The Role of Prescreening in Auctions with Predictions

2025-02-17

AI Summary

This paper studies how an auction designer can leverage noisy value predictors to implement pre-screening—selecting bidders ex ante based on posterior expected valuations—to maximize revenue in all-pay and first-price auctions. Method: We formulate a Bayesian game model, characterize conditions for the existence of symmetric strictly monotone equilibria, and establish, for the first time, an equivalence between pre-screening auctions and affiliated-type auctions. Contribution/Results: We show that pre-screening substantially improves revenue in all-pay auctions: under high prediction accuracy, the optimal mechanism retains only two bidders; moreover, the optimal number of qualified bidders varies continuously with prediction precision. In contrast, pre-screening is generically redundant in first-price auctions. This work provides the first systematic theoretical framework for prediction-augmented mechanism design and delivers fundamental boundary results on when and how predictive signals enhance auction performance.

Technology Category

• Game Theory and Economic Paradigms: Mechanism Design
• Multiagent Systems: Mechanism Design
• Machine Learning: Learning Preferences or Rankings

Application Category

• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

We consider an auction environment with i.i.d. privately known valuations. Equipped with a noisy predictor, the auction designer receives a coarse signal about each player's valuation, where the signal is fully informative with a given probability. Based on the posterior expectation of the valuations, the designer selects the top players to admit -- a procedure we call emph{prescreening}. We show that this prescreening game is equivalent to a standard auction without prescreening but with emph{correlated} types. Besides, when the signals are always fully informative, these correlated types are emph{affiliated}. We characterize conditions for the existence of a symmetric and strictly monotone equilibrium strategy in both all-pay and first-price auctions. Our results reveal that prescreening can significantly improve the designer's revenue in all-pay auctions; in fact, when the prediction accuracy is one, admitting only two players is optimal. In contrast, prescreening is usually unnecessary in first-price auctions.

Problem

• Analyzes prescreening in auctions with predictors
• Determines optimal strategies for auction revenue
• Compares prescreening effects in different auction types

Innovation

• Uses noisy predictor for valuations
• Applies prescreening to select top players
• Analyzes equilibrium in all-pay auctions