S7 DIMENSIONLESS TIMING PROTOCOL — 29 October 2026

This is a retrospective integration. The original fixed-unit result, S5 free-scale scores, S6 geometric selections, preliminary second-place Cygnus result and preliminary Tucana horizon failure were known before this packaged audit was specified. The audit reproduces and broadens those checks; it is not preregistered discovery or independent replication.

Scope fixed before this packaged execution:
1. Recompute exactly one J2000 geometric role assignment for all 88 existing Hipparcos five-brightest regional sets, with the S6 Kruskal/medoid and tangent-bearing rule. Membership and apparent-magnitude ranks stay fixed. Do not filter by a four-arm MST.
2. Use only the ratio vector (7,18,4). Compute absolute wrapped right-ascension intervals relative to the equator of date. Minimize E=population_std(interval/target)/mean(interval/target). Do not convert counts into time units. Keep the S5 propagation engine and astronomical-year search [-8000,2000].
3. Search all bracketed minima and endpoints using 0.5- and 1-year grids and local scalar refinement. Verify every unrestricted score against S5 and every role against S6. Verify invariance under five positive unit/target rescalings at every sampled epoch for all 88 groups.
4. Apply the same necessary local-observation condition to all 88 groups: each star can attain a positive upper culmination at latitude 51.18 degrees N. Identify feasible epoch segments, refine boundary crossings and optimize within every segment on both grids. This is stronger than simply deleting groups whose unrestricted optimum fails visibility; feasible alternative epochs are considered. Tangent boundary scores are limiting infima. The finite grids do not constitute a theorem about every epoch or every stellar subset.
5. For Cygnus and Tucana, refine maxima of each star's upper culmination over the full search. Report simultaneous altitude feasibility at the global optimum as a separate diagnostic, without a season, twilight, extinction or terrain model. Report the original Stage 2 comparison (2620–2480 BCE) within the Cygnus branch.
6. Preserve all outcomes. Do not interpret ranks, sample counts or error ratios as probabilities of historical intention. The local-observation restriction tests direct observation at Stonehenge, not knowledge transmitted from another latitude.

ADDENDUM BEFORE THE ANGULAR RELABELLING AUDIT
After the initial S7 results, a reviewer-style suggestion proposed replacing sidereal-minute terminology with angular units and writing the scale divisor as D=4. Recompute the fixed Cygnus objective directly in degrees against (7,18,4), verify its exact factor-four relation to the archived fixed-unit score and unchanged rank, and count the unrestricted free-scale competitors. The D=4 identity is documented as a retrospective parametrization, not evidence that the builders used that angular convention. Unit relabelling and removal of a fixed scale are different operations.
