Papers · Preprint · v4

Adequacy Residuals and Blind-Spot Currency

How much does a second set of probes see that the first cannot explain?

In plain words

Two families of measurements look at the same system. One family is native to a layer. The other tries to dissolve the layer and see through it. The adequacy residual is the leftover: what the dissolving probes measure that the native probes cannot account for, at a given audit budget.

A unit disk of vectors with cost at most one, axes u1 and u2. The native probe reads u1; the target reads u1 plus u2, with largest squared response 2. The vertical diameter, where u1 is 0, is invisible to the native probe, yet the target's squared response there reaches 1: residual 2 minus 1 equals 1. Below: adding a probe that reads u2 makes the residual 0; adding a probe that reads 2 u1 leaves it at 1.
The paper's two-dimensional example. The native probe sees only the first coordinate, so on the vertical line it reads nothing, while the target still responds there. That leftover is the residual. A new probe helps only if it couples to the unexplained direction: reading u₂ removes it, and reading the first coordinate twice as loud does not.

The paper turns that leftover into a calculus. It says how residuals compose, how they shrink under post processing, and how they move across bridges between layers.

When a declared budget fails, the failure becomes a concrete object: a vector with positive excess. That is a blind spot you can point at.

What it shows

  • The residual is a Schur complement in energy scaled coordinates.
  • Chain rule, data processing and transfer laws for residuals.
  • A witness theorem that turns a failed budget into an explicit blind spot.

What it does not claim

It claims no novelty for the classical operator theory it packages, and does not claim residuals vanish in general.

Cite

Tsiokos, I. (2026). Adequacy Residuals and Blind-Spot Currency. Zenodo. https://doi.org/10.5281/zenodo.23086980