method of moments.

On the Error of Approximation by Means of Scenario Trees with Depth 1

Let¤n denote the set of scenario trees with depth 1 and n scenarios. LetX = (0 · x1 < . . . < xn · 1) and let¤n(X) denote

the set of all scenario trees of depth 1 with the scenarios X = (0 · x1 < . . . < xn · 1). Let G be a probability distribution

defined on [0, 1] and H be a subset of measurable functions defined on [0, 1]. Let dH,X(G) = inf ˜G∈¤n(X) dH(G, ˜ G) and

dH(G) = inf ˜G∈¤n

dH(G, ˜ G), where dH(G, ˜ G) := suph∈H

¯¯¯

R h dG − R h d˜G

¯¯¯

. The main goal of the paper is to estimate