In the standard model, scalar perturbations generate only E-mode polarization, and B-modes arise mainly from gravitational lensing of E-modes. EE and lensing-induced BB patterns are therefore predicted to be tightly correlated. Planck and ground-based experiments report hints of partial decorrelation and subtle frequency and spatial mismatches between EE and BB that are hard to reconcile with a single coherent lensing-and-foreground model (Planck XI 2020; BICEP/Keck 2018).
The standard model assumes statistically isotropic scalar perturbations with parity symmetry. Under those assumptions, EE and lensing-induced BB share the same underlying matter distribution and must be highly correlated. Persistent decorrelation has nowhere to go in the model except as foreground-physics complexity or systematic error.
SCT replaces the hot-dense-center with a superluminal collision and the thermalized debris field. From this single change, the cascade is intrinsically directional and parity-asymmetric. The original collision deposited an angular momentum vector J = μ(b × v_rel) into our patch (P22, P31, P32). E-modes (even parity) and B-modes (odd parity) respond differently to the J-axis-aligned cascade structure, reducing the EE/BB correlation below the ΛCDM standard expectation.
The cascade-stream filament network (P34, P36) propagates the J vector to all CMB multipoles, generating axis-aligned polarization response that introduces a parity-asymmetric component into both EE and BB power spectra. The two are still correlated by the underlying matter distribution and lensing, but with a residual cascade-axis-aligned decorrelation component on top. The Plasma Equivalence Theorem (P29, P30) preserves the geometric imprint through to recombination so that the decorrelation appears in observable polarization data.
The decorrelation axis is predicted to align with the Axis of Evil (recid 24), the hemispherical-asymmetry axis (recid 28), the bipolar power axis (recid 29), the parity-odd TB/EB axis (recid 35), the odd-parity preference axis (recid 38), and the low-l power deficit axis (recid 32). Same parameter set, six observable projections from the same cascade-deposited J vector. There is no need for non-standard polarized-dust foregrounds or modifications to gravitational lensing.
If precision LiteBIRD or CMB-S4 polarization finds the EE/BB decorrelation axis statistically incompatible with the other six low-l anomaly axes at greater than 3σ, the M10 common-collision-axis explanation fails. Equivalently, if frequency-cleaned EE and BB are recovered to be perfectly correlated to the noise floor, the cascade-induced decorrelation is refuted.