0.999… = 1

Arithmetic anyone can check settles it. Let x be 0.999… Multiply by ten and you get 9.999… Subtract the first line from the second: the infinite tail cancels exactly, leaving 9x = 9, so x = 1. Nothing here is rounded and nothing is approximated. The same result falls out of the familiar fact that one third is 0.333… — multiply both sides by three and the left side is 1 while the right side is 0.999…

Readers always raise the same objection, and it is worth taking seriously rather than waving away: surely there is some gap, however small. So name it. If 0.999… and 1 are different real numbers, their difference is a positive real number, and between any two distinct real numbers there is a third. Write down a number strictly between 0.999… and 1. It cannot be done, because any candidate you propose is smaller than 0.999… at some decimal place. Two real numbers with nothing between them are the same number.

The deeper answer is that decimal notation is a definition, not a picture. The string 0.999… is defined as the limit of the sequence 0.9, 0.99, 0.999, and so on — that is what the ellipsis means in the real numbers. That limit is 1, not something near 1. The notation does not describe a process that is still running; it names the value the process converges to.

This is where the intuition actually goes wrong. People read 0.999… as a number that is forever approaching one, because the way we write it looks like an unfinished action. But a decimal expansion is a completed object, and the real numbers contain no infinitesimals — no positive quantity smaller than every fraction 1/n. Without infinitesimals there is no room for the gap the intuition insists on.

One honest caveat, because the shirt says EXACTLY and it should earn the word. There are number systems that do contain infinitesimals — the hyperreals of Abraham Robinson's non-standard analysis, and the surreal numbers. In those systems you can define objects that differ from 1 by an infinitesimal amount. But 0.999… is not one of them: the standard decimal expansion still denotes 1 there too, and the infinitesimal quantities have no decimal representation at all. The equality holds in the reals, which is where the shirt is written.

The number of people who have argued about this online is itself a documented phenomenon. It appears in mathematics-education literature as a standard example of a correct result that conflicts with a robust and wrong intuition — which is exactly the reason it is in this archive.