E.13:2 - Problem

Typical failures include:

1. Measures replace objectives. Teams speak as if the score, metric, benchmark, assurance level, or all-5 posture is the value.
2. Evaluation loops become reward functions. A checking reader asks for improvement; the author adds fields, guards, source rows, proof sketches, and relation catalogues until the visible evaluation looks better.
3. Unmeasured value is damaged. Usability, safety margin, maintainability, learning, domain fit, affordability, or operator action quality gets worse while the proxy improves.
4. Proxy use is not typed. The same metric is treated as orientation cue, target, incentive, gate, and release proof without saying which use is live.
5. No value slice exists. The text claims practical payoff, but no minimally viable slice shows the value being realized in a case.