Smooth Partial Lotteries for Stable Randomized Selection

Alexander Goldberg, Giulia Fanti, Nihar B. Shah · 2026

Read Smooth Partial Lotteries for Stable Randomized Selection by Alexander Goldberg, Giulia Fanti, Nihar B. Shah (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-20069.txt.

Description

Alexander Goldberg, Giulia Fanti and Nihar Shah design partial lotteries that stay stable when scores shift. Lottery designs let a small change in one candidate's score cause large swings in selection probability, undermining the goal of softening fine distinctions near the cutoff. They formalise smoothness as a Lipschitz condition mapping scores to selection probabilities. Their Clipped Linear Lottery scales probability linearly with estimated quality between always-accept and always-reject thresholds. It matches the best regret for smooth rules up to a factor, and beats alternatives on peer-review data.

Catalog record

CREATORAlexander Goldberg, Giulia Fanti, Nihar B. Shah
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE84031
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.20069

Browse the full library · About the Tower of Babel