No benefit from adding forecast uncertainty when the robust Stackelberg strategy is a singleton
Demonstrate that in the robust commitment multi-follower setting, if the optimal robust Stackelberg commitment reduces to a single distribution (a singleton), then releasing uncertainty estimates around predictions and having followers respond via maximin using these intervals does not improve outcomes for the principal.
References
CONJECTURE Given a robust commitment multi-follower game. If the Robust Stackelberg strategy is a singleton, then giving uncertainty estimates to predictions and maximin-responses to that the predictions is NOT making better.
— The Value of Ambiguous Commitments in Multi-Follower Games
(2409.05608 - Collina et al., 9 Sep 2024) in Conjecture, The (Dis)Advantage of Imprecise Forecasts