Documentation

MIPRE.Background.Repetition.CommutingRepetition.Prerounding.CostsLemmas

noncomputable def CommutingRepetition.scenMartA {n : } {X Y A : Type} [Fintype X] [Fintype Y] [Fintype A] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [Nonempty X] [Nonempty Y] [Nonempty A] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : X, μ v (yw c) 0) :
EffectMartingale (Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)) L.card

The Alice-side scenario reveal martingale: the path space is the Alice values on the progressively revealed block L, the law is the product of the pinned-Bob conditional weights (normalized by the total block mass), and the index path is the canonical Alice label at the growing revealed set SA ∪ π[1..j], with the off-block values frozen to the background g.

Equations
  • One or more equations did not get rendered due to their size.
Instances For
    theorem CommutingRepetition.scenMartA_law {n : } {X Y A : Type} [Fintype X] [Fintype Y] [Fintype A] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [Nonempty X] [Nonempty Y] [Nonempty A] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : X, μ v (yw c) 0) (ω : LX) :
    (scenMartA SA L hdisj π μ yw g za hM).law ω = (∏ c : L, μ (ω c) (yw c)) / c : L, v : X, μ v (yw c)
    theorem CommutingRepetition.scenMartA_idx {n : } {X Y A : Type} [Fintype X] [Fintype Y] [Fintype A] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [Nonempty X] [Nonempty Y] [Nonempty A] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : X, μ v (yw c) 0) (j : Fin (L.card + 1)) (ω : LX) :
    (scenMartA SA L hdisj π μ yw g za hM).idx j ω = mkALabel D (SA ordPrefix π j) ((wordSplit L X).symm (ω, g)) yw za
    theorem CommutingRepetition.scenMartA_univ_eq {n : } {X Y A : Type} [Fintype X] [Fintype Y] [Fintype A] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [Nonempty X] [Nonempty Y] [Nonempty A] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : X, μ v (yw c) 0) :

    The martingale's path-space Fintype instFintypeΩ is definitionally the canonical one on ↥L → X, but only at default transparency, so this provable (non-rfl-trivial) univ equality is not skipped by simp and canonicalizes the summation index after simp only [scenMartA] leaves the projected instance behind.

    theorem CommutingRepetition.mem_ordPrefix_iff {n : } {L : Finset (Fin n)} (π : Fin L.card L) {m : } {j : Fin n} :
    j ordPrefix π m ∃ (t : Fin L.card), t < m (π t) = j
    theorem CommutingRepetition.mem_ordPrefix_of_lt {n : } {L : Finset (Fin n)} (π : Fin L.card L) {t : Fin L.card} {m : } (ht : t < m) :
    (π t) ordPrefix π m
    noncomputable def CommutingRepetition.scenMartB {n : } {X Y B : Type} [Fintype X] [Fintype Y] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty B] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : Y, μ (xw c) v 0) :
    EffectMartingale (Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)) L.card

    The Bob-side scenario reveal martingale: the path space is the Bob values on the progressively revealed block L, the law is the pinned-Alice conditional weight (normalized by the block mass), and the index path is the canonical Bob label at the growing revealed set SA ∪ π[1..j], freezing the off-block values to a background g.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For
      theorem CommutingRepetition.scenMartB_law {n : } {X Y B : Type} [Fintype X] [Fintype Y] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty B] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : Y, μ (xw c) v 0) (ω : LY) :
      (scenMartB SA L hdisj π μ xw g zb hM).law ω = (∏ c : L, μ (xw c) (ω c)) / c : L, v : Y, μ (xw c) v
      theorem CommutingRepetition.scenMartB_idx {n : } {X Y B : Type} [Fintype X] [Fintype Y] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty B] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : Y, μ (xw c) v 0) (j : Fin (L.card + 1)) (ω : LY) :
      (scenMartB SA L hdisj π μ xw g zb hM).idx j ω = mkBLabel D (SA ordPrefix π j) xw ((wordSplit L Y).symm (ω, g)) zb
      theorem CommutingRepetition.scenMartB_univ_eq {n : } {X Y B : Type} [Fintype X] [Fintype Y] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty B] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : Y, μ (xw c) v 0) :

      Bob-side Fintype canonicalization (mirror of scenMartA_univ_eq).

      theorem CommutingRepetition.scenMartA_law_mul_M0 {n : } {X Y A : Type} [Fintype X] [Fintype Y] [Fintype A] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [Nonempty X] [Nonempty Y] [Nonempty A] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : X, μ v (yw c) 0) (ω : LX) :
      (scenMartA SA L hdisj π μ yw g za hM).law ω * c : L, v : X, μ v (yw c) = c : L, μ (ω c) (yw c)

      Block-law renormalization, Alice (node 1.2.9 assembly step (2)): the scenario law times the total block mass M₀ is the unnormalized block product weight ∏_{c∈L} μ(ω c, yw c). This converts the martingale's law-weighted block-value sum into the alignCostA reference-word prior sum after the wordSplit at L.

      theorem CommutingRepetition.scenMartB_law_mul_M0 {n : } {X Y B : Type} [Fintype X] [Fintype Y] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty B] {D : Finset (Fin n)} (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (μ : XY) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) ( : ∀ (x : X) (y : Y), 0 μ x y) (hM : c : L, v : Y, μ (xw c) v 0) (ω : LY) :
      (scenMartB SA L hdisj π μ xw g zb hM).law ω * c : L, v : Y, μ (xw c) v = c : L, μ (xw c) (ω c)

      Block-law renormalization, Bob (mirror of scenMartA_law_mul_M0).

