldn’t simply ‘guess’ words if they missed a letter. He suggested that it was just a practice text.
2020: The ‘Keyboard Mashing’ Realization
Over the next couple of years, Schmeh, alongside German IT security expert Tobias Schrödel and other blog readers, looked closer at the letter groupings. They noticed the high frequency of spatial sequences (like qwer and aswq) and realized the text wasn’t randomly generated with dice or cryptography, but by someone simply rolling their fingers around a German QWERTZ typewriter. Schmeh eventually published an update officially re-classifying the mystery as ‘probably solved’ and removing it from his list of unsolved ciphers.
August 2022: The Mathematical Proof
The final nail in the coffin came from academia. In August 2022, a data scientist and statistician named Floe Foxon published a peer-reviewed paper titled ‘A treatise on the Rilke cryptogram’ in the journal Cryptologia.
Foxon mapped the letters of the cryptogram to a standard 1940s German typewriter layout. Using statistical analysis, Foxon mathematically proved that the physical distances between the consecutive keystrokes were drastically shorter than what would occur in natural language, a known substitution cipher, or true randomness.
Thanks, Logan. The solution is even more interesting than the puzzle!
