Inflation's flagship signature, primordial B-modes from tensor fluctuations, has retreated for forty years: BICEP2's detection dissolved into dust, the bound sits at r < 0.036 and tightening, the classic models predicting 0.01 to 0.1 are executed, and the survivors are engineered to predict what has not been seen. The discriminator that separated "inflation happened" from "inflation is a parameterization" keeps migrating below each null.
The model needs inflation for horizon, flatness, and perturbations, and inflation at the energies that do those jobs naturally radiates tensors. The persistent silence forces the paradigm downward into plateau designs whose chief virtue is invisibility, a mechanism increasingly defined by its ability to evade its own test.
SCT generates the primordial perturbations without inflating anything, and its tensor budget follows from the same arithmetic as its other primordial numbers: the cascade's many stages produce gravitational waves incoherently, suppressing the integrated background by the event count, with the registered prediction r below 10⁻⁵ for every physically plausible parameter combination (Paper 10, From Chaos To Confirming Falsifiability; the cascade statistics in Paper 3). The forty-year null is therefore not an embarrassment to be out-designed but the framework's standing forecast: every tightened bound, from BICEP2's retraction through r < 0.036 toward the 0.001 frontier, is the cascade's silence being measured at increasing precision. All observed B-modes are lensing-processed E-modes, permanently.
The same event count that silences the tensors sets the scalar tilt (n_s = 1 - 1/L) and the non-Gaussianity floor (1/√N_coll): three primordial observables from one counting, none adjustable, which is the structural difference from a paradigm whose r migrates with each null. The perturbations' origin in collision geometry (P56, P36-P38) does the work inflation was hired for, with the horizon and flatness problems dissolved by the same single replacement (P24).
The asymmetry of risk is total and welcome: SCT cannot survive a detection, and inflation cannot be killed by a null. One of these is a falsifiable position.
The registered kill is the cleanest in the catalog: a confirmed primordial B-mode detection at r above 0.01 at 3σ or better (CMB-S4, LiteBIRD, Simons Observatory) falsifies the cascade origin outright, no fallback, no migrated model space. SCT's tensor floor sits four orders below the coming sensitivity, so for the foreseeable future every measurement is a pass-fail test the framework has pre-committed to.