The impact of uncertainty on regularized learning in games
Pierre-Louis Cauvin c,∗, Davide Legacci ∗, Panayotis Mertikopoulos ∗ · 2025
Read The impact of uncertainty on regularized learning in games by Pierre-Louis Cauvin c,∗, Davide Legacci ∗, Panayotis Mertikopoulos ∗ (2025) 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-2506-13286.txt.
Introduction
Pierre-Louis Cauvin, Davide Legacci and Panayotis Mertikopoulos study how random shocks change learning in games. Players follow a perturbed form of follow-the-regularized-leader, a classic rule that tracks past payoffs with a smoothing term. Noise hits both what players observe and how they update, at any strength.
The striking result is that uncertainty favours extremes: every player's play reaches a near-pure strategy in finite time, and keeps returning near pure strategies forever. The only stable limits are pure Nash equilibria, and a set of pure strategies attracts play exactly when it is closed under better replies. In games where calm play would cycle, noise instead pushes play toward the edge on average.
Catalog record
| CREATOR | Pierre-Louis Cauvin c,∗, Davide Legacci ∗, Panayotis Mertikopoulos ∗ |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 163 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-13286