1. What this unit proves 2
2. What it uses 3
3. Strategy 4
4. Steps 4
5. Where the difficulty lies 4
6. Notation 5
7. What remains unexplained 5
Code dependencies 5
Lean source 6
Additional source locations 7
Key statements 11
Part III — Ultraviolet construction (renormalization group).
Organizes observable and activity bounds, including additional comparison estimates carried by the selected run.
Status: edited. Condensed from the unit's detailed notes, which were checked against the linked Lean in two rounds and edited; the local limitations still apply. Status meanings.
Overview · All units · Detailed notes · ← Previous: III.14 · Next: III.16 →
Results · Inputs · Strategy · Proof steps · Notation · What remains unexplained · Code dependencies · Lean source · Key statements
Reading scope. The pair-step and observable-derivative bounds feed III.2, III.9 and III.10. Additional comparison bounds reach the selected run in III.14 as a separate conjunct, which the final construction-run theorem discards; this does not establish that the associated code can be removed.
Part: III, Ultraviolet construction (renormalization
group). Lean: unit III.15: YM/RemainderComparison/,
YM/Progress/DIRECTED_RECIPE/, YM/Next/SCHEDULE/, with readout and integration
fragments. Size: 7,382 lines, 1,053 declarations;
about 4,200 lines concern launch data.
Source inventory and conventions
Proposition III.15.1 (pair-step rates). Fix \(\rho > 0\), \(C \geqslant 0\), \(r > \rho\), \(d \in \mathbb{N}\), \(e \in \mathbb{R}\), \(q > 0\). Put
For all sufficiently large \(y > 1\), uniformly for \(\beta \in [y / 4, 4 y]\) and integers \(0 \leqslant B \leqslant \mathcal{B} (y)\),
The threshold depends only on the fixed parameters. Status: theorem. Lean: YM.Progress.DirectedRecipe.Payments.Pairing.currency.
Proposition III.15.2 (hard–soft merging). Fix a parameter point with \(K \geqslant 6\), direction power \(P_0 \geqslant 2\), \(B_P \geqslant 1\), \(\sigma > 0\), \(0 < t_P \leqslant 1\). Assume countably indexed regular terms \(E_i\), forced activities \(F_i\) on \(X_i\), and real corrections \(\ell_i\) on nonempty \(X'_i\). Require support regularity, flat-field vanishing and vanishing affine Hessian on nonwrapping supports for \(E_i\); gauge invariance, measurability and dependence on supported links and final flags for \(F_i\), their eligibility tests and \(\ell_i\); eligibility whenever \(F_i \ne 0\); translation/reflection covariance for all families, with prescribed permutations, support and inventory transformations for regular and forced families; and \(\ell_i = 0\) on its small-field region. Suppose the regular family's nonnegative derivative bounds through \(K\) on small-field supports, in prescribed directions, have weighted anchored sum \(\leqslant \delta\); \(|F_i (U) | \leqslant p_i\), \(| \ell_i (U) | \leqslant q_i\) for all \(i, U\), with \(p_i, q_i \geqslant 0\); and
Assume III.15.1's inequalities with \(B = B_P\), replacing its last bound by \(\text{quote} (\kappa, \gamma) \leqslant w\). Then the fusion input condition holds with \((\delta, w)\): original bounds \(\delta, a\) persist, the soft family has a summable uniform majorant, and \((F_i) \sqcup (e^{\ell_i} - 1)\) has weights \((e^{\kappa} p_i) \sqcup (2 q_i)\) and hit bound \(\gamma\), with \(|F_i | \leqslant e^{- \kappa} (e^{\kappa} p_i)\) and \(|e^{\ell_i} - 1| \leqslant 2 q_i\). III.10's fusion gives norm \(\leqslant w\). Status: theorem. Lean: YM.Progress.DirectedRecipe.Payments.tender.
For III.15.3–4 fix \(K \geqslant 6\), \(P \geqslant 2\), \(0 < t \leqslant 1\). Directions satisfy \(\|A_e \| \leqslant t (1 + \text{dist} (e, o))^P\).
