Scoring Rules for Performative Binary Prediction

TL;DR

This paper investigates how traditional scoring rules can incentivize manipulation in performative binary prediction settings and proposes a new class of scoring rules that prevent such manipulation.

Contribution

It introduces a model of performative prediction where predictions influence outcomes and designs scoring rules that discourage manipulation.

Findings

Proper scoring rules can incentivize manipulation in performative settings

A simple class of scoring rules is constructed to avoid manipulation

Theoretical and numerical results support the proposed scoring rules

Abstract

We construct a model of expert prediction where predictions can influence the state of the world. Under this model, we show through theoretical and numerical results that proper scoring rules can incentivize experts to manipulate the world with their predictions. We also construct a simple class of scoring rules that avoids this problem.