The Black-Scholes model prices options by assuming a single, known volatility. In reality, volatility is uncertain — it fluctuates, and nobody knows its exact value. The standard response is to estimate it and hope for the best. But what if the uncertainty itself were built into the pricing framework?
The paper on option pricing under the G-expectation framework (arXiv: 2603.22831) does exactly this. G-expectation, developed by Shige Peng, replaces the classical linear expectation with a nonlinear one that captures volatility ambiguity. The asset price evolves under a G-Brownian motion where the volatility lies in an interval [σ_min, σ_max] rather than at a point.
The result is a G-Black-Scholes equation — a nonlinear PDE that generalizes the classical Black-Scholes equation. The option price becomes an interval rather than a point, reflecting genuine uncertainty about the underlying volatility process.
The authors solve this numerically using finite difference schemes (explicit and implicit) that are provably consistent, stable, monotone, and convergent to the viscosity solution. A logarithmic transformation of asset prices yields an alternative PDE that relaxes stability constraints, improving computational efficiency.
The through-claim: the G-expectation framework doesn't add complexity to capture something optional. It removes an assumption that was never justified — that volatility is precisely known. The nonlinearity of the pricing equation is not a modeling choice but a consequence of honest uncertainty. When you don't pretend to know the volatility, the mathematics becomes nonlinear. The linearity of classical Black-Scholes was the artifact, not the generalization.
2603.22831. Mathematical finance / option pricing / G-expectation / nonlinear PDE / volatility uncertainty.