Nim vs Chomp: Which Combinatorial Game to Play?

Nim

Chomp

The Verdict

Play Nim if you want a game you can genuinely master: Charles Bouton solved it completely in 1901 — XOR the heap sizes and always hand your opponent a zero. Play Chomp for the beautiful open problem: David Gale's 1974 strategy-stealing argument proves the first player wins, yet on large grids nobody knows how. Nim rewards learned technique; Chomp rewards live exploration.

Side by Side

	Nim	Chomp
Players	2	2
Equipment	Heaps of counters (classically 3, 5 and 7)	Rectangular grid with one poisoned corner
Win condition	Take the last counter (normal play)	Avoid taking the poisoned square
Average game length	2–5 min	2–5 min
Luck vs skill	100% skill	100% skill
Solved status	Fully solved 1901 (Bouton's nim-sum)	First-player win proven (Gale, 1974); moves unknown in general
Misère play	Standard variant — last counter loses	Built in — the final square is the losing one
First appeared	Ancient folk roots; analysed and named 1901	1974, David Gale (equivalent to Schuh's 1952 divisor game)

How the two games work

Nim could hardly be simpler to state: several heaps of counters sit on the table, each turn you remove any number of counters from one heap, and in normal play whoever takes the last counter wins. Chomp is played on a rectangular grid — picture a chocolate bar whose top-left square is poisoned: each turn you pick a square and eat it together with every square below and to its right, and whoever is finally forced to eat the poisoned square loses. Both are two-player, perfect-information games with no luck, and both finish inside five minutes.

Solved versus proven-but-unknown

Nim is completely solved. In 1901 Harvard's Charles Bouton proved a position is lost for the player to move exactly when its nim-sum — the XOR of the heap sizes in binary — is zero; the winning strategy is simply to restore the nim-sum to zero every turn. Chomp's status is stranger. David Gale proved in 1974 that the first player wins on any grid bigger than 1×1 by strategy-stealing: if the second player had a winning reply to the smallest possible first bite, the first player could have made that winning move directly. The proof names no moves — computers have found them for small boards, but no general rule is known.

Which should you learn first?

Start with Nim: it is the front door of combinatorial game theory, and the payoff for one evening's study is total — once you can XOR heap sizes in your head, you beat anyone who cannot. Then let Chomp humble you: knowing you should win as the first player while owning no formula for the moves is a genuinely different experience, and even 3-wide boards get hard fast. Together they bracket the whole field — one game where mathematics answers everything, one where a two-line proof answers almost nothing.

Pick Nim if…

you want a game with a complete, learnable winning formula
you enjoy mental binary arithmetic as a party trick
you want the canonical entry point to combinatorial game theory
you like playing with matchsticks, coins or any counters at hand

Pick Chomp if…

you want a game even its inventor could not play perfectly
you enjoy fresh puzzles over memorised technique
you like the drama of a poisoned square closing in
you want to feel a non-constructive proof from the inside

FAQ

What is the winning strategy in Nim?

Compute the nim-sum: write each heap size in binary and XOR them together. If the nim-sum is non-zero, a winning move always exists — remove counters so the nim-sum becomes zero, and repeat every turn. Charles Bouton published this complete solution in 1901.

Who wins Chomp with perfect play?

The first player, on any grid larger than 1×1. David Gale's strategy-stealing argument proves it: if the second player had a winning answer to the first player eating just the bottom-right square, the first player could have played that winning move themselves — a contradiction. The proof reveals no actual moves.

What is the difference between normal and misère Nim?

Normal play means taking the last counter wins; misère play means it loses. The famous match-stick game in Last Year at Marienbad is misère Nim. The optimal strategies are identical until the endgame, where misère play flips one decision about leaving heaps of size one.

Is Chomp a solved game?

Only partially. The first player is proven to win any grid bigger than 1×1, and computers have calculated exact winning moves for small boards, but no general winning strategy is known — an unusual gap between knowing who wins and knowing how.

What is a combinatorial game?

A two-player game with perfect information, no chance, and alternating moves — think Nim, Chomp, Hex or Go endgames. Nim is the family's cornerstone: the Sprague–Grundy theorem shows every impartial combinatorial game is equivalent to some Nim heap.

Are Nim and Chomp good games for children?

Excellent ones. Both teach in under a minute, run 2–5 minutes per game, and reward pattern-hunting; Nim doubles as a stealth introduction to binary numbers. A chocolate bar makes Chomp self-enforcing — eaten squares stay eaten.

← All comparisons