Benford's Law

An empirical regularity that for many naturally occurring numerical datasets the distribution of leading (first) decimal digits d = 1,...,9 follows P(d) = log10(1 + 1/d), so smaller digits (especially 1) appear as leading digits far more often than larger ones, typically when data span multiple orders of magnitude and are not artificially constrained.

An empirical regularity about the distribution of leading (most significant) decimal digits in many naturally occurring numerical datasets: the probability that the first significant digit equals d (1≤d≤9) is log10(1 + 1/d). The law applies under conditions such as scale invariance and datasets spanning multiple orders of magnitude.