The most common letters in English: letter frequency table

The most common letter in English is E, at about 12 to 13% of the letters in ordinary text. Then come T (about 9%), A (about 8%), O and I (7 to 8% each) and N (about 7%). Those six letters alone make up close to half of everything you read. At the other end, J, X, Q and Z together account for well under 1%. The exact percentages depend on which texts were counted, so the tables below show three published counts side by side. To count the letters in your own text, paste it into the letter frequency counter.

English letter frequency table (A to Z)

The table is sorted by the first column, which is the set of figures from Robert Lewand’s Cryptological Mathematics (2000) as reproduced on Wikipedia’s letter frequency page. Our letter frequency counter uses these same figures in its “In English” column. The second column is Peter Norvig’s count of the Google Books English corpus (about 3.56 trillion letters), and the third is a sample of 40,000 words published by Cornell University’s mathematics department.

Rank	Letter	Lewand (%)	Google Books, Norvig (%)	Cornell sample (%)
1	E	12.702	12.49	12.02
2	T	9.056	9.28	9.10
3	A	8.167	8.04	8.12
4	O	7.507	7.64	7.68
5	I	6.966	7.57	7.31
6	N	6.749	7.23	6.95
7	S	6.327	6.51	6.28
8	H	6.094	5.05	5.92
9	R	5.987	6.28	6.02
10	D	4.253	3.82	4.32
11	L	4.025	4.07	3.98
12	C	2.782	3.34	2.71
13	U	2.758	2.73	2.88
14	M	2.406	2.51	2.61
15	W	2.360	1.68	2.09
16	F	2.228	2.40	2.30
17	G	2.015	1.87	2.03
18	Y	1.974	1.66	2.11
19	P	1.929	2.14	1.82
20	B	1.492	1.48	1.49
21	V	0.978	1.05	1.11
22	K	0.772	0.54	0.69
23	J	0.153	0.16	0.10
24	X	0.150	0.23	0.17
25	Q	0.095	0.12	0.11
26	Z	0.074	0.09	0.07

A few totals from the Lewand column put the numbers in perspective:

• The five vowels A, E, I, O and U make up 38.1% of letters.
• The top five letters, E, T, A, O and I, make up 44.4%.
• The top twelve, E T A O I N S H R D L C, make up 80.6%.
• The four rarest, J, X, Q and Z, make up 0.47%.

Why the published tables disagree

All three counts agree on the top six letters, E T A O I N, and on Z being last or near last. Below that the order shifts, because each table counted different texts:

• Lewand’s figures come from his 2000 book on the mathematics of cryptography.
• Norvig’s count covers the English books in the Google Books Ngram data (version 20120701): about 743.8 billion word occurrences. Book text has its own vocabulary, and here R moves ahead of H, L ahead of D, and P ahead of G, W and Y.
• Cornell’s sample is only 40,000 words, so the rare letters in particular are less stable. It puts U ahead of C and Q ahead of J.

H is the clearest example of how much the choice of text matters. It is 6.09% in Lewand’s table and 5.92% in the Cornell sample, but only 5.05% in Google Books. Because H appears in very frequent words such as “the”, “he” and “her”, how often those words occur in the counted texts moves its share noticeably.

The practical point: treat any frequency table as an approximation. If your text is short, or is a list of names, product codes or medical terms, its letter counts can be far from all three columns.

ETAOIN SHRDLU and the order in other counts

Typesetters used ETAOIN SHRDLU as the order of the twelve most frequent English letters. Norvig notes that the Google Books count “slightly violated” it: L, R and C each move up one place, giving ETAOIN SRHLDCU.

Two other orders are worth knowing, because they answer different questions:

Question	Most common first	Source
Which letters appear most in running text?	E T A O I N	All three tables above
Which letters appear most in dictionary entries, each word counted once?	E A R I O T N S L C U D P M H G B F Y W K V X Z J Q	Concise Oxford Dictionary analysis, cited on Wikipedia
Which letter starts the most words in running text?	T	Norvig, Google Books count

The dictionary order is different because it ignores how often words are used. In running text, “the”, “that”, “this” and “to” repeat thousands of times, which lifts T and H. In a word list each of those counts once, so letters common in longer words, such as R and S, move up. If you are playing a word game against a dictionary, the dictionary order is the more useful one; for decoding or analyzing real writing, use running-text frequency.

The most common letter pairs

Letters are not independent. Some pairs (bigrams) are far more frequent than the single-letter figures would suggest. These are the top ten in Norvig’s Google Books count:

Rank	Pair	Share of all pairs
1	TH	3.56%
2	HE	3.07%
3	IN	2.43%
4	ER	2.05%
5	AN	1.99%
6	RE	1.85%
7	ON	1.76%
8	AT	1.49%
9	EN	1.45%
10	ND	1.35%

TH and HE come mainly from “the”. Pairs like these are what make frequency analysis work on longer ciphertexts: once you have guessed E and T, a frequent two-letter group that ends in E is likely HE, and a frequent three-letter word is likely THE.

How to count the letters in your own text

The letter frequency counter runs in your browser, so the text never leaves your device.

1. Paste or type your text into the box.
2. Leave Letters only (skip digits and punctuation) ticked to count letters alone. Untick it to include digits and punctuation. Spaces and line breaks are never counted.
3. Tick Case-sensitive (A and a counted apart) if capitals matter; otherwise A and a are counted together as a.
4. Tick Count accented letters as plain (é → e) to merge accented letters into their base letter. Without it, “Café crème” gives separate rows for e, é and è; with it, all three count as e (3).
5. Choose the Order: Most frequent first or Alphabetical.

The results table shows each letter’s count, its share of all counted characters, the typical English percentage (the Lewand figures above) and a bar. The text version below the table lists letter, count and percent separated by tabs, ready to paste into a spreadsheet.

Short texts drift from the averages quickly. The tool’s sample, the opening of A Tale of Two Cities (“It was the best of times, it was the worst of times…”, 129 letters), gives these top letters:

Letter	Count	Share in sample	Typical English (Lewand)
e	17	13.18%	12.70%
t	17	13.18%	9.06%
s	14	10.85%	6.33%
i	13	10.08%	6.97%
o	12	9.30%	7.51%
h	9	6.98%	6.09%
a	8	6.20%	8.17%

E is still on top, but the repeated “it was” and “of times” push T, S and I far above their usual share, and N, R and D almost disappear (two each in the sample). A few hundred letters are rarely enough for the averages to settle.

Using letter frequency to break a Caesar cipher

A Caesar cipher replaces each letter with the one a fixed number of places further along the alphabet. Because every E becomes the same letter, the most common letter in a long enough ciphertext is usually the encrypted E.

Here is a sentence encoded with the Caesar cipher tool at shift 3:

Phhw ph dw wkh rog eulgjh dw qrrq. Wkh hdvw jdwh zloo eh rshq diwhu vhyhq, vr zdlw wkhuh dqg nhhs wkh ohwwhu vdih.

Pasting that into the letter frequency counter puts h on top with 21 of 88 letters (23.86%), followed by w (14.77%) and d (9.09%). If h stands for e, the shift is 3 (e → f → g → h). Decoding with Shift 3 and Decode (shift back) returns the original:

Meet me at the old bridge at noon. The east gate will be open after seven, so wait there and keep the letter safe.

The second most common letter, w, decodes to t and d decodes to a, matching the top of the frequency table. When the most frequent letter is not the encrypted E, which happens with short messages, tick Show all 25 shifts (to crack a message) in the Caesar tool and read down the list for the line that makes sense. The same counting approach helps with simple substitution puzzles such as cryptograms, where you start from E, T and A and use common pairs like TH and HE to fill in the rest.