Proposition III.15.3 (slots). For every \(\ell \in \mathbb{N}\), slot \(s\) with \(|s| \leqslant \ell\), finite \(S \subset \mathbb{Z}^4\) containing its stencil, and site \(o\), the slot is \(K\) times left-differentiable. If \(U\) is small-field on wrapped \(o + S\), every derivative of order \(m \leqslant K\) in admissible directions has norm \(\leqslant C_s (P, K, \ell) t\), where
At order zero the bound is \(t\); constants are independent of volume and \(t\). Status: theorem. Lean: YM.Next.Scheduling.Wages.slotPays.
Proposition III.15.4 (atoms). For every \(k \in \mathbb{N}\), \(k\)-leg atom \(a\), site \(o\), and \(U\) small-field on its footprint, \(a (\cdot, o)\) is \(K\) times left-differentiable, with every admissible derivative through \(K\) bounded by
Status: theorem. Lean: YM.Next.Scheduling.Wages.fin.
Proposition III.15.5 (launch data). Let \(G\) be simple. Fix a sheet of profiles \(P_j\), kernel, compatible price table, targets \(R_{ji}\) with \(\rho_{R_{ji}} < - 3 / 8\) for all \(j, i\), tolerances \(h_{ji} > 0\), libraries \(C_{ji}\) of order \(P_i\), and marks \(r_{ji}\). There exist \(f : \mathbb{N} \to \mathbb{N}\) and \(C'_{ji} \geqslant 1\) giving quotation bounds with fine/coarse exponents \((\sigma_j, \sigma_i + 2)\) and factor \(C'_{ji} L^{- 5}\). These require support enlargement by \(r_{ji} \mathcal{B} (x)\), coarse power equal to that of \(P_i\), fine exponent \(\sigma_j\), \(B_P =\mathcal{B} (x)\), \(B_B \leqslant 2 B_P\), \(t_P \leqslant 2 t_B\), and acceptance. For every \(u \geqslant 0\) satisfying local analytic smallness and every \(\tau \geqslant 0\), the bound holds. The quotation inequality is \(e^{(\sigma_i + 2) N_{B_P} (Y)} \tau \hspace{0.17em} \text{returnFee} \leqslant C'_{ji} L^{- 5} e^{\sigma_j N_{B_B} (X)} \tau\). For each stage pair \((j, i)\) and every block side \(L \geqslant f_j\), III.15.6 holds and
Status: theorem. Lean: YM.RemainderComparison.Chron.sails.
Proposition III.15.6 (large-side comparison). For simple \(G\), kernel \(K\) and target \(Q\) with \(\rho_Q < - 3 / 8\), there exists \(n\) such that for every \(L \geqslant n\) and branching patch, every \(W, \mu > 0\), order datum at \(W\), songbook and scrollwork at side \(L\), \(\chi\), stages \(E, E'\), rays satisfying the heart condition for \((E', \text{e}_Q)\), library of order \(E'\), record with parameter \(\leqslant 1 / 4\), \(j \in \mathbb{N}\), reserve \(f\), and \(\eta, \delta > 0\), there exist a financed quarry \(D\) at rate \(\rho_Q\), with \(d (D) > 0\) and paint supply, and \(P_1\). For all \(x \geqslant P_1\), poster comparison, chart conditions and instructions hold; active, resonance and tolerance conditions hold for every \(\beta \in [\beta_F / 4, 4 \beta_F]\) and every point tempered to \(E'\). The threshold \(n\) is independent of auxiliary data; \(D, P_1\) may depend on them. Status: theorem. Lean: YM.RemainderComparison.Cadenza.buoys.
