The Recursive Pattern — Prospective Prediction Registry (Protocol)

Prospective prediction registry protocol for The Recursive Pattern (working paper v1, DOI 10.5281/zenodo.23246408). Predictions are logged before their outcomes, with HIT/MISS defined in advance and base rates stated before each window opens; the theory's test is whether timestamped predictions beat their own pre-registered base rates. Both kill criteria are operational: Kill 1 (definitional — hindsight-only recognition is story, not structure) and Kill 2 (statistical — hit rate not above base rate kills it, tie included). The §10 proposal layer was adopted November 9, 2026 (base-rate sources; minimum N = 20 scored entries before any claim; exact Poisson-binomial test as supporting evidence; VOID category; public-scale sub-tally; role-label anonymization; the October draft RP-001 left in October, never locked, never backdated). The registry opens empty: no predictions are registered in this record. This deposit is the timestamped protocol lock, not a results publication.