Scoring Protocol v2.2 · Historical Research Branch
The text and mechanical tests once passed, but empirical testing on real samples did not pass as a whole; this route never entered the current ranking or public data.
“Final” in the filename was only the version name at the time. This proposal did not pass all real-sample tests and is not used in the current ranking.
v2.2 attempted to handle object levels, duplicate accounting, AND/OR routes, probability, and attribution shares in complex systems. Textual red-team review and nine deterministic regressions once passed, but six classes of real samples exposed insufficient evidence for success probabilities and attribution shares, as well as confusion between node and system objects.
This route therefore remains a historical research branch. It did not enter current person, Dark Star, or Nebula data and is not mixed with any current score. Any future restart would first need to correct the golden objects, complete probability and share evidence, and validate real assessment cost; this page records historical status and does not say it has achieved certification.