M x N categorical outcome matrix with integer entries in
{0,...,C}.
Selection count; defined for any integer k >= 1. k = 1 matches
bayesCi.
Optionalw: readonly number[] | null
Optional reward vector of shape (C+1,). If omitted, R must be
binary and [0, 1] is used.
OptionalR0: Matrix | null
Optional M x D matrix of prior outcomes.
Credibility level for the normal-approximation interval.
Optionalbounds: Bounds | null
Optional [lo, hi] clipping bounds. If omitted, the interval
is clipped to the minimum and maximum reward levels in w.
[mu, sigma, lo, hi].
Bayesian posterior summary for maxAtK, returning
[mu, sigma, lo, hi].The posterior uses the same Dirichlet-plus-one construction as
bayes. Whenk = 1, Max@1 reduces to the single-draw expected score, so this function agrees withbayesCi. This uncertainty model is ascorioextension and is not part of Bagirov et al. (2025).