Plotting positions close to the exact unbiased solution: application to the Pozzuoli's bradeysism earthquake data

TL;DR

This paper introduces a new plotting position formula that closely approximates the exact unbiased solution, improving the accuracy of graphical analysis for critical data like earthquake magnitudes, and unifies existing methods.

Contribution

A novel general plotting position formula that approximates the exact unbiased positions, applicable to various distributions, and clarifies ongoing debates in the literature.

Findings

Monte Carlo simulations confirm high accuracy for symmetrical and skewed distributions.

Application to 1983-1984 Campi Flegrei earthquake data demonstrates practical utility.

The new formula unifies and clarifies existing plotting position methods.

Abstract

Graphical techniques are recommended for critical applications in order to share information with non-statisticians, since they allow for a visual analysis and helpful understanding of the results. However, graphical estimation methods are often underestimated because of their minor efficiency with respect to the analytical ones. Therefore, finding unbiased plotting positions can contribute to rise their reputation and to encourage their strategic use. This paper proposes a new general plotting position formula which can be as close as needed to the exact unbiased plotting positions. The ability of the new solution in estimating quantiles for both symmetrical and skewed location-scale distributions is shown via Monte Carlo simulation. An applicative example shows how the proposed formula enables to perform, with known accuracy, the graphical analysis of critical data, such as the earthquake magnitudes registered during the serious 1983-1984 bradyseismic crisis in Campi Flegrei (Italy). Moreover, the proposed formula gives a unified look at existing plotting positions and a definitive insight into plotting position controversies recently renewed in the literature.