Popoviciu's inequality

Not to be confused with Popoviciu's inequality on variances.

For a more general discussion, see Bookmaker and Odds.

In convex analysis, Popoviciu's inequality is an inequality about convex functions. It is similar to Jensen's inequality and was found in 1965 by Tiberiu Popoviciu, a Romanian mathematician. It states:

Let f be a function from an interval $I \subseteq \mathbb{R}$ to $\mathbb{R}$. If f is convex, then for any three points x, y, z in I,

$\frac{f(x)+f(y)+f(z)}{3} + f\left(\frac{x+y+z}{3}\right) \ge \frac{2}{3}\left[ f\left(\frac{x+y}{2}\right) + f\left(\frac{y+z}{2}\right) + f\left(\frac{z+x}{2}\right) \right].$

If a function f is continuous, then it is convex if and only if the above inequality holds for all x, y, z from $I$. When f is strictly convex, the inequality is strict except for x = y = z.

It can be generalised to any finite number n of points instead of 3, taken on the right-hand side k at a time instead of 2 at a time:

Let f be a continuous function from an interval $I \subseteq \mathbb{R}$ to $\mathbb{R}$. Then f is convex if and only if, for any integers n and k where n ≥ 3 and $2 \leq k \leq n-1$, and any n points $x_1, \dots, x_n$ from I,

$\frac{1}{k} \binom{n-2}{k-2} \left( \frac{n-k}{k-1} \sum_{i=1}^{n}f(x_i) + nf\left(\frac1n\sum_{i=1}^{n}x_i\right) \right)\ge \sum_{1 \le i_1 < \dots < i_k \le n} f\left( \frac1k \sum_{j=1}^{k} x_{i_j} \right)$

Popoviciu's inequality can also be generalised to a weighted inequality.