Proposition III.15.7 (delay enlargement). Fix a sheet, compatible target/expense tables \(R, X\), sides \(L_j\), protection, register, firmament, tolerances, libraries and marks as in III.15.5, and records \(u_{ji}\). The register supplies positive \(W_{ji}, \mu_{ji}\) and orders on stage pairs; each finite request family supplies scrollwork, songbook, grade/orders, reserve, positive constants, and an extra condition eventually true for every financed quarry. Assume launch thresholds \(k_j \leqslant L_j\), \(\varepsilon (u_{ji}) \leqslant 1 / 4\) on stage pairs with \(i\) at least the first stage, and \(\sigma_i + 3 < \text{e}_{R_{ji}}\) for all \(j, i\). For every initial delay \(T_0\), there exist \(W \geqslant T_0\) and quarries \(V_{jig}\) for every request slot at every \((j, i)\), chosen before any \(U \geqslant W\). For every such \(U\), launch data persist; on branching stage pairs with \(j \geqslant 4\) and \(i\) at least the first stage, every request has the heart property, positive \(d\), paint supply, III.15.6's allotment at threshold \(\beta_j^{\ast} (U)\), and its extra condition above that threshold. On all stage pairs with \(i\) at least the first stage, requests are tiled whenever \(x \geqslant \beta_j^{\ast} (U)\) and
Tiling means \(2 z_g (x) < c\) and \(4 \text{halo}_g (x) + 16 \leqslant c\). Delay ordering increases entry couplings, scale-entry functions, shell counts and growth exponents componentwise. Status: theorem. Lean: YM.RemainderComparison.Chron.grows.
III.1: scales, slots, atoms; II.1: left-derivative calculus; II.2: hit bounds; III.5: logarithmic cluster tails.
III.10: fusion, quotation, connected-activity and tree estimates; III.2: selected targets and delay ordering.
Leibniz and chain rules, Mayer expansion, Kotecký–Preiss bounds, elementary asymptotics, \(|e^z - 1| \leqslant 2 |z|\) for \(|z| \leqslant 1\).
Retain \(e^{- \rho s^2}\) from the hard bound and use the remaining Gaussian decay to absorb support counting; soft bounds decay by their negative coupling power. Express slots as single group words: unitarity controls derivatives, and Leibniz gives \(t^k\) for atoms.
Choose the block-side threshold before auxiliary data, then comparison functions and entry couplings. Negative activity rate survives the tree weight \(\beta_F^{3 / 8}\); finitely many requests per stage permit common delay enlargement.
Fusion. Rescaled hard weights and doubled soft weights add to \(\gamma\); the exponential inequality bounds the Mayer terms. YM.Progress.DirectedRecipe.Payments.tender.
Observable bounds. Differentiate the detour word, use unitarity, then multiply slot estimates. YM.Next.Scheduling.Wages.slotPays; YM.Next.Scheduling.Wages.fin.
Comparison functions. Choose reserve functions conditional on local derivative estimates, charts and coefficient matching. YM.RemainderComparison.Herbal.compile (proof read; analytic inputs not expanded).
Marked outputs. For the conditional marked-output estimate, separate single remaining occurrences from clusters with further occurrences and from large supports. The \(L^{- 5}\) quotation summed over \(L^4\) sites supplies the \(C_0 / L\) term; the other contributions have negative rate. Occurrences are counted outside inserted leaves, with at least one marked. YM.RemainderComparison.Solo.atlas (proof read; singleton input not expanded).
Launch and enlargement. Combine comparison and tail bounds with quotation thresholds, then take finite stagewise maxima. YM.RemainderComparison.Chron.sails; YM.RemainderComparison.Chron.grows.
The bounds must be uniform in coupling bands and volume; the block-side threshold precedes all auxiliary choices, while comparison functions precede the entry coupling. Part C's conditional comparison conclusions reach the selected run but are discarded by the final construction-run theorem.
\(g (\beta) = \beta^{- 1 / 2}\), \(p (x) = e^{(\log \log x)^2}\), \(t (x) = g (x) p (x)\), \(\mathcal{B} (x) = \lceil (\log x)^2 \rceil\), \(\beta_F = x + b_G \log L\); \(x\) is coarse inverse coupling. \(\text{quote} (\kappa, \gamma) = e^{- \kappa} (e^{\gamma} - 1)\). Hit means \(\sum_{i : z \in X_i} f_i e^{aN_M (X_i)} \leqslant \gamma\) at every site \(z\), for weights \(f_i \geqslant 0\); \(N_M\) counts a minimal connected covering by \(M\)-balls. Stage pairs have \(i \in \{j, j - 1\}\); \(\sigma_j\) is the profile's support exponent. A quarry is a record of bounding functions; financing specifies reserve/rate bounds. Rate \(\leqslant a\) means \(|w| \leqslant \beta_F^v\) eventually for every \(v > a\); a reserve satisfies \(|w| \leqslant p (\beta_F)^k\) eventually for some \(k \in \mathbb{N}\).
A slot is curvature transported along a walk \(\omega\), with length \(| \omega | + 4\); atoms are coefficient-weighted products of slots, with coefficient norm \(|a| = \sum_b |c_b |\) and total length \(\text{len} (a)\). Small-field means every touching plaquette has chord distance \(\leqslant t\) from identity; \(\kappa_G\) is the colour scale and \(\|1\|\) the Frobenius norm of identity.
For a target \(Q\), \(\text{rel}_Q = \text{e}_Q + 2 \varepsilon_Q + \delta_Q\); \(h_Q\) is its halo and \(c_{h_Q + 5}\) its pre-cost. Targets carry their chart/local-ratio conditions, \(\rho_Q \in (- 1 / 2, 0)\), \(\text{e}_Q \geqslant 0\), \(\varepsilon_Q, \delta_Q > 0\), \(h_Q \geqslant 500\), \(4 \varepsilon_Q\) below the top support exponent and \(8 \varepsilon_Q\) below both chart widths. YM.ConnectedActivityE6.Target.
The heart condition is
with starred ray parameters. Tempered points have radius \(\mathcal{B} (x)\), threshold \(t (x) \leqslant 2 t (\beta_F)\), and the stage's order, power and support exponent.
Local analytic, compatibility and readiness conditions remain to be written out. Comparison-function existence does not supply coefficient matching. The singleton, weighted-cluster, full-bill and analytic-expansion inputs are not all checked here; readout, RG12 and R2I fragments also remain outside the detailed verification.
Detailed notes: the full account of this unit, with proof sketches for every step, is on the detailed notes page.
Code dependencies (from the compiled Lean). These lists follow references between declarations allocated to units; they can include shared definitions, record fields and alternative constructions. They are not a mathematical reading order. Only unit pairs with at least 25 references are shown; see the counting rules.
References to: III.10, III.1, I.1, II.2, III.5, III.6, III.11, III.2, II.1, III.16, III.9, VIII.1, III.4.
Referenced by: III.2, III.12, III.13, III.10, III.9, III.11, III.14, VIII.1, III.4, III.3.
Source allocation: 7,382 attributed lines; 1,053 line-counted declarations; 1,745 assigned declaration names. Of the assigned names, 1,568 are in the compiled closure of the seventeen headline results. These quantities have different scopes.
The folders and result chains below are source inventories. Their line counts cover each whole group, including any share allocated to another unit; percentages use attributed lines plus one per assigned name. They need not sum to the unit total.
proof of sails and the results it chains to — 4,201 lines.
work package YM/Progress/DIRECTED_RECIPE/ — 1,251 lines.
work package YM/Next/SCHEDULE/ — 933 lines.
work package YM/RemainderComparison/Cadenza/ — 186 lines.
work package YM/RG05Readout/Test/ — 62 lines.
work package YM/RemainderComparison/Auxiliary/ — 61 lines.
work package YM/RemainderComparison/Funk/ — 61 lines.
work package YM/RG05Readout/Offsets/ — 49 lines.
work package YM/RemainderComparison/Access/ — 47 lines.
work package YM/RemainderComparison/Herbal/ — 42 lines.
work package YM/RemainderComparison/Gaussian/ — 41 lines.
work package YM/RemainderComparison/Chron/ — 40 lines.
work package YM/RemainderComparison/Muse/ — 38 lines.
work package YM/RemainderComparison/Portal/ — 38 lines.
work package YM/RG05Readout/Grants/ — 36 lines.
work package YM/RemainderComparison/Lyric/ — 36 lines.
work package YM/R2I/Campaign/ — 34 lines.
work package YM/RG05Readout/Address/ — 26 lines.
work package YM/Integration/RG05Preconstruction/ — 25 lines.
work package YM/RG12/Shadows/ — 25 lines.
work package YM/RG12/Spins/ — 18 lines.
work package YM/R2I/Buckram/ — 17 lines.
work package YM/RemainderComparison/Arc/ — 16 lines.
work package YM/RemainderComparison/Cascade/ — 14 lines.
work package YM/RemainderComparison/Sleet/ — 13 lines.
work package YM/Integration/ChronPreconstruction/ — 12 lines.
work package YM/RemainderComparison/Halo/ — 10 lines.
work package YM/OPE/Therm/ — 9 lines.
work package YM/RemainderComparison/Flues/ — 9 lines.
work package YM/RemainderComparison/Exchange/ — 8 lines.
work package YM/RemainderComparison/Solo/ — 5 lines.
work package YM/RG12/Burst/ — 4 lines.
work package YM/RemainderComparison/Passion/ — 4 lines.
work package YM/RemainderComparison/Gear/ — 4 lines.
work package YM/RemainderComparison/Ethereal/ — 3 lines.
work package YM/RemainderComparison/WickReps/ — 2 lines.
work package YM/RemainderComparison/Apo/ — 1 lines.
work package YM/RemainderComparison/Breeze/ — 1 lines.
Further pinned locations for the definitions and comparisons discussed in this unit.
In the selected statement order. K- identifiers name this generated list; proposition and lemma numbers in the outline are separate. Each sentence is an automatically drafted summary of a Lean statement; follow its link for the exact hypotheses and conclusion. The outline status does not certify these summaries. Main result and key result describe importance, not correctness. Full names are in the statement table.
K-III.15.1 currency (numbered result) — Eventually in y, uniformly over beta in band, pair-step rates hold with the exponentially small strong and mild costs.
K-III.15.2 tender (numbered result) — Separate hard and soft transfer costs with a rate relation suffice to obtain the combined Ingredients bound.
K-III.15.3 slotPays (key result) — A local slot observable satisfies a derivative bound with explicit slot cost on a healthy region of a finite site set.
K-III.15.4 sails (key result) — For a simple gauge group, if all rates satisfy 3/8+rate<0, there exist data meeting the pre-stage launch (pledge) conditions for the all-order asymptotic proof.
K-III.15.5 buoys (key result) — For a simple group and rate with 3/8+rate<0 the quay condition holds for all sufficiently large side lengths.
K-III.15.6 rollBefore (key result) — Beyond a size threshold, any financed poster extends to a financed quarry satisfying the instruction conditions eventually in x.
K-III.15.7 quorum (key result) — For positive μ and a suitable heart condition there is a positive-parameter financed budget that is paid for all large x.
K-III.15.8 compile (key result) — Given admissible weights and index bounds, there exist four admissible functions such that the airstrip comparison estimate holds eventually in t.
K-III.15.9 run (numbered result) — A weighted sum over revealed semantic supports satisfies a quantitative retention bound, giving a uniform hit estimate under a veto condition.
K-III.15.10 atlas (key result) — For large L and every recorded setup there are rates and constants so that a captain structure exists and scale-data inquiries are paid.
K-III.15.11 seeds (key result) — For large side L there exist rates and constants such that eventually in x every setup admits a captain structure with the heavy, heroes and quarry properties.
K-III.15.12 generation (key result) — Under 3/8+rate<0, for large L every certified test is paid, with pad structure and role bounds with negative-rate errors.
K-III.15.13 ramp (key result) — With inherited recurrence budgets, a captain structure exists and, for any scale data, yields horizon-type bounds with a negative exponent.