Kryha Pocket cipher machine · 1924
Sold as the unbreakable cipher of the future. Friedman’s team solved a 1,135-letter test message in 2 hours 41 minutes.
Broken (Friedman, 1933)
Why This Matters
Alexander von Kryha marketed clockwork cipher machines in the 1920s. Models varied: the small Liliput should not be confused with the larger Standard machine. Commercial claims of strength were later challenged by practical cryptanalysis.
The documented 1933 challenge tested a specific message and setup. Its solution provides stronger evidence than sales claims about the number of possible settings, but it should not be generalized to every model or deployment.
Two concentric wheels: a fixed outer ring of plaintext letters and an inner wheel carrying a mixed alphabet. The operator locates a letter on the outer ring, reads off the cipher letter from the inner wheel, then triggers a spring-driven escapement that advances the inner wheel by an irregular number of positions — between 1 and roughly 26 — according to a stepping pattern fixed by the wheel design.
The advertising claimed astronomical key spaces by counting wheel orientations and stepping pawls. The reality is that the inner wheel is a single mixed alphabet rotated by a finite, repeating sequence of step sizes — a polyalphabetic cipher with a long but eventually periodic key.
William Friedman, Solomon Kullback, Frank Rowlett and Abraham Sinkov solved a 1,135-character Kryha test message in 2 hours 41 minutes in 1933. The result concerns that challenge and its conditions; it is not a universal running-time bound for every message or key.
The repeated stepping structure creates related substitutions that can be exploited. How much ciphertext is needed depends on the settings, message and available information; “a few hundred characters” is not a guaranteed threshold.
| Kryha lesson | Modern echo |
|---|---|
| Counting wheel orientations ≠ key strength | Key length ≠ entropy in modern key analysis |
| Vendor unbreakability claims with no public review | “Trust us” crypto vs. open peer-reviewed standards |
| Fixed stepping pattern = finite period | A long period alone does not establish keystream unpredictability |
| 2 hours 41 minutes is a long time to admire a cipher | A demonstrated attack is stronger evidence than a marketing keyspace count |
| Origin | Alexander von Kryha, Berlin |
| Year | Patented 1924; sold widely 1924–1930s |
| Mechanism | Single mixed alphabet wheel, clockwork-driven |
| Marketed as | Unbreakable; recommended by German banks and diplomats |
| Broken by | Friedman, Kullback, Rowlett and Sinkov, US Army SIS, 1933 |
Sources and Reading
Lambros D. Callimahos, “Q. E. D.—2 Hours, 41 Minutes”, NSA Technical Journal, pp. 13–34, especially p. 18 on the four analysts. The time applies to the documented 1933 test message.