Breaking Determinism: Stochastic Modeling for Reliable Off-Policy Evaluation in Ad Auctions
Hongseon Yeom, Jaeyoul Shin, Soojin Min, Jeongmin Yoon, Seunghak Yu, Dongyeop Kang
TL;DR
The paper tackles the zero-propensity problem in off-policy evaluation for deterministic ad auctions by modeling the market price distribution through a Discrete Price Model. This environment-centric approach yields a robust approximate propensity score, enabling SNIPS-based offline evaluation that aligns closely with online results. Key contributions include the first validated framework for deterministic auctions using adaptive binning and DPM, with strong empirical support on AuctionNet and NAVER data showing improved directional accuracy and trend tracking. The work promises practical reductions in online experimentation and offers a pathway to extend OPE to other deterministic decision systems.
Abstract
Online A/B testing, the gold standard for evaluating new advertising policies, consumes substantial engineering resources and risks significant revenue loss from deploying underperforming variations. This motivates the use of Off-Policy Evaluation (OPE) for rapid, offline assessment. However, applying OPE to ad auctions is fundamentally more challenging than in domains like recommender systems, where stochastic policies are common. In online ad auctions, it is common for the highest-bidding ad to win the impression, resulting in a deterministic, winner-takes-all setting. This results in zero probability of exposure for non-winning ads, rendering standard OPE estimators inapplicable. We introduce the first principled framework for OPE in deterministic auctions by repurposing the bid landscape model to approximate the propensity score. This model allows us to derive robust approximate propensity scores, enabling the use of stable estimators like Self-Normalized Inverse Propensity Scoring (SNIPS) for counterfactual evaluation. We validate our approach on the AuctionNet simulation benchmark and against 2-weeks online A/B test from a large-scale industrial platform. Our method shows remarkable alignment with online results, achieving a 92\% Mean Directional Accuracy (MDA) in CTR prediction, significantly outperforming the parametric baseline. MDA is the most critical metric for guiding deployment decisions, as it reflects the ability to correctly predict whether a new model will improve or harm performance. This work contributes the first practical and validated framework for reliable OPE in deterministic auction environments, offering an efficient alternative to costly and risky online experiments.