Non-probabilistic odds and forecasting with imperfect models

TL;DR

This paper proposes non-probabilistic odds forecasts as an alternative to probability forecasts, aiming to better represent uncertainties including model errors, using a game-theoretic approach for scientific evaluation and practical applications.

Contribution

It introduces a novel game-theoretic framework for deriving odds forecasts that account for uncertainties beyond randomness, addressing limitations of traditional probability forecasts.

Findings

Odds forecasts can be calculated for various applications.

They provide a less distorted view of forecast uncertainties.

The approach is applicable to investment, loss mitigation, and weather forecasting.

Abstract

Probability forecasts are intended to account for the uncertainties inherent in forecasting. It is suggested that from an end-user's point of view probability is not necessarily sufficient to reflect uncertainties that are not simply the result of complexity or randomness, for example, probability forecasts may not adequately account for uncertainties due to model error. It is suggested that an alternative forecast product is to issue non-probabilistic odds forecasts, which may be as useful to end-users, and give a less distorted account of the uncertainties of a forecast. Our analysis of odds forecasts derives from game theory using the principle that if forecasters truly believe their forecasts, then they should take bets at the odds they offer and not expect to be bankrupted. Despite this game theoretic approach, it is not a market or economic evaluation; it is intended to be a scientific evaluation. Illustrative examples are given of the calculation of odds forecasts and their application to investment, loss mitigation and ensemble weather forecasting.