The best first guess
Computed from 301,132 words at build time · presence, not frequency
Everybody opens with E. For most word lengths that is right, and for some it is not. The difference comes from a distinction that gets skipped: how often a letter appears in English text is not how often it appears in a distinct word.
E is the most common letter in written English partly because a handful of very common words carry it. Hangman is not played against text. It is played against one word, pulled from a dictionary, and every word counts once no matter how often anybody says it. Measured that way the picture changes, and it changes most at short lengths where the field is smallest.
By number of blanks
The share of words of each length containing the letter at least once.
| Letters | Words | Best opening | Then | E ranks |
|---|---|---|---|---|
| 3 | 2,044 | A27% | E 21% · O 20% · I 17% | #2 |
| 4 | 6,798 | A37% | E 33% · S 27% · O 27% | #2 |
| 5 | 15,034 | A44% | E 43% · S 38% · R 31% | #2 |
| 6 | 28,447 | E58% | A 47% · R 39% · S 37% | first |
| 7 | 40,230 | E63% | A 50% · I 47% · S 45% | first |
| 8 | 49,802 | E67% | I 54% · A 53% · S 50% | first |
| 9 | 50,748 | E72% | I 61% · A 57% · R 54% | first |
| 10 | 43,773 | E74% | I 67% · A 60% · R 59% | first |
| 11 | 36,128 | E75% | I 73% · A 64% · R 63% | first |
| 12 | 28,128 | E77% | I 77% · A 67% · N 67% | first |
Advertisement
How to use it at the board
Count the blanks before you say anything. That number is free information and it points at a different opening than the one you were about to make.
After the first guess the table stops being the best tool. Two or three revealed positions narrow the field far faster than any frequency table, and at that point you should be reading the pattern rather than the odds. The solver does exactly that: it counts which unguessed letter appears in the most of the words that still fit.
Overall, across every length
Pooled across all 301,132 words the order is E 67%, A 55%, I 55%, R 45%, S 29%. That is the answer if you have no other information, and you almost always do have other information.
Where this differs from our word lists
The strategy page reports frequency across this site's own curated lists, which are words a person would plausibly choose for a game. This page reports the whole dictionary. The two disagree slightly, and the curated numbers are the better guide when you know somebody picked the word rather than a machine.
Questions
What is the best first letter to guess in hangman?
It depends on how many blanks you are looking at. Across the 301,132 words measured here the strongest opener changes with length, and A, E each top the table at some length. The table on this page gives the answer for each.
Why is E not always the answer?
Because letter frequency in running text is not the same as letter frequency across distinct words. E dominates English prose partly because of very common short words like THE. Hangman is played against one word chosen from a dictionary, and in that setting the distribution is different, especially at short lengths.
What does the percentage mean?
The share of words of that length that contain the letter at least once. That is the number that matters when guessing, because a letter appearing twice in a word is still one correct guess.
Is this the same as the site's difficulty score?
No, but they are built from the same idea. The difficulty score asks how much of a word a standard opening gives away. This page asks what that standard opening should actually be.