I've been studying the game of hangman (somewhat obsessively) for the last 6 or so years. I only recently discovered this Knuth paper via An Optimal Mastermind (4,7) Strategy and
More Results in the Expected Case [1].
I think the reason it has captivated me for so long is because of the unexpected insight that choosing a known short-term failure or an increased chance of short-term failure (when optimizing for information gain) will result in quicker paths to your final goal AND decreased amounts of short-term failures.
I think the reason it has captivated me for so long is because of the unexpected insight that choosing a known short-term failure or an increased chance of short-term failure (when optimizing for information gain) will result in quicker paths to your final goal AND decreased amounts of short-term failures.
My repo (currently being reworked) for simulating different strategies in hangman and battleship: https://github.com/chrisconley/hangman
[1] - https://arxiv.org/pdf/1305.1010.pdf