Hangman Word Bank Audit: The 250 Words Behind Our Engine

Most hangman guides quote letter frequency from English corpora. That advice fails here because this site ships a fixed bank of 250 uppercase words in src/games/hangman/HangmanEngine.ts. This audit counts the actual bank, so every percentage below is checkable against the source file.

What the hangman word bank contains

The engine holds five categories: animals, countries, foods, sports, and colors. Each contributes exactly 50 entries, for 250 total. Five entries repeat, so the bank holds 245 distinct words. SALMON, PEACH, LEMON, and OLIVE each appear in two categories, and ROWING is listed twice inside sports.

The game shows the category label above the puzzle before your first guess. Use it, because letter odds differ sharply per category.

Word lengths run from 4 to 10 letters. BASKETBALL is the only 10 letter word. Five more reach 9: ARGENTINA, AUSTRALIA, ATHLETICS, TURQUOISE, and TANGERINE. Only 6 words reach 9 letters or more, while 58 words sit at 5 letters or fewer. Short words are the common case, so plan for tight patterns, not long ones.

Every one of the 250 words contains at least one of A, E, I, O, U. Vowel hunting always pays eventually. A total of 207 words contain A or E, which leaves 43 words that dodge both.

Letter coverage you can spend a life on

Counting how many words contain each letter, across all 250 entries, the ranking is:

1. A appears in 147 words, 58.8 percent.

2. E appears in 125 words, 50.0 percent.

3. N appears in 113 words, 45.2 percent.

4. R appears in 111 words, 44.4 percent.

5. I appears in 104 words, 41.6 percent.

O drops to 88 words, L to 84, and T to 72. U sits in only 46 words, 18.4 percent, so guess it late.

The per category numbers matter more than the global ones. In countries, A appears in 41 of 50 words, 82 percent, and the next best letters are N at 27 and I at 26. In sports the order flips: N leads at 32 of 50, then I at 30, then G at 26, while A manages only 25. Animals favor A at 29, E at 28, R at 25. Colors repeat that shape with A at 29, E at 28, R at 24. Foods spread out, with A at 26, E at 25, and C the surprise third at 19.

The deterministic opening

Follow these numbered rules in order.

1. Read the category label before pressing any key.

2. For countries, guess A first, then N, then I.

3. For sports, guess N first, then I, then G.

4. For animals, colors, and foods, guess A first, then E, then R, or C when the category is foods.

5. After each hit, recompute against the visible pattern instead of continuing the fixed order.

6. Cap blind guessing at three letters with no hit. Each miss burns one of six lives, and a three miss blind run leaves you half dead.

One mechanic rewards covered letters twice. The engine reveals every occurrence of a guessed letter, so a letter that appears twice in the word opens two slots for one press. High coverage letters tend to repeat inside words, which compounds their value.

Six lives, counted exactly

The engine ends the game at six wrong guesses and prints the count as wrongGuesses out of 6. Guessing a letter you already tried costs nothing, because guessLetter returns early when the letter is already in the guessed set. The keyboard grays out used keys and colors wrong ones red, so track your misses there.

The loss screen reveals the word gradually over one second, so read it before clicking Play Again. A loss resets your streak to zero and saves that reset immediately. A win increments the streak by one.

Honest limits of this audit

The bank is static and small. Play enough rounds and you memorize entries, which inflates your streak beyond what raw guessing skill explains. The duplicate words slightly skew odds in two categories at once, since SALMON in animals and SALMON in colors are two separate rolls of the same word. There is no hint button, no difficulty setting, and no way to lock the category. The streak lives in localStorage under hangman_streak, so clearing site data for the domain erases it.

The category conditional opening above also assumes the word fits the labeled category, which the code guarantees by picking words only from the matching list. If a future engine version swaps the bank, rerun the count before trusting these percentages.

Hangman pre guess checklist

  • [ ] Category label read before the first key press.
  • [ ] Opening letter matches the category table above.
  • [ ] Wrong letter counter checked before every risky guess.
  • [ ] No repeated guesses on grayed out keyboard keys.
  • [ ] Pattern reconciled with the 245 word bank before the final letter.

Open the [hangman game](/en/hangman) and run the category opening for ten rounds. If your hit rate on the first guess lands below the percentages above, send us your round log with categories and first letters. Mismatched numbers against this bank are the most useful correction you can contribute.