Canonical: https://knowlpedia.com/probability/pinsker-inequality/

# Pinsker's inequality

An inequality bounding total variation distance by the square root of Kullback–Leibler divergence.

**Pinsker's inequality:** Let \(P\) and \(Q\) be [probability measures](https://knowlpedia.com/probability/probability-measure/) on the same \((\Omega,\mathcal F)\). Then their [total variation distance](https://knowlpedia.com/probability/total-variation-distance/) satisfies
\[
d_{\mathrm{TV}}(P,Q)\;\le\;\sqrt{\frac12\,D(P\|Q)},
\]

where \(D(P\|Q)\) is the [Kullback–Leibler divergence](https://knowlpedia.com/probability/relative-entropy-kl-divergence/) computed with natural logarithms (if another log base is used, the constant changes accordingly). The inequality is understood to hold trivially when \(D(P\|Q)=+\infty\).

Pinsker's inequality formalizes that small relative entropy forces two laws to be close in a strong, event-wise sense. Together with [Gibbs' inequality](https://knowlpedia.com/probability/gibbs-inequality-kl/), it highlights KL divergence as a nonnegative discrepancy measure that quantitatively controls \(d_{\mathrm{TV}}\).
