Bounds for the accuracy of invalid normal approximation

TL;DR

This paper investigates the accuracy of normal approximation in situations where classical conditions are not met, providing bounds and explanations for its reliability and limitations in applied probability contexts.

Contribution

It offers theoretical bounds for the accuracy of the normal approximation when classical assumptions are invalid, and explains the behavior of convergence in stable law domains.

Findings

Bounds for the accuracy of normal approximation in invalid cases

Explanation of the initial decrease and subsequent increase in approximation error

Insights into the behavior of sums in the domain of attraction of stable laws

Abstract

In applied probability, the normal approximation is often used for the distribution of data with assumed additive structure. This tradition is based on the central limit theorem for sums of (independent) random variables. However, it is practically impossible to check the conditions providing the validity of the central limit theorem when the observed sample size is limited. Therefore it is very important to know what the real accuracy of the normal approximation is in the cases where it is used despite it is theoretically inapplicable. Moreover, in some situations related with computer simulation, if the distributions of separate summands in the sum belong to the domain of attraction of a stable law with characteristic exponent less than two, then the observed distance between the distribution of the normalized sum and the normal law first decreases as the number of summands grows and begins to increase only when the number of summands becomes large enough. In the present paper an attempt is undertaken to give some theoretical explanation to this effect.