Decision tree analysis
Valuation & CompsValuing each outcome with the management decision that would follow it built in, then probability weighting, which captures optionality without option pricing.
Also written: decision tree, scenario tree
A decision tree separates the two things a scenario analysis usually blends: what happens to the world, and what management does about it. Each branch carries a probability and its own operating response, so the bad state is valued with the project stopped rather than with the project ploughing on.
That single change is what captures the option. Averaging outcomes under one fixed plan destroys the asymmetry that gives flexibility its value; averaging outcomes under the plan that would actually be followed preserves it.
It is the version practitioners use, and the reason is defensibility rather than elegance. Every input is a forecast or a probability that can be argued about explicitly, where a lattice or Black Scholes valuation of a real asset rests on an unobservable volatility and is therefore easy to attack in a room.
Its weakness is honest and worth stating: the probabilities are judgements, and the answer moves a long way when they move. Making them explicit rather than burying them is the whole discipline, which is the same argument that governs a probability weighted pipeline valuation in pharmaceuticals.
Worked example
A mine faces an equal chance of €9,000 and €5,000 copper, with a cash cost of €7,000 on 10,000 tonnes.
One plan for both states averages 20 and negative 20, giving nothing. A tree that stops production in the low state averages 20 and negative 3, giving 8.5.
Same forecasts, same probabilities. The only difference is that the second version lets management act, and that difference is the option value.