W Win Board Games
首页

Hangman

Hangman at expert level is Bayesian inference on word patterns. You update letter probabilities after every reveal, weighting by corpus frequency and positional constraints.

获胜要点

  • Open with E, T, A, O, I, N, S—covering ~65% of English letter mass.
  • After first hits, switch to positional frequency (R in _R_, NG endings) not global frequency.
  • Avoid rare letters until pattern narrows; Q and Z are last-resort guesses.
  • Count blanks and remaining wrong guesses before risking a low-probability letter.

专家策略

  1. 1

    Corpus weighting: expert players use Scrabble or ENABLE frequency lists, not naive ETAOIN. Position-conditioned tables improve hit rate 12-18%.

  2. 2

    Pattern branching: with _ A _ E _, prioritize consonants that appear in C_A_E, M_A_E, W_A_E patterns before guessing L or D.

  3. 3

    Endgame risk calculus: with 2 lives and 4 blanks, compute expected value per letter across all matching dictionary words, not single best guess.

  4. 4

    Anti-human bias: when playing as setter, choose words with repeated rare letters (J, V, X) and uncommon digraphs (PH, GH) to break frequency openers.

  5. 5

    Digraph completion: once TH or CH is partially revealed, complete the digraph before branching—TH__ almost always resolves before testing W.