A registered result in its source context

GNET.MARTINGALE.LIMIT.01

Original ledger classification: Measured [MEASURED] martingale hypotheses under frozen gates (all met; rate one power better than the dim-law envelope, reported not forced) + typed analytic remainder; the QGEO majorant residues relocated into the normality step, not proven; no marker move (Python-only per GATE.WOLFRAM.02).

Original claim text

the noncommutative-martingale route to the G_net GNS limit -- its finite hypotheses MEASURED, all met (15/15 checks, controls 4/4, ~7 s; discovery probe gnet_martingale_probe.py, 2026-08-06, verdict MARTINGALE-HYPOTHESES-MET). THE PROGRAMME: replace continuum heuristics by martingale convergence -- if the local-state defects along the certified Haar inductive system are SUMMABLE in trace norm, the limit state exists uniquely by elementary Banach telescoping; the operator-algebra content becomes a typed list of analytic steps with finite evidence in hand. S1 FILTRATION EXACT: Haar isometry V*V = 1 with support coherence iota(win(N,0,l)) inside win(2N,0,2l) EXACT (leak 0); the Watatani quasi-basis explicit at the predeclared depths (185 elements at l = 2, 3041 at l = 3, sum v v* = 4*1); martingale compatibility omega_k o E = omega_k at every level; window data literally level-independent (H_W identical at N = 48 and N = 3072 -- one index-4 expectation for the whole tower). HONEST TYPING: the omega_k are an ALMOST-martingale (approximate, with measured defects); the convergence mechanism is norm summability + dual-space completeness (Umegaki expectations; Cuculescu cited as context, not used). S2 SUMMABLE DEFECTS (load-bearing): delta_k = ||rho_{k+1} - rho_k||_1 geometric with fitted rates 0.250 / 0.263 / 0.265 per doubling for l = 1/2/3 -- ONE POWER BETTER than the 2^{-l} dim-law envelope (reported against 0.5, not forced; bar 0.55); envelope delta_k <= 8 delta_1 2^{-k} quantifies summability (l = 2 tail sum 2.9e-2). S3 TELESCOPING / UNIQUE LIMIT: the Cauchy tail bound holds on every rung against the Richardson extrapolant; the frozen 6-observable battery gaps fall at <= 0.6; extrapolation uniqueness within 2.25 delta_5. S4 SPLIT/NORMALITY PRECURSORS uniform along the tower: sector-resolved cross-window profiles anchored with shrinking tails; determinant witness D_k = det(1 - X X*) >= 0.985 and stabilizing at 0.992756 -- the uniform estimate a normality/split argument consumes. STEPS 5-7 TYPED, NOT EXECUTED (the analytic remainder): GNS reconstruction of omega_inf, local normality (S4 witnesses as the finite input), split property + the mu4 simple-current extension (GATE.METRIC.11) on the limit net. THE QGEO ARCHITECTURAL NOTE (typed, no gate): the open majorant residues (b) 'L2 uniform FH/RH' and (c) 'L3 uniformity' of QGEO.GRIDSUP.01/v745 are NOT needed for limit EXISTENCE (summable state-norm defects suffice) -- the route relocates them into the typed normality step: an architectural reallocation, not a proof of (b)/(c). CONTROLS: the shifted-bond isometry leaks out of the doubled window (0.707 > 0.3 -- the filtration breaks); the Z2-only average has lambda = 1/2 != 1/4; the scrambled sea gives O(1) defects (summability breaks); DECIMATION CONVERGES TO THE WRONG DEGENERATE LIMIT (||rho_dec - 1/16||_1 < 1e-3 vs Haar 0.1+) -- a summable filtration with the wrong bond data forgets the sea: the coherent Haar bond is load-bearing. Every measurement is finite-level evidence FOR the cited continuum steps, never the theorems. No marker move, NO RH claim. Python-only per GATE.WOLFRAM.02. CONTRACT (registered OPEN [O], the analytic remainder = steps 5-7): GNS reconstruction of omega_inf on the quasi-local algebra (existence/uniqueness finite-level delivered here); local normality of the GNS representation on each A(W) (nuclearity-type estimate; the uniform S4 witnesses are the finite input); split property and the mu4 simple-current extension (GATE.METRIC.11, [B:A] = 4 = |mu4|) on the limit net; kill: non-summable defects at a deeper reachable rung or a determinant-witness collapse.

Original classification detail

MARTINGALE-HYPOTHESES-MET: exact Haar filtration with one index-4 expectation for the whole tower; state defects summable at measured rate ~0.25 per doubling (<= the 2^-l envelope, one power better); unique limit by telescoping verified on every rung; split/normality precursors uniform (determinant witness >= 0.99); decimation converges to the WRONG degenerate limit (coherent bond load-bearing); QGEO residues (b)/(c) bypassed for existence, relocated into normality; steps 5-7 typed open; no RH statement.

Dependencies as recorded

GNET.GNSLIMIT.01; GATE.METRIC.11; QGEO.GRIDSUP.01; v779; v746