We consider two versions: sequential, where Bob observes Alice's cut

In this section we provide a detailed proof for the main theorem. First we state some facts about the learning rate and the algorithm. This bound contains three parts. The first is an upper bound for the first step when there is no data. The third part is an "average" of the estimated future regret.

This proves the first part of the theorem.

Adaptive experiments improve efficiency by adjusting treatment assignments based on past outcomes, but this adaptivity breaks the i.i.d. assumptions that underpins classical asymptotics. At the same time, many questions of interest are distributional, extending beyond average effects. Kernel treatment effects (KTE) provide a flexible framework by representing counterfactual outcome distributions in an RKHS and comparing them via kernel distances. We present the first kernel-based framework for distributional inference under adaptive data collection. Our method combines doubly robust scores with variance stabilization to ensure asymptotic normality via a Hilbert-space martingale CLT, and introduces a sample-fitted stabilized test with valid type-I error. Experiments show it is well calibrated and effective for both mean shifts and higher-moment differences, outperforming adaptive baselines limited to scalar effects.

We consider two versions: sequential, where Bob observes Alice's cut

The persistence of disparity is thus attributed to classifier policy.