The Setting
Let and be real random variables on the same finite probability space, with the same distribution. For a real function on the cube with the uniform measure and a coordinate , and , where is with its -th coordinate reversed, are such a pair. Suppose that exceeds by at most wherever the probability is positive, with no bound on . The question is how large the mean square of can be in terms of and the spread of , measured by its second moment about a point .
Definitions
A finite probability space is a finite set with a weight at each point , the weights summing to . A real random variable on a finite probability space is a function , and the expectation of is (1)
Two real random variables and on have the same distribution if for every function . An inequality between random variables on holds almost surely if it holds at every point with . The second moment of about a point is , the variance of when .
For a point , the coupling potential about is the function given by (2)✦
The Claim
Theorem 1. (a) Let be a finite probability space and let and be real random variables on it with the same distribution. Let , and , and suppose that the second moment of about is and that almost surely. Then (3)✦
(b) Let and let be the coupling potential about . For all real numbers , (4)✦ and for all real numbers and , (5)✦
In the paper, the result reads: if and are real random variables with the same distribution, finite mean and finite variance , and almost surely, where , then .
What Is Checked
The claim is stated over a finite probability space given by nonnegative weights summing to one, two real random variables on it whose expectations and agree for every function , and the coupling potential (2).