A Sealed Log-Gram Test of Chiral Spin Amplification Across Molecular Length and Facial Defects

Can chiral spin log-odds grow by one fixed increment per independently certified molecular completion? This preregistration seals a two-tier molecular-length and facial-defect test of chiral spin amplification. The measured object is a gain-free Log-Gram coordinate built from four positive raw currents. The frozen decision fork is \[ \boxed{ \begin{aligned} H_S:&\quad \Xi_{\rm odd}=\beta N_{\rm eff}(1-2f),\ H_K:&\quad \Xi_{\rm odd}=\frac{N_{\rm eff}(1-2f)}{4\pi\cos(\pi/8)}, \end{aligned} } \] where \(N_{\rm eff}\) is an independently certified effective serial count and \(f\) is an independently measured signed facial-defect fraction. Neither quantity may be inferred from the spin data or from a displayed molecular drawing. The observable \[ \boxed{ \Xi_{\rm odd} =\frac14\log\frac{I_{+,R}I_{-,S}}{I_{-,R}I_{+,S}}. } \] Independent positive gain changes on the two enantiomer rows cancel exactly. The analysis starts from raw currents, propagates their full covariance, and enforces a frozen current floor rather than inserting post-data pseudocounts. Two distinct scientific questions Tier S tests the theorem-supported structural claim that independently certified serial packets add in Log-Gram rapidity. One common material slope is learned only from blinded training strata; held-out lengths and defect cells are predicted without refitting. Tier K is harder: it freezes the slope \[ \boxed{ \xi_K=\frac{1}{4\pi\cos(\pi/8)} =0.0861340345203237\ldots } \] with zero fitted slope, intercept, saturation, length offset, or normalization. Tier K is explicitly a sealed QTT conjecture because the source-word-to-material-transfer bridge has not yet been proved. What counts as a valid verdict The campaign requires at least four structurally characterized length fractions spanning a factor of two, both enantiomers at every length, racemic/achiral controls, independent preparations, repeated field cycles, blind labels, and ordinary transport controls. Rail rotation, unresolved packing/contact confounding, current-floor failure, or post-unblinding calibration gives INELIGIBLE-NO-THEORY-VERDICT, not a QTT falsification. An eligible Tier-K pass is registered as QTT-ZERO-TARGET-FIT-EMPIRICAL-VICTORY:CISS-LOGGRAM-GAP. If Tier S passes but Tier K fails, only the fixed coefficient is rejected; the serial law survives. If eligible held-out data reject serial additivity with a stable rail, the registered material application fails. A reproducible third shape is preserved as an anomaly packet rather than erased as noise or used for a post-hoc rescue. Public anchors Sealed-test concept DOI Parent CISS source-graph and Log-Gram theorem Bifacial-ladder discovery paper Quantum Traction Theory: Main Book QTT Observatory