Answers are grounded in this paper and cite their sources.
The Fréchet mean can exhibit unexpected behavior for a wide class of random variables. For instance, it can stick to a point under resampling (sample stickiness) or for topologically nearby distributions (topological stickiness, such as total variation or Wasserstein stickiness). It can stick to a point for slight but arbitrary perturbations (perturbation stickiness). We explore various flavors of stickiness, also, sticking to a closed set, and their relationships in varying scenarios, for instance on CAT( κ κ ) spaces, κ ∈ R κ ∈ R , leading to a theorem on equivalence of flavors. Modulation stickiness and directional stickiness (generalizing moment stickiness) are promising candidates for new statistical methods building on asymptotic fluctuation, where, due to stickiness, the mean itself features no asymptotic fluctuation. We provide asymptotic rates for the former and a central limit theorem for the latter. We also rule out sticky flavors on sufficiently regular spaces, e.g., on manifolds with curvature bounds.
No takes yet. Share an insight, caveat, or question.
Lammers et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: