Research

Paper

TESTING March 13, 2026

Breaking the Winner's Curse with Bayesian Hybrid Shrinkage

Authors

Richard Mudd, Abbas Zaidi, Rina Friedberg, Ilya Gorbachev, Anchal Choubey, Houssam Nassif

Abstract

The widespread adoption of randomized controlled trials (A/B Tests) for decision-making has introduced a pervasive "Winner's Curse": experiments selected for launch often exhibit upwardly biased effect estimates and invalid confidence intervals. This selection bias leads to over-optimistic impact projections and undermines decision-making, particularly in low-power regimes. We propose Bayesian Hybrid Shrinkage (BHS), an empirical Bayes (EB) framework that leverages data-driven priors to mitigate selection bias and provides accurate uncertainty quantification. Unlike traditional EB methods that apply uniform shrinkage, BHS introduces an experiment-specific "local" shrinkage factor that incorporates individual experiment characteristics, improving robustness against prior misspecification. We also derive a closed-form inference strategy designed for high-throughput production environments. Extensive simulations and real-world evaluations at Meta Platforms demonstrate that BHS outperforms existing methods in terms of bias reduction and interval coverage, even under substantial violations of modeling assumptions.

Metadata

arXiv ID: 2603.12867
Provider: ARXIV
Primary Category: stat.ME
Published: 2026-03-13
Fetched: 2026-03-16 06:01

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.12867v1</id>\n    <title>Breaking the Winner's Curse with Bayesian Hybrid Shrinkage</title>\n    <updated>2026-03-13T10:13:38Z</updated>\n    <link href='https://arxiv.org/abs/2603.12867v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.12867v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>The widespread adoption of randomized controlled trials (A/B Tests) for decision-making has introduced a pervasive \"Winner's Curse\": experiments selected for launch often exhibit upwardly biased effect estimates and invalid confidence intervals. This selection bias leads to over-optimistic impact projections and undermines decision-making, particularly in low-power regimes. We propose Bayesian Hybrid Shrinkage (BHS), an empirical Bayes (EB) framework that leverages data-driven priors to mitigate selection bias and provides accurate uncertainty quantification. Unlike traditional EB methods that apply uniform shrinkage, BHS introduces an experiment-specific \"local\" shrinkage factor that incorporates individual experiment characteristics, improving robustness against prior misspecification. We also derive a closed-form inference strategy designed for high-throughput production environments. Extensive simulations and real-world evaluations at Meta Platforms demonstrate that BHS outperforms existing methods in terms of bias reduction and interval coverage, even under substantial violations of modeling assumptions.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='stat.ME'/>\n    <published>2026-03-13T10:13:38Z</published>\n    <arxiv:primary_category term='stat.ME'/>\n    <author>\n      <name>Richard Mudd</name>\n    </author>\n    <author>\n      <name>Abbas Zaidi</name>\n    </author>\n    <author>\n      <name>Rina Friedberg</name>\n    </author>\n    <author>\n      <name>Ilya Gorbachev</name>\n    </author>\n    <author>\n      <name>Anchal Choubey</name>\n    </author>\n    <author>\n      <name>Houssam Nassif</name>\n    </author>\n  </entry>"
}