Home / Blog / Background

What Hangman Teaches You About Probability and Patterns

A paper game that quietly runs small experiments in statistics. How letter frequency, conditional guessing and information all show up in a round of hangman.

5 November 2026 · 5 minute read · Zaher Khan

Hangman looks like a children's word game, and it is. It is also, without anybody quite noticing, a small laboratory for some of the most useful ideas in probability. Every time you guess a letter, you are making a decision under uncertainty, using partial information and updating your beliefs when you learn the answer. People who have never studied statistics do this instinctively in hangman, and the game is a pleasant way to see the ideas working.

This article makes no claim that playing hangman will make anybody better at mathematics. It is more modest than that. It looks at four ideas that show up naturally in the game, explains each in plain language and uses hangman as the example. If you teach, the same examples work as a gentle introduction to the topics in a classroom.

Idea one: frequency

Some letters occur more often than others. In ordinary English text, E accounts for roughly thirteen letters in every hundred, while Z accounts for less than one in a thousand. This is a distribution, and the shape of it, a few common letters and a long tail of rare ones, is a pattern found all over the world in language and elsewhere. When you open with E, you are playing the odds, using the distribution to pick the guess most likely to succeed.

Idea two: conditional probability

A letter's chance of appearing is not fixed. It depends on what you already know. Before any guess, the chance that a word contains a Q is tiny. After a Q lands, the chance that the next letter is U is nearly certain. That change, from a general probability to one conditioned on new information, is the heart of conditional probability, and hangman shows it again and again. Every letter that lands changes the odds for every other letter on the board.

Idea three: information

Not all guesses are equally useful, even when they are equally likely to be right. A good guess is one whose answer, whichever way it goes, tells you the most. This is the idea at the centre of information theory. A guess with a roughly even chance of hitting or missing gives you the most information in either case, while a guess that is almost certain to hit, or almost certain to miss, tells you little. That is why the very first guess is worth taking carefully, and why experienced players sometimes choose a letter that is not the most likely, because it splits the possibilities more cleanly.

Idea four: updating

A strong player does not fix a belief and cling to it. They hold a rough picture of what the phrase might be and revise it with each new letter. If a guess that should have landed misses, the picture changes: perhaps the word is not what was assumed. That habit of updating a belief when evidence arrives, without being attached to the earlier guess, is one of the most valuable thinking habits there is, and hangman rewards it directly, since stubbornness loses lives.

• Hold two or three possible readings of the phrase at once.
• Choose a guess that helps tell them apart.
• When the answer arrives, discard the readings it rules out.
• Be willing to abandon a favourite theory the moment the board contradicts it.

Why the setter has the advantage, and how they use it

The setter knows the answer, and so has a hidden advantage. A clever setter picks a phrase that exploits the guesser's likely assumptions, such as a phrase with few common letters. This is, in effect, a game of two people reasoning about each other's reasoning. The guesser follows the statistics and the setter chooses a phrase that makes the statistics fail, and the best rounds happen when both are thinking about what the other expects.

A small worked example

Suppose the board shows a five letter word that ends in E, with the other four letters still blank, and you have already ruled out A. Every guess now has a different value than it did at the start. Letters that commonly sit just before a final E, such as T, L and R, become more promising, and a letter that rarely appears in that position falls away. You are not computing anything exact. You are weighing, by feel, how much of the remaining possibility each letter would clear away. That is probability in its everyday form: a rough ranking of options that you revise every time the board changes.

A good decision and a good outcome are different things

A guess can be sensible and still miss. If you open with E on a phrase that happens to contain none, the decision was sound and the outcome poor. The reverse also happens: a wild guess of Z might land by luck and look brilliant. Over many rounds the sensible guesser wins more, but in any single round luck has a say. Recognising this is useful well beyond the game, because it stops you from judging a choice only by how it turned out. After a lost round, ask whether you chose well with what you knew, not just whether you won.

A note on honesty

It would be easy to overstate this. Playing hangman does not train a person in statistics, and nobody becomes numerate from a pen and a napkin. What it does is give a feel for ideas that are otherwise abstract: that a distribution has a shape, that probabilities change with information and that the best action is the one that learns the most. That feel is a useful thing to have before the formal ideas arrive, and a genuinely enjoyable way to build it.

Using it in teaching

A teacher can use a round of hangman as a short demonstration. Ask the class to predict which letters will appear, tally their guesses against the result and discuss why the common ones usually do. Then reveal a phrase that defeats the usual strategy and ask why it did. The discussion covers distributions, prediction and the difference between a good decision and a good outcome, in about ten minutes and with no equipment. The hard list on this site contains plenty of phrases that defeat the usual strategy, with explanations of why.