ECR-000002 Expected Outcomes

Expected Stronger Signals

• Canonical packets should produce the highest recognition rates.
• Paraphrased packets should reduce easy recognition while preserving some family-level structure.
• Structural packets should provide the strongest test of extraction without canonical cues.
• Reasoning and execution families should remain more procedurally rich than the static family.

Expected Weaker Signals

• Primitive wording is likely to drift more than structural or constraint layers.
• Product relevance may appear, but it should remain secondary and observational.

Non-Claims

• This run does not resolve recognition bias by itself.
• This run does not validate Framework Engineering.
• This run does not justify automatic confidence updates.