A frequency count is the first thing a codebreaker does with an unknown message, and it's just as useful for spotting odd characters hiding in data. The tool counts every character in your text, ranks them, and shows a percentage and a small bar for each, so you can see at a glance that, say, e makes up about one letter in eight of ordinary English.
How to use it
- Paste the text to analyse.
- Choose whether upper and lower case count as the same letter, and whether to look at letters only.
- Read down the ranked list, which shows the top 60 characters.
- Check the unique count to see how varied the text is.
When it helps
- Working on substitution ciphers and word puzzles, where the most common symbol is probably e or t.
- Spotting stray symbols, non-breaking spaces or odd quotes in imported data.
- Comparing letter distribution between languages.
- Teaching basic text statistics.
Good to know
- Tick "letters only" to take spaces, digits and punctuation out of the picture; spaces otherwise sit near the top of every list.
- Whitespace is labelled (space, new line, tab) so you can tell them apart from visible characters.
- Emoji count as single characters, even though they take several bytes to store.
Why letter frequencies are so useful
In ordinary English, e is the most common letter, followed by t, a, o, i, n and s. Other languages have their own profiles: in French e is still on top but a and s follow, and in Spanish a and e lead. That regularity is why simple substitution ciphers fall to frequency analysis, and why a count like this is the first thing a codebreaker does.
Outside puzzles, frequency counts reveal problems in data. A sudden spike in a character that should be rare, such as a non-breaking space or a curly quote, tells you where an import went wrong.
Worked example: a short cipher
Take the ciphertext "WKLV LV D WHVW". Counting gives L three times, V three times, W three times and H, D, K once each. The most frequent letters, L, V and W, are likely to stand in for common letters. A shift of three, which is the Caesar cipher, turns W into T, K into H and L into I, giving "THIS IS A TEST".
With a longer message the counts are far more reliable. A hundred characters is a rough guide, five hundred is solid. The percentage column helps you compare your text against known English frequencies without doing arithmetic.
How to read the results
The bar beside each character is scaled to the most common one, so it shows relative size rather than an absolute amount. The unique count at the top tells you how varied the text is: a paragraph of English prose typically uses 30 to 40 different characters once you count capitals and punctuation, while a string of random hex digits uses just 16.
If you only care about letters, tick the letters-only box so spaces, digits and punctuation don't crowd the list. If you are checking for hidden characters in pasted data, leave it off and look for unfamiliar entries lower in the table.
Questions
Are accented letters counted separately?
Yes, é and e are different characters with their own counts.
Why is the space so high?
There's one between nearly every pair of words, so it usually outranks every letter.