Invest, Wait or Abandon (2D)

Decision tree and a regime map of where each strategy wins. Click the map to move the project.

Switch to 3D

Presets

Project

Model check

The science

A project can be started now (cost I), started after one period once the demand has been revealed, or never. Its present value V0 moves to u V0 or d V0 with u = e^sigma, d = 1/u. Waiting forgoes a fraction delta of the value (cash flow lost in the first period).

Risk-neutral probability and backward induction

q = (1 + r - delta - d)/(u - d). Invest now = V0 - I. Wait = [q max(uV0 - I, 0) + (1-q) max(dV0 - I, 0)]/(1+r). The decision is the best of invest now, wait, or 0 (abandon).

Break-even value V* (the white curve on the map)

If only the up state pays: V* = I (1 - q/(1+r)) / (1 - q u/(1+r)). If both states pay: V* = r I / delta (invest now iff the cash yield delta V0 is at least the interest r I). Higher volatility pushes V* up: uncertainty makes waiting more valuable.

The 2D view

The map scans the whole (V0, sigma) plane with the same tree, colouring each cell by its best action. The checks compare the boundary found on the map with the closed form V*, plus the three checks of the 3D page.

Honest simplifications

One period, two states, one source of uncertainty. Real projects have many periods, staged costs and competitors. Pricing with a risk-neutral q assumes the risk can be hedged with traded assets; otherwise q and r are modelling choices. All numbers ($M) are illustrative and not drawn from any real company or market. This is a teaching model, not advice.

Value of each strategy against V0
Value of waiting against volatility
Payoffs at t = 1 by branch
Made with 🟠 Blender