A hypothesis-testing perspective on the G-normal distribution theory

TL;DR

This paper investigates the tail behavior of the G-normal distribution using a nonlinear heat equation approach, providing asymptotic results that improve tail probability estimation and reveal vulnerabilities in heteroscedastic hypothesis testing.

Contribution

It introduces a novel analysis of the G-normal distribution's tail behavior via a nonlinear heat equation, impacting hypothesis testing for heteroscedastic data.

Findings

Asymptotic tail probability estimates with high accuracy.

Potential for manipulating heteroscedastic data to falsely attain significance.

Insights into the limitations of current hypothesis testing methods.

Abstract

The G-normal distribution was introduced by Peng [2007] as the limiting distribution in the central limit theorem for sublinear expectation spaces. Equivalently, it can be interpreted as the solution to a stochastic control problem where we have a sequence of random variables, whose variances can be chosen based on all past information. In this note we study the tail behavior of the G-normal distribution through analyzing a nonlinear heat equation. Asymptotic results are provided so that the tail "probabilities" can be easily evaluated with high accuracy. This study also has a significant impact on the hypothesis testing theory for heteroscedastic data; we show that even if the data are generated under the null hypothesis, it is possible to cheat and attain statistical significance by sequentially manipulating the error variances of the observations.