Evaluating the Discrimination Ability of Proper Multivariate Scoring Rules

TL;DR

This paper compares the discrimination ability of energy and variogram scores in multivariate forecasting, using realistic simulation based on diverse financial data, and finds the variogram score with p=0.5 performs best.

Contribution

It provides a comprehensive simulation-based comparison of proper multivariate scoring rules, highlighting the superior discrimination ability of a specific variogram score.

Findings

Variogram score with p=0.5 outperforms energy score.

Simulation design is more realistic with diverse financial data.

The study offers a thorough evaluation of scoring rules in various settings.

Abstract

Proper scoring rules are commonly applied to quantify the accuracy of distribution forecasts. Given an observation they assign a scalar score to each distribution forecast, with the the lowest expected score attributed to the true distribution. The energy and variogram scores are two rules that have recently gained some popularity in multivariate settings because their computation does not require a forecast to have parametric density function and so they are broadly applicable. Here we conduct a simulation study to compare the discrimination ability between the energy score and three variogram scores. Compared with other studies, our simulation design is more realistic because it is supported by a historical data set containing commodity prices, currencies and interest rates, and our data generating processes include a diverse selection of models with different marginal distributions, dependence structure, and calibration windows. This facilitates a comprehensive comparison of the performance of proper scoring rules in different settings. To compare the scores we use three metrics: the mean relative score, error rate and a generalised discrimination heuristic. Overall, we find that the variogram score with parameter p=0.5 outperforms the energy score and the other two variogram scores.