aiqbench analysis
This paper resolves the theoretical gap in sequential prediction by establishing a tight $\Omega(\sqrt{T})$ lower bound for smooth calibration error against adaptive adversaries. By forcing forecasters to incur higher costs when compensating for past residuals, the authors settle the minimax rates for these robustness measures, with proof concepts partially aided by AI tools.
A Matching Lower Bound for Online Smooth Calibration
From arXiv
Full textarXiv:2611.16097v1 Announce Type: new Abstract: We study smooth calibration error in sequential binary prediction, which measures how far a sequence of probabilistic forecasts is from being calibrated and, unlike the classical expected calibration error, is robust to small perturbations of the forecasts. Against adaptive adversaries, the minimax smooth calibration error is known to lie between $\Omega(T^{1/3})$ and $O(\sqrt{T})$ (Qiao and Zheng, 2024), where $T$ is the number of rounds. In this paper, we close this gap by constructing a new adversary whose outcome mean moves with the forecaster's cumulative residual.
This design makes compensating for past residuals costly, and forces every forecaster to incur an expected smooth calibration error of $\Omega(\sqrt{T})$. This lower bound matches the known upper bound up to universal constants, and further implies $\Omega(\sqrt{T})$ lower bounds for calibration distance and lower calibration distance. Overall, our results settle the minimax rates of these calibration measures in sequential prediction. The proof ideas are partially developed by GPT-6 Astra under the authors' guidance.
Abstract:We study smooth calibration error in sequential binary prediction, which measures how far a sequence of probabilistic forecasts is from being calibrated and, unlike the classical expected calibration error, is robust to small perturbations of the forecasts. Against adaptive adversaries, the minimax smooth calibration error is known to lie between $\Omega(T^{1/3})$ and $O(\sqrt{T})$ (Qiao and Zheng, 2024), where $T$ is the number of rounds. In this paper, we close this gap by constructing a new adversary whose outcome mean moves with the forecaster's cumulative residual.
This design makes compensating for past residuals costly, and forces every forecaster to incur an expected smooth calibration error of $\Omega(\sqrt{T})$. This lower bound matches the known upper bound up to universal constants, and further implies $\Omega(\sqrt{T})$ lower bounds for calibration distance and lower calibration distance. Overall, our results settle the minimax rates of these calibration measures in sequential prediction. The proof ideas are partially developed by GPT-6 Astra under the authors' guidance.
Machine Learning (cs.LG); Machine Learning (stat.ML)
arXiv:2611.16097 [cs.LG]
(or arXiv:2611.16097v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2611.16097 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 255 words
Related coverage
Strength Controls Change How Interventions Compare Across Sparse Autoencoder Codes
arXiv · Oct 21, 2026
Extractable Memorization From First Principles
arXiv · Sep 30, 2026
Winner's Curse Drives False Promises in Data-Driven Decisions
arXiv · Oct 27, 2026
Damage Predicts Recovery: When Calibration Data Matters in Compressing Financial LLMs
arXiv · Sep 23, 2026