      theorem CommutingRepetition.TracialStrategy.scenMartA_isMeanTower {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af : Type} [Fintype Af] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} (htotF : ∀ (s : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)), a : Af, Ffam s a = S.setEffectA D s.1 μ s.2.1 s.2.2.1 s.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) (hM : c : L, v : X, μ v (yw c) 0) :
      (scenMartA SA L hdisj π μ yw g za hM).IsMeanTower fun (i : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)) => a : Af, Ffam i a

      The scenario martingale is a mean tower for the arena's Alice totals: on each fiber of the current canonical label, the law-weighted average of the next revealed-set effect is the current one (setEffectA_reveal, fiber-refined).

      theorem CommutingRepetition.TracialStrategy.scenMartA_colBudget {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hcol : R.ColEntropyBudget) (htotF : ∀ (s : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)), a : Af, Ffam s a = S.setEffectA D s.1 μ s.2.1 s.2.2.1 s.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) (hM : c : L, v : X, μ v (yw c) 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (j : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)) (ω₀ : LX) :
      s : Fin L.card, ω : LX, (scenMartA SA L hdisj π μ yw g za hM).law ω * R.branch σ ((scenMartA SA L hdisj π μ yw g za hM).idx s.succ ω) j - R.branch σ ((scenMartA SA L hdisj π μ yw g za hM).idx s.castSucc ω) j ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam ((scenMartA SA L hdisj π μ yw g za hM).idx 0 ω₀) a) * σ * b : Bf, Gfam j b))).re.negMulLog

      Alice-column budget, consumed on the scenario martingale (node 1.2.6 → 1.2.9 bridge): the reveal-martingale mean tower feeds the signed column budget, so the total law-weighted squared L²-increment of the branch vectors along the Alice scenario reveal (block L, background SA, pinned Bob index j) is at most the scalar entropy H₁ of the initial branch pairing. This is the per-scenario budget consumption that node 1.2.9 telescopes at the uniform cut.

      theorem CommutingRepetition.TracialStrategy.scenMartB_isMeanTower {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Bf : Type} [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (htotG : ∀ (t : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)), b : Bf, Gfam t b = S.setEffectB D t.1 μ t.2.1 t.2.2.1 t.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) (hM : c : L, v : Y, μ (xw c) v 0) :
      (scenMartB SA L hdisj π μ xw g zb hM).IsMeanTower fun (j : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)) => b : Bf, Gfam j b

      The Bob scenario martingale is a mean tower for the arena's Bob totals (setEffectB_reveal, fiber-refined).

      theorem CommutingRepetition.TracialStrategy.scenMartB_rowBudget {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hrow : R.RowEntropyBudget) (htotG : ∀ (t : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)), b : Bf, Gfam t b = S.setEffectB D t.1 μ t.2.1 t.2.2.1 t.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) (hM : c : L, v : Y, μ (xw c) v 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (i : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)) (ω₀ : LY) :
      s : Fin L.card, ω : LY, (scenMartB SA L hdisj π μ xw g zb hM).law ω * R.branch σ i ((scenMartB SA L hdisj π μ xw g zb hM).idx s.succ ω) - R.branch σ i ((scenMartB SA L hdisj π μ xw g zb hM).idx s.castSucc ω) ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam i a) * σ * b : Bf, Gfam ((scenMartB SA L hdisj π μ xw g zb hM).idx 0 ω₀) b))).re.negMulLog

      Bob-row budget, consumed on the scenario martingale (mirror of scenMartA_colBudget): feeding the Bob mean tower to RowEntropyBudget bounds the total law-weighted squared L²-increment of the branch vectors (second slot) along the Bob scenario reveal by the scalar entropy H₁ of the initial branch pairing.

      theorem CommutingRepetition.TracialStrategy.scenMartA_cut_le {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hcol : R.ColEntropyBudget) (htotF : ∀ (s : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)), a : Af, Ffam s a = S.setEffectA D s.1 μ s.2.1 s.2.2.1 s.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) (hM : c : L, v : X, μ v (yw c) 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (j : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)) (ω₀ : LX) (k : Fin L.card) :
      ω : LX, (scenMartA SA L hdisj π μ yw g za hM).law ω * R.branch σ ((scenMartA SA L hdisj π μ yw g za hM).idx k.succ ω) j - R.branch σ ((scenMartA SA L hdisj π μ yw g za hM).idx k.castSucc ω) j ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam ((scenMartA SA L hdisj π μ yw g za hM).idx 0 ω₀) a) * σ * b : Bf, Gfam j b))).re.negMulLog

      Single cut-step column budget (node 1.2.9 assembly step (1)): dropping the other nonnegative reveal steps, the law-weighted squared increment at any single step k is bounded by the same scalar entropy H₁. This is the per-datum cut increment that node 1.2.9 identifies with the alignCostA integrand.

      theorem CommutingRepetition.TracialStrategy.scenMartB_cut_le {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hrow : R.RowEntropyBudget) (htotG : ∀ (t : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)), b : Bf, Gfam t b = S.setEffectB D t.1 μ t.2.1 t.2.2.1 t.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) (hM : c : L, v : Y, μ (xw c) v 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (i : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)) (ω₀ : LY) (k : Fin L.card) :
      ω : LY, (scenMartB SA L hdisj π μ xw g zb hM).law ω * R.branch σ i ((scenMartB SA L hdisj π μ xw g zb hM).idx k.succ ω) - R.branch σ i ((scenMartB SA L hdisj π μ xw g zb hM).idx k.castSucc ω) