Continuation of Track B, frozen before new tests. Proposition target: actual fixed-(H,A) endpoint fibres have source diameter at most H/3. Positive weights inside such fibres lie in a short phase arc at binary conductor m~H, hence their amplitude/weight ratio is at least 1-O((|b|H/2^m)^2). Test exact affine identities and cylinder lifts of the two coalescent words, verify quota dependence at one common state. Compare this physical grouping to Paper115's valid state/amplitude grouping. Derive an approximate-group covariance bound, retaining the exception/correlation term. Test a finite opposite-phase fixture demonstrating why unconditional cancellation alone cannot control a rare future event. A further positive direction is certified finite Fourier norms of completion indicators; preset horizons j<=18, physical endpoint residues 0..2^j-1, heights from exact powers with H in {16,32}, deficits {1,2,3,4}, last state both. Compare total-variation bound to l1 Fourier weights and their dyadic conductor spectrum. The indicator's mean and spectrum must be checked independently by word DP. This can guide a weighted bound but numerical norms alone prove no asymptotic estimate.
