The large-angle CMB carries the signatures topology would produce, suppressed low multipoles, aligned axes, hemispheric asymmetry, which is why the COMPACT collaboration (PRL 2024) argues the anomalies may encode topological information. Yet matched-circle searches and Planck's likelihood analyses are null: no identified circles, no detected wrapping. Every indirect statistic whispers finite; every direct test shouts unbounded.
ΛCDM assumes infinite, simply connected space as an unexplained default, and inflation makes any identification scale generically unobservable. The model therefore cannot explain topology-like anomalies except as flukes, and could not accommodate real topology without breaking its own initial-conditions story. It has nowhere to put a sky that looks bounded but is not.
SCT replaces the hot-dense-center with a superluminal collision, and that one change splits the question the data could not: the manifold and the patch are different objects. Space itself is infinite and simply connected (P2), so there are no identified points, no wrapped light paths, and no matched circles, ever; the null direct searches are a permanent prediction. But the thermalized region whose products we inhabit is finite, the collision overlap volume (P26), and its perturbation deposit has a largest characteristic scale, Λ_max of about 2 R_pocket, near 5 Gpc (P55), with an orientation set by the collision geometry (P48).
A finite, oriented deposition domain inside an infinite space produces precisely the half-and-half sky observed: mode suppression at scales above the deposit boundary, preferred axes from the collision geometry, and asymmetries from the deposition pattern, all the statistical signatures of finiteness, with none of the circle-pair signatures of identification. The hints are real physics and the nulls are real physics; they describe the patch and the manifold respectively. ΛCDM conflates the two because a universe-wide Big Bang leaves no room for the distinction.
This is the same finite-domain geometry that produces the low-multipole deficit and the aligned anomaly family, and the same embedding that predicts parent-pocket structure beyond the horizon rather than empty space. There is no need to invoke compact topology, and no need to explain away the anomalies as variance.
Symmetric kills. A confirmed detection of cosmic topology, matched circles or a COMPACT-style identification signature in polarization or 21-cm data, would falsify SCT's infinite simply connected manifold (P2) outright. In the other direction, if the large-scale anomaly statistics dissolve under LiteBIRD polarization into full consistency with an isotropic Gaussian sky, the finite-deposit signature is removed; and wide-field surveys finding no boundary-scale structure in the matter distribution at k below 0.01 Mpc⁻¹ (DESI, Euclid) would undercut the 5 Gpc deposit scale the mechanism requires.