Why scoring functions cannot assess tail properties

DOI10.1214/19-EJS1622zbMath1429.62045arXiv1905.04233MaRDI QIDQ2326990

Kirstin Strokorb, Jonas R. Brehmer

Publication date: 11 October 2019

Published in: Electronic Journal of Statistics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1905.04233

zbMATH Keywords

consistencyextreme value indexelicitabilityscoring functionsproper scoring ruletail equivalenceelicitation complexitymax-functional

Mathematics Subject Classification ID

Decision theory (91B06) Statistics of extreme values; tail inference (62G32) General considerations in statistical decision theory (62C05)

Related Items (5)

Modeling Nonstationary Extreme Dependence With Stationary Max-Stable Processes and Multidimensional Scaling ⋮ Evaluating Forecasts for High-Impact Events Using Transformed Kernel Scores ⋮ Scoring interval forecasts: equal-tailed, shortest, and modal interval ⋮ Forecast evaluation of quantiles, prediction intervals, and other set-valued functionals ⋮ Scoring predictions at extreme quantiles

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Making and Evaluating Point Forecasts
• Higher order elicitability and Osband's principle
• Likelihood-based scoring rules for comparing density forecasts in tails
• The geometry of proper scoring rules
• New results for tails of probability distributions according to their asymptotic decay
• Focusing on regions of interest in forecast evaluation
• Forecaster's dilemma: extreme events and forecast evaluation
• Elicitability and identifiability of set-valued measures of systemic risk
• On the indirect elicitability of the mode and modal interval
• Order-sensitivity and equivariance of scoring functions
• Comparing Density Forecasts Using Threshold- and Quantile-Weighted Scoring Rules
• Should Scoring Rules be “Effective”?
• Statistics for near independence in multivariate extreme values
• Scoring Rules for Continuous Probability Distributions
• On elicitable risk measures
• Elicitation complexity of statistical properties
• Strictly Proper Scoring Rules, Prediction, and Estimation
• The mode functional is not elicitable
• DISTRIBUTION‐INVARIANT RISK MEASURES, INFORMATION, AND DYNAMIC CONSISTENCY

This page was built for publication: Why scoring functions cannot assess tail properties