Assumption: Complete validation contract

For each setting $(\nu ,k,P)$:

1. conditional on a declared calibration record $\mathcal{Z}$, the positive sample count $N_{\nu kP}$ is fixed before validation;

2. the scores $Z_{\nu kP,i}\in [-1,1]$ are conditionally independent with a common mean $m_{\nu kP}$;

3. all settings for $(\nu ,k)$ test the same complete channel;

4. on an event $G$, with $\mathrm{Pr}(G^c)\le \beta$,

   $$|m_{\nu kP}-t_{\nu kP}|\le b_{\nu kP};$$

5. on that event, the deployment channel obeys

   $$d_{\diamond }({{\Phi }^{\star }}_{\nu ,k},{\Phi }_{\nu ,k})\le {b^{\mathrm{tr}}}_{\nu ,k}.$$

Preparation, decoding, losses, failure branches, and score assignment are specified completely. No success-postselected channel is substituted for the deterministic channel being claimed.