Large deviations theory on a t-shirt — Cramér's theorem, distilled to one line. The sample mean of N i.i.d. draws almost never strays far from its expectation, but when it does, the probability decays exponentially in N: P( (1/N) sum X_i >= x ) ~ exp( -N I(x) ) where I(x) is the rate function. Topped with the theory's famous heuristic — "any large deviation is done in the least unlikely of all the unlikely ways" (Frank den Hollander) — and signed off with: let your deviation count. For probabilists, quants, statisticians, data scientists, and anyone who has ever computed a tail bound for fun. Minimal monochrome line art. Vector design — crisp at every print size.