Letter Frequency: How the Percentages Are Calculated

Contents
When you paste English text into a letter frequency tool to see which letters appear and how often, most people misread the results at one of two points. Either they assume the percentage denominator is the total character count, or they don't realize the letter set shifts from 26 to 52 depending on the case-sensitivity option. This post covers exactly what the calculations are based on, then shows how to check the numbers by hand without a tool.
Summary ① The denominator for percentages (%) isn't the total character count (including spaces and digits) — it's the number of alphabetic letters only. ② With case sensitivity off, everything is grouped into 26 letters, A–Z. With it on, results split into 52 categories: a–z and A–Z. ③ Bar length isn't measured against 100% — it's a relative comparison based on the most frequent letter in that text.
Why These Calculations Get Confusing
There are usually three situations where you'd want to check English letter frequency. You're working through a substitution cipher exercise (like a Caesar cipher) and trying to guess which letter stands in for E or T in the plaintext. You're building typing practice material and want to know which letters come up most often. Or you're running linguistic statistics on a body of text.
All three cases run into the same mistake. You count characters yourself, divide by the total length of the text, and then the numbers don't match what the tool shows — so you assume your math was wrong. In reality, the two are just measured against different baselines. Neither one is wrong.
For example, say a 100-character text with spaces and digits mixed in contains only 80 letters. If E appears 10 times, its percentage isn't 10 out of 100 (10%) — it's 10 out of 80 (12.5%). Without knowing that, the tool's output and your own calculation will seem to disagree every time.
How to Calculate It by Hand Without a Tool
You can do all of this by hand or in a spreadsheet. Here's the order.
Step 1 — Decide on case sensitivity first. Whether you merge "A" and "a" into one letter or count them separately changes every result that follows.
Step 2 — Count how many times each letter appears. In a spreadsheet, a formula like =LEN(A1)-LEN(SUBSTITUTE(A1,"e","")) will count occurrences of a specific letter. Note that string functions like this are case-insensitive by default, so counting upper and lower case separately means combining them with the EXACT function.
Step 3 — Add up the total letter count. Sum the occurrence counts for all 26 (or 52) letters. That total is the "pure letter count" — spaces, digits, and punctuation excluded — and it becomes the denominator for every percentage from here on.
Step 4 — Calculate each letter's percentage. Divide that letter's occurrence count by the total from Step 3, then multiply by 100.
Step 5 — Sort the results. If you want the most frequent letters first, sort by percentage descending. If you want to look letters up, list them from A to Z.
For a few lines of text, doing this by hand is fine. But once you're working with a paragraph or more, counting 26 (or 52) letters one at a time gets tedious fast. Check it with our tool.
Paste your text and the A-Z counts and percentages are calculated instantly. Free to use, with no login or install.

The Two Baselines People Get Wrong Most
Here are the two most common points of confusion when reading results, laid out in a table.
| Item | Common misconception | Actual basis |
|---|---|---|
| Percentage denominator | Total character count (including spaces and digits) | Alphabetic letters only |
| Case option | Always grouped into 26 letters | Default is 26 (case-insensitive); turning the option on splits into 52 |
| Bar length | A full-length bar equals 100% | Relative length based on the most frequent letter in that text |
Spaces, digits, and punctuation each show up in the summary stats at the top of the screen (total characters, letters, digits, spaces), but they aren't part of the per-letter percentage calculation. Non-alphabetic characters are excluded, and letters are compared only against other letters.
How Case Sensitivity Changes the Shape of Your Results
The case option also changes the shape of the output itself. With it off, the T in "The" and the t in "the" merge into one, giving you 26 rows from A to Z. With it on, uppercase and lowercase letters each get their own row, producing up to 52 rows.
Which one is right depends on what you're trying to check. If you want to see how much weight proper nouns and sentence-initial capitals carry, turn case sensitivity on. If you only care about how often each letter appears regardless of case, leave the default (off).
Comparing Against the Standard Order (E-T-A-O-I-N)
It's widely known in cryptography and linguistics that the most common letter in English text overall is generally E, followed by T, A, O, I, and N in that order. Put your text's ranking from the tool next to that sequence, and you can get a sense of whether your text has a distribution close to ordinary English prose, or whether specific words and repeated patterns stand out.
Keep in mind that this order is only a tendency for general English prose. Short texts, or ones heavy with proper nouns, code, or repeated words, can deviate a lot.
The Short Version
- Percentages (%) are calculated against the number of alphabetic letters only (spaces, digits, and punctuation excluded), not the total character count.
- The case-sensitivity option determines whether you get 26 categories (the default) or 52.
- Bar length isn't measured against 100% — it's a relative comparison against the most frequent letter in that text.
- You can sort by frequency (the default) or alphabetically.
- Comparing against the widely known standard order (E, T, A, O, I, N…) tells you whether your text is close to a typical English distribution.
Paste your text and the A-Z counts and percentages are calculated instantly. Free to use, with no login or install.
Frequently Asked Questions
Do spaces and numbers count toward the percentages?
No. Total characters, space count, and digit count are shown separately in the summary stats, but per-letter percentages are calculated from the number of English letters only. Even if your text has a lot of spaces and digits mixed in, they don't affect the ratios between letters.
When should I turn case sensitivity on?
There's no single right answer — it depends on what you're checking. If you want to look separately at capitalized words or the share of sentence-initial letters, turn it on and view up to 52 categories. If you only care how often each letter appears regardless of case, just leave the default, case-insensitive setting (26 categories).
The bars are all different lengths: does the longest one mean 100%?
No. Bar length isn't drawn against 100%. It shows each letter's value relative to the most frequent letter (the maximum) in that text. The actual percentage is printed as a number next to each row, so read the number beside the bar for the precise value.
Does the sort order (frequency vs. alphabetical) change the results themselves?
No. Sorting only changes the order in which the same results are displayed — it doesn't affect any letter's occurrence count or percentage. Frequency order is handy for seeing at a glance which letter is used most, and alphabetical order is handy for looking up a specific letter from A to Z.
This article was adapted from our Korean original. Read the Korean original