33.76 turns structure formation into a retainable endgame scorecard: at matched redshift and resolution, S_matter must nest inside S_field with high N_in, field-skeleton coverage C must rise monotonically with maturity M and toward lower redshift while leaving reproducible unfilled segments at earlier stages, low-contrast regions must still show nonzero orientation alignment A_align before or beyond count enhancement, and skeleton / redshift / mask / sparsification nulls must fail cleanly while K_multi across field probes exceeds cross-tracer matter agreement; under V08/V09-compatible retain, this remains one road-network-first structure-formation ledger rather than a free-standing field-skeleton ontology.
Use this section as a compact machine-readable EFT reference.
33.76 turns structure formation into the volume’s terminal retain court. The admissible claim is that a field skeleton S_field should organize first, so the matter skeleton S_matter should embed within it, field-skeleton coverage should fill in later as maturity rises, and low-contrast regions should already carry an orientation prior before matter counts fully catch up.
The scorecard is built from two independently extracted skeletons in matched redshift slices and at matched angular resolution. S_field comes from ridge extraction on δκ or an equivalent residual tidal-shear field, while S_matter comes from at least two matter tracers such as galaxy-density, gas, SZ, radio, or intensity-mapping ridges. The hard verdict metrics are the embedding rate N_in, the coverage rate C, an independent maturity label M, a low-contrast orientation-alignment statistic A_align, and a cross-probe field-consistency metric K_multi that should exceed cross-tracer matter-skeleton agreement.
The workflow is designed to prevent one skeleton from leaking into the other. All probes are projected to the same resolution and tomographic bins, skeletonization thresholds and smoothing scales are frozen before measurement, S_field and S_matter are generated by separate teams or pipelines without cross-feeding, N_in, C, and A_align are computed in matched windows with a frozen low-contrast aperture, maturity stratification is defined independently of skeleton extraction, a held-out sky region or redshift layer is reserved for adjudication, and at least two weak-lensing and two matter-tracer pipelines must agree on the verdict direction.
The nulls are explicitly destructive. Skeleton permutation or rotation must return N_in and A_align toward chance, redshift-label permutation must weaken embedding and coverage trends, mask-boundary perturbations test whether the main structure is robust rather than an obscuration artifact, low-maturity controls should preserve embedding while lowering coverage, and forced matter sparsification should fragment S_matter more readily than S_field if the “roads first” asymmetry is real.
Support requires three linked lines to survive together: N_in must remain well above permutation baselines, C must rise monotonically with maturity and toward lower redshift while leaving earlier unfilled segments, and A_align must stay significantly nonzero in low-contrast regions after control for local count strength. Falsification follows if nesting is near random, if coverage trends disagree in sign across probes or pipelines, if the orientation prior appears only in high-contrast density peaks, if nulls remain comparably strong, or if strict removal of matter information makes the main field skeleton vanish. The main systematics are weak-lensing reconstruction artifacts, matter-tracer bias and incompleteness, and algorithm dependence in skeleton extraction.
So the volume closes on one retained road-network-first structure-formation scorecard aligned with the V08/V09 maturity, null, and readout discipline. S_field may survive only as the audited organizing skeleton of this endgame ledger, not as a free-standing ontology owner or a one-chapter cosmology verdict. The endgame remains one clean terminal route into structure-formation adjudication.