# C251 — A shifted apparent anchor does not preserve the native endpoint result

23 September 2026 · Step 46 of 60, continuing C205.

As a control, move only the Creation anchor 30 years earlier while keeping the same native Conquest point. The double-expansion output changes by 30[(25/23)²−1]=2880/529. Every formerly integral Flood output becomes fractional.

This does not rerun the apparent-Creation campaign or reject its separately anchored results. It shows only that the present native 11-fold endpoint correspondence cannot be transferred by changing one anchor and leaving the other input fixed. The new input width is then 11K+30, so it is no longer the same native 11-fold interval.

## Evidence and scope

K=529 and r=25/23. BC base B means (B+1)t/Bn BC; reflected AD base A means AD At/(A+1)n. Their civil span is B+A in both seasonal components, with no year zero. Signed paired centers are −B and +A; the center model uses components ±1/4 and restores this separation after dilation. An interval center is geometry, not an event or a proposed universal Mirror axis.

Regular Gears are restricted to Noah/Shem/Flood. Creation and Conquest retain fixed source-field members; no fourth Gear is admitted. Generated coordinates, source-field membership and stored graph incidence are distinct. File_18/File_09 and the frozen C127 graph are unchanged. Earlier N01/N03 qualifications, the missing Shem join, closed primers and unestablished original Reports 11–25 remain unchanged. This packet selectively uses prior evidence; it does not rerun completed campaigns. Assertions and alternative placements are not independent historical witness counts.

**12 step-specific checks and 19 frozen-input checks passed.** Exact inputs, results and predecessor hashes are in the accompanying JSON records; calculation code is in the evidence archive. Source dependencies are identified in the body and frozen in `sources/`.