Skip to content

DCT Lexicon · Optimization

Expected value

The average of an outcome over the declared law's futures, estimated by the mean over simulated paths; with curvature it differs from the outcome computed at average inputs.

Drafted for the author's approval from Vol. I, Theorem 8.5 and Section 8.9

In depth

In the book

With curvature, the average of the outcomes differs from the outcome computed at average inputs: for a convex quantity, such as a cost that rises ever faster with volume, the average outcome lies above the plan computed at average inputs (Jensen’s inequality), and for a concave quantity, below it.

Vol. I, Theorem 8.5(ii) (Expected Enterprise Evolution Theorem), verbatim

Executive example

The planned trough is not the average trough. Averaged over paths, each path’s lowest quarter is $384 million, almost $100 million below the plan’s $482 million, because the low point of a path is a concave function of the path and Jensen’s inequality works against the plan (Theorem 8.5(ii)).

Vol. I, Worked Example 8.1, Step 4, verbatim

Common misreading

A common mistake is to present the base case as the expected outcome without checking for curvature.

Vol. I, LOS 8.4, verbatim

How AXIOM uses it

Wherever AXIOM shows an expected value, it is the mean over the simulated paths of the declared law, with its standard error; the plan computed at average inputs is never labelled the expected outcome.

Drafted for the author's approval

In the book: Vol. I, Ch. 8: Stochastic Enterprise Dynamics

Related: Expected value uplift · Monte Carlo simulation · Percentiles (P10, P50, P90) and bands

V1.Thm.8.5 · V1.WE.8.1 · V2.Def.9.1