Problem
When a system evaluates an approach using evidence produced only after that approach's vocabulary and expected structure are known, apparent validation can become self-referential.
IPB
IPB preserves one or more bounded independent first-pass attempts before a developing method, project or capability definition can shape those attempts, then reconciles the preserved baseline against the later formalized approach.
Core model
original problem -> preserve independent first pass -> formalize the developing approach -> compare against preserved baseline -> reconcile, do not retrofit
Protections
baseline != truth
baseline != control group by default
independence != correctness
comparison != retroactive rewriting
Use
Useful in agentic method development, prompt-system design, architecture work and evaluation where framing lock-in or contamination could otherwise make later evidence look stronger than it is.
Scope
IPB preserves independent predefinition evidence for later comparison. It does not turn the baseline into truth, guarantee correctness, or permit the later formalized approach to rewrite the preserved first pass.