A Caesar cipher applies one fixed alphabet shift to every letter. When the shift is missing, you can test the entire A–Z key space rather than guessing one letter at a time. There are 26 residue classes, including shift 0, which leaves letters unchanged. The other 25 are the nontrivial possibilities. This small search is why Caesar is a teaching puzzle rather than secure encryption.
Name the direction before testing
CipherSteps labels a candidate with the shift used to encrypt the original message. To test encryption shift 3, subtract 3 while decoding. For example, K moves back to H. Adding 3 again would take K to N and test a different hypothesis. Shift 29 is equivalent to shift 3 modulo 26; testing both adds no new candidate.
Recover a familiar greeting
HELLO, WORLD!Caesar shift 3; preserve punctuationKHOOR, ZRUOG!| Encryption shift tested | Decoded candidate | Observation |
|---|---|---|
| 0 | KHOOR, ZRUOG! | Baseline: the input letters remain unchanged. |
| 2 | IFMMP, XPSME! | Almost-looking words are not enough. |
| 3 | HELLO, WORLD! | A readable candidate worth verifying. |
| 4 | GDKKN, VNQKC! | One shift farther changes every letter again. |
Solve it with a full candidate list
- Open the ciphertext analyzer and paste KHOOR, ZRUOG! into its input. Press Analyze text.
- Open All 26 Caesar shifts. The rows are ranked by an English frequency score, but every shift from 0 to 25 is present.
- Choose shift 3 to read the full candidate. Check its punctuation, any proper names and the puzzle clue.
- Open the Caesar tool in Decode mode with shift 3 and confirm the complete plaintext.
- Switch to Encode with HELLO, WORLD! and the same shift. Compare its output with the original ciphertext, character by character.
Why the first ranked row can be wrong
The analyzer compares letter counts with a general English reference using chi-square. It does not understand sentences. A two-letter answer, a place name, a pangram or a list of initials may rank below gibberish. Even a longer message can have an unusual distribution. A score should decide what to inspect first, not what to accept.
For a paper exercise, write the ciphertext under an alphabet strip and move the plaintext strip one position at a time. Record each candidate rather than stopping at the first recognizable fragment. A likely word in one position must coexist with sensible text everywhere else, using one consistent shift. If two candidates make sense, the puzzle needs another clue.
If no row makes sense
| Check | Why it matters |
|---|---|
| Transcription | One missing character or a mistaken symbol can hide a short answer. |
| Alphabet | This tool rotates only A–Z; accents are not extra positions in the alphabet. |
| Language | English rankings are not suitable evidence for another language. |
| Cipher family | Atbash, a general substitution or a transposition is not necessarily a Caesar shift. |
| Multiple stages | A correct shift may reveal another puzzle, but do not invent a second stage without evidence. |
The round-trip test verifies that the chosen shift reproduces the input under the stated normalization. It does not by itself prove that your readable sentence is the intended answer. Keep the key, alphabet convention and supporting clue with your solution, especially when preparing an answer key for someone else.
A small independent practice
Try decoding PHHW DW QRRQ without being given a shift. Inspect the entire result before opening the callout answer below. Then re-encode your proposed answer to check it.