Assessing Extrapolation of Peaks Over Thresholds with Martingale Testing
By: Joseph de Vilmarest, Olivier Wintenberger
Potential Business Impact:
Predicts rare, heavy rainstorms with better accuracy.
We present the winning strategy for the EVA2025 Data Challenge, which aimed to estimate the probability of extreme precipitation events. These events occurred at most once in the dataset making the challenge fundamentally one of extrapolating extreme values. Given the scarcity of extreme events, we argue that a simple, robust modeling approach is essential. We adopt univariate models instead of multivariate ones and model Peaks Over Thresholds using Extreme Value Theory. Specifically, we fit an exponential distribution to model exceedances of the target variable above a high quantile (after seasonal adjustment). The novelty of our approach lies in using martingale testing to evaluate the extrapolation power of the procedure and to agnostically select the level of the high quantile. While this method has several limitations, we believe that framing extrapolation as a game opens the door to other agnostic approaches in Extreme Value Analysis.
Similar Papers
Statistical Prediction of Peaks Over a Threshold
Methodology
Predicts future extreme events, like heatwaves or crashes.
Extrapolating into the Extremes with Minimum Distance Estimation
Methodology
Predicts extreme weather events more accurately.
A novel statistical workflow for nonstationary modelling of successive Fréchet extremes
Statistics Theory
Predicts when big energy needs will happen.