DecipherTechnical reference

Cryptanalysis

Cryptanalysis

How to turn counts, patterns, repetitions, and guesses into evidence you can test.

Also found as cipher analysis · breaking ciphers

Evidence before answers

Cryptanalysis starts with things you can check: which units repeat, where they occur, which symbols meet, and what changes under a different arrangement. None of those observations identifies a cipher by itself. A useful result narrows the next question or exposes a contradiction.Military Cryptanalytics, Part I

A count becomes useful only after you say what was counted. The frequency and pattern pages use letters, words, coordinate pairs, and other units for different jobs. Changing the unit can change the result without changing the ciphertext.

Observations and claims

KindExampleWhat it means
ObservationABC occurs at positions 1, 4, and 7a statement taken directly from the selected units
Hypothesisthe period may be 3an explanation to test against more evidence
Contradictionone ciphertext letter would need two plaintext valuesthe active claims cannot all hold under that model

The distinction matters most when a plausible guess begins to look familiar. A crib or hypothesis can be internally consistent and still be wrong. Consistency earns another test, not a verdict.

Short texts leave room for accidents

Short samples often make several explanations look reasonable. A repeated gap may be accidental; a high-ranked symbol may not stand for E; a statistic may move sharply when one unit is added. Treat independent clues that agree as stronger than one striking number.Military Cryptanalytics, Part IGaines, Cryptanalysis

Sources

  1. William F. Friedman and Lambros D. Callimahos, Military Cryptanalytics, Part I
  2. Helen Fouché Gaines, Cryptanalysis