I was rather surprised that it wasn't discrete convolution in the article, after the hospital analogy. Perhaps even with finite summation first, and generalized afterwards; otherwise there's a bit of a leap.