I wonder if there is a name or reference for the following fact. It is not the proof I am looking for.

Let $s_1, s_2, ...,s_n$ be non-negative real numbers ordered in a non-increasing way. Let $b_1,b_2,...,b_n$ be non-negative real numbers ordered in a non-decreasing way (so opposite of the $s_i$).

Then the average value $$\frac{s_1+s_2+\cdots +s_n}{n}$$ is at least as large as the weighted average $$\frac{b_1 s_1+\cdots+b_n s_n}{b_1+\cdots+b_n}.$$



This is known as Chebyshev's Sum Inequality. (I've only ever seen it used in the context of competition math, but Wikipedia gives a reference to Hardy-Littlewood-Polya.)

  • 1
    $\begingroup$ I've seen it used at least once by K. Matomäki (in "On signs of Fourier coefficients of cusp forms", Math. Proc. Cambridge Philos. Soc. 152 (2012), no. 2, 207–222, if I remember correctly). $\endgroup$ – Denis Chaperon de Lauzières Mar 9 '18 at 9:07

It is called often called "Chebyshev's other inequality". And this makes more sense than the name "Chebyshev's sum inequality" proposed by Wikipedia, since even Wikepedia knows that there is a version for integrals.

  • 7
    $\begingroup$ Integrals aren’t sums? $\endgroup$ – Geoffrey Irving Mar 9 '18 at 0:06
  • $\begingroup$ So one of my physics teachers told us. But I could understand it only after a transposition. $\endgroup$ – Lutz Mattner Mar 9 '18 at 0:14
  • $\begingroup$ @LutzMattner : ... so integrals are permutations of sums? :-) $\endgroup$ – Eric Towers Mar 9 '18 at 2:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.