Hangman Letter Frequency

Most letter-frequency charts count every letter in a body of text. Hangman needs a different number. When you guess a letter, you only learn whether it is in the word, so the useful measure is the share of words that contain the letter at least once. We counted this across the 15,000 most common words in the Hangman Lab dictionary. The full table is below, followed by how the best guess changes as words get longer.

The share of words containing each letter

Here is every letter, ranked from the highest to the lowest. Read each line as the chance that a random common word contains that letter:

• e: 61.9% of words
• a: 51.7% of words
• i: 49.1% of words
• t: 47.8% of words
• r: 47.3% of words
• n: 43.0% of words
• o: 41.4% of words
• l: 36.8% of words
• s: 35.6% of words
• c: 30.4% of words
• u: 25.2% of words
• d: 22.2% of words
• p: 20.4% of words
• m: 20.1% of words
• y: 17.5% of words
• h: 16.0% of words
• g: 14.4% of words
• b: 13.0% of words
• f: 10.7% of words
• v: 9.6% of words
• w: 6.3% of words
• k: 5.6% of words
• x: 2.3% of words
• z: 1.8% of words
• q: 1.5% of words
• j: 1.1% of words

Why share of words beats share of letters

Standard frequency counts add up every appearance of a letter, so a word with three e's counts three times. Hangman does not work like that. Guessing e in eleven reveals all its e's in a single guess, and gives you no extra credit for the repeats.

The two rankings overlap but differ. A letter that repeats a lot inside a few words, like s in a plural-heavy text, can look stronger by raw count than it is for a guesser. The share-of-words figure is the chance your guess is not wasted, and that is the number to follow.

Reading the table

Three bands stand out. The top nine letters, e through s, each appear in more than a third of words. The middle group, c through b, appear in 13% to 30%. The tail, f through j, appears in under 11% of words.

The gap between the top and bottom is large. A guess of e is right 56 times more often than j, which appears in 1.1% of words. Even the fifth letter, r at 47.3%, is more than 40 times as likely to hit as j.

Consonants and vowels split neatly. The vowels a, e, i, o and u all appear in the top thirteen, with u the lowest at 25.2%. Y, at 17.5%, ranks fifteenth and is the letter to remember for short words that appear to have no vowel at all.

How the best first guess changes with length

The all-words ranking is only the average. Filter by word length and the order moves.

For 3-letter words (473 in our set), the best first guesses are a at 30.2%, e at 24.1% and o at 23.7%. For 5-letter words (1,765 in our set), they are e at 44.6%, a at 41.4% and r at 34.9%. For 8-letter words (1,979), they are e at 67.8%, a at 55.8% and i at 54.9%.

At the long end, the leaders change. In 11-letter words, i leads at 77.7%, ahead of e at 75.0% and t at 74.7%. In 12-letter words, i is at 79.4%, t at 78.0% and n at 73.3%. Long words are built from suffixes like -ation and -ing.

Why the percentages rise with length

Notice that the top figures climb from 30.2% for 3-letter words to nearly 80% for 12-letter words. A longer word has more slots, so it is more likely to contain any common letter. That is the reason long words are easy to guess and short words are not.

Our solver confirms this. It averages 7.51 wrong guesses on 3-letter words and 0.28 on 12-letter words.

The vowel-free words

Vowels carry most of the strategy, so words without them are the exception worth knowing. In our word set, 14 of the 473 three-letter words have no a, e, i, o or u, as do 5 four-letter words, 4 five-letter words and 1 six-letter word. None of the words of seven letters or more qualify.

Nearly all rely on y: sky, why, try, fly, dry, cry, spy, shy, gym, hymn, myth, rhythm. They score between 5 and 10 wrong guesses for our solver, high for their length. Our list of hangman words without vowels collects them.

Using the numbers

For a game on paper, memorize the opening order e, a, i, o, r, t, n, s, and shift to the length-specific order if you know the length. For a classroom, put the table on the wall and let students predict which letter will hit before each guess.

Setters can read the table backward. A word built from the bottom of the list, with letters like j, q, x, z, k and w, has the lowest chance of being hit. Our j, q, x, z list does that for you.

A caution on scope: these figures come from the 15,000 most common words in our dictionary. A list of rare or technical words would give a slightly different order, so for a themed game such as animals or foods, trust the pattern once you have a few letters showing and let the topic guide the rest.