How difficulty is scored

Score version 1.1 · 1792 words across 24 categories · every figure on this page computed from the live scorer

Every hangman word list explains what makes a word hard and then sorts by feel anyway. The explanation is always about right. This page is that explanation turned into arithmetic, published so you can check it.

The question the score answers

Not how obscure a word is. How much of it a normal opening gives away.

A guesser does not start from nothing. They open with E, then A, then R, and work through the letters that carry most of English. If those letters cover half the word, the board lights up in three guesses and the rest is a formality. If they cover none of it, the guesser is four wrong before they have learned anything.

The four measurements

Common cover

The share of the word's letter positions taken by the eight letters people guess first: E T A O I N S R. High cover means an easy word. This is the single strongest term in the formula and it is weighted at 10.

Rare share

The share of the word's distinct letters that are rare: J Q X Z V K W. These get guessed late or never, so they survive to the end of a game. Weighted at 6.

Distinct ratio

Distinct letters divided by length. A word with repeated letters hands over several positions from a single correct guess; a word where every letter is different hands over one. Weighted at 3.

Vowel share

Vowels as a share of the word. Vowels are cheap hits that light up the shape of a word, so this term is subtracted, at 8.

Putting it together

The four are combined into one number, and multi-word answers get a discount because the gaps between words are information a single word never gives you:

score = 10 × (1 − commonCover)
      + 6 × rareShare
      + 3 × distinctRatio
      − 8 × vowelShare
      − min(2, 0.6 × extraWords)

easy    score < 3.6
medium  3.6 ≤ score < 6.4
hard    score ≥ 6.4

The weights are a judgement call and we would rather say so than bury them. They were set so that the 1792 words on this site land at roughly 901 easy, 646 medium and 245 hard, which puts a usable number of words in each band. Nothing about them is derived from data we do not have.

Six worked examples

Run through the live scorer at build time. Check any of them against the formula above.

duck

score 12.50 · hard

0% of its letters are among the eight people guess first.
It contains the rare letter K.
Every letter is different, so each correct guess reveals one position.
Vowels make up 25% of it.

rhinoceros

score 1.20 · easy

80% of its letters are among the eight people guess first.
It contains no rare letters.
2 letters repeat, so some guesses reveal more than one position.
Vowels make up 40% of it.

rhythm

score 9.17 · hard

33% of its letters are among the eight people guess first.
It contains no rare letters.
1 letter repeats, so some guesses reveal more than one position.
Vowels make up 0% of it.

banana

score -0.83 · easy

83% of its letters are among the eight people guess first.
It contains no rare letters.
3 letters repeat, so some guesses reveal more than one position.
Vowels make up 50% of it.

kyrgyzstan

score 8.23 · hard

50% of its letters are among the eight people guess first.
It contains rare letters K, Z.
1 letter repeats, so some guesses reveal more than one position.
Vowels make up 10% of it.

photosynthesis

score 2.50 · easy

71% of its letters are among the eight people guess first.
It contains no rare letters.
5 letters repeat, so some guesses reveal more than one position.
Vowels make up 29% of it.

The result that surprises people

Duck scores harder than rhinoceros, and by a wide margin. Nothing has gone wrong. A guesser opening with the eight commonest letters hits nothing in duck, and by the time they have tried them all they have used most of their allowance on a four letter word. In rhinoceros almost every letter is one they try early, so the word collapses.

Every hand-sorted list in this niche would put rhinoceros under hard and duck under easy, because rhinoceros is a stranger animal. That is a fact about zoology and not about hangman.

What the score does not do

It does not know whether a guesser has seen a word before, whether the category was announced, or how good the guesser is. It does not know that a word means something embarrassing, or that everyone at the table works in the field it comes from.

It measures one thing, which is what the spelling gives away to a standard opening. That is a smaller claim than the lists it competes with, and it is the one that can be checked.

Questions

How is hangman difficulty calculated here?

Four measurements are taken from the spelling: what share of the word's letters are among the eight people guess first, what share of its distinct letters are rare, how many of its letters are different from each other, and what share are vowels. They are combined with fixed weights into one number, and two thresholds turn that number into easy, medium or hard.

Why does a short word sometimes score harder than a long one?

Because length helps the guesser. A long word gives up more of the board with every correct letter, and it almost always contains several common letters. A four-letter word made of awkward letters gives a guesser nothing at all in their opening.

Is the score based on how rare the word itself is?

No. It never looks at meaning, familiarity or dictionary frequency, only at the letters. That is deliberate: obscurity affects the last moment of a game, and letters decide everything before it.

Can I check the score of my own word?

Yes. The create page scores whatever you type as you type it, using this exact formula, before you turn it into a game.

Will the formula change?

It might, and if it does the version number changes with it. This page is currently on version 1.1, and it is generated from the same code that scores every word on the site, so the two cannot disagree.

The hardest words on the site Browse the scored lists