EconBase
← All papers

Selecting the Best Arm in One-Shot Multi-Arm RCTs: The Asymptotic Minimax-Regret Decision Framework for the Best-Population Selection Problem

Joonhwi Joo

arXiv 4 Sep 2025 · Econometrics

arXiv:2509.03796 · PDF · Extracted main text

Abstract

We develop a frequentist decision-theoretic framework for selecting the best arm in one-shot, multi-arm randomized controlled trials (RCTs). Our approach characterizes the minimax-regret (MMR) optimal decision rule for any location-family reward distribution with full support. We show that the MMR rule is deterministic, unique, and computationally tractable, as it can be derived by solving the dual problem with nature's least-favorable prior. We then specialize to the case of multivariate normal (MVN) rewards with an arbitrary covariance matrix, and establish the local asymptotic minimaxity of a plug-in version of the rule when only estimated means and covariances are available. This asymptotic MMR (AMMR) procedure maps a covariance-matrix estimate directly into decision boundaries, allowing straightforward implementation in practice. Our analysis highlights a sharp contrast between two-arm and multi-arm designs. With two arms, the empirical success rule ("pick-the-winner") remains MMR-optimal, regardless of the arm-specific variances. By contrast, with three or more arms and heterogeneous variances, the empirical success rule is no longer optimal: the MMR decision boundaries become nonlinear and systematically penalize high-variance arms, requiring stronger evidence to select them. This result underscores that variance plays no role in optimal two-arm comparisons, but it matters critically when more than two options are on the table. Our multi-arm AMMR framework extends classical decision theory to multi-arm RCTs, offering a rigorous foundation and a practical tool for comparing multiple policies simultaneously.

Citation extraction

47
references
83
in-text mentions
47
distinct cited
1
self-citations
11,589
main-text words

appendix boundary found by appendix_command · 42% of the source is main text. Read the extracted text to check this.

Most heavily cited references

The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.

ReferenceIntensityMentionsSectionsMain text
1Tetenov, A (2012) Statistical treatment choice based on asymmetric minimax regret criteria0.87462100%
2Hirano, K. and J. R. Porter (2009) Asymptotics for statistical treatment rules0.87452100%
3Joo, J. and K. X. Chiong (Forthcoming): Getting the Most Out of A/B… self0.87452100%
4van der Vaart, A. W (1998) Asymptotic Statistics0.7374275%
5Stoye, J (2009) Minimax regret treatment choice with finite samples0.73732100%
6Stoye, J (2012) Minimax regret treatment choice with covariates or with limited validity of experiments0.73732100%
7Liese, F. and K.-J. Miescke (2008) Statistical Decision Theory - Estimation, Testing, and Selection0.6936433%
8Kitagawa, T., S. Lee, and C. Qiu (2023) Treatment choice with nonlinear regret0.64422100%
9Manski, C. F (2004) Statistical treatment rules for heterogeneous populations0.64422100%
10Blackwell, D. and M. Girshick (1954) Theory of Games and Statistical Decisions0.5113233%

Showing the top 10 of 47 scored citations.

Cited by, within the corpus

arXiv econ.EM papers that cite this one, ranked by how heavily they lean on it.

Citing paperIntensityMentionsSections
1Valuing Winners: When and How to Correct for Selection Bias in Randomized Experiments0.40511