Keep claims within the observed channel

These examples and illustrative results are independently authored teaching materials, not measured model results.

Use case

A browser teaching probe logs one console Error: quota without pageerror/crash evidence. Headlines cannot claim a crash, and proposed regressions cannot listen only for exceptions.

Mechanism

Create observed, expected, evidence, confirmation and separate hypothesis fields with frequency/channel. Reuse bounded facts across headline/summary/conclusion and check consistency. Regressions capture the original channel/message; probe crash hypotheses separately. Report measured timing boundaries, not budgets as durations or single observations as reproduction.

Bad example

Call one log a recurring crash, add exception-only listeners and qualify uncertainty in a footnote.

Good example

Every section says one console Error: quota observed with no observed uncaught exception/crash. Propose console capture; separately test UI/exception behavior for crash hypotheses. Retain original evidence and single-observation status rather than stronger headlines.

Why the change matters

Summaries shape interpretation, and detailed caveats cannot fix exaggerated headlines. Channel-matched regressions detect original behavior instead of unproved alternatives.

Observable expectation

A teaching console-only fixture trips console checks but not pageerror, without establishing crash. New uncaught-exception evidence may update claims with a new record. Headlines/tests retain frequency limits.

Limits

Consistency is not causation or proof of no crash, only no observation. Capture may miss channels and later evidence can broaden claims. Timing/execution/frequency fields require corresponding evidence.