Georges-Louis Leclerc, Comte de Buffon, worked out in 1777 that dropping a needle onto a floor of parallel lines could estimate the value of pi, using nothing but the chance of where the needle landed. The technique, now known as Buffon’s needle, links a purely random event to one of mathematics’ most exact constants.
more
Buffon started from a game rather than a laboratory. He had a long-standing interest in probability and wanted to show that some probability problems answer to geometry rather than algebra. He imagined two men throwing stick-shaped loaves of bread backward, so they landed on a floor of parallel wooden planks, betting on whether a loaf would fall across the crack between two of them. Working out the loaf length that made the bet fair, he found that the odds of a crossing were tied to pi.
He then swapped the bread for something easier to repeat indoors: a needle and a sheet of paper ruled with evenly spaced parallel lines. The probability that the needle lands touching a line is linked to pi. Cast the needle many times, count the hits, and pi falls out of the arithmetic. As long as the needle is no longer than the gap between the lines, twice the total number of drops divided by the number of times the needle touches a line gives an approximate value for pi.
Why one needle isn’t enough
A single drop tells a reader almost nothing. The method only converges on pi as the number of throws grows, and with too few throws the estimate can be wildly off. When a team running an online version of the experiment for Pi Day tried it themselves with 100 drops, they got 2.63, a result that would make for noticeably wonky circles. Marcus du Sautoy of the University of Oxford, who ran that event, put it plainly: the method carries errors because it is probabilistic, and only a larger pool of drops narrows in on the true value.
The idea keeps finding new floors
The needle is not the only shape that works. A 2024 paper describes a variant that swaps the needle and the planked floor for an equilateral triangle cast onto a grid of square tiles, testing the same probabilistic route to pi with a different geometry and reporting a computer simulation of the method.
The same reasoning turned up somewhere Buffon never imagined. Scientists at the University of Bath studying the ant species Temnothorax albipennis found that scouting ants estimate the area of a potential nest site using two exploratory visits: the first lays down a pheromone trail, the second crosses it repeatedly, and the ant seems to gauge area from how often the second path meets the first. Researchers realised that is the same geometric reasoning as Buffon’s needle, used by an ant with a brain smaller than a grain of salt.
Nearly 250 years after a French naturalist bet on falling bread, the method he devised to justify a wager still describes how a tiny insect measures a home it cannot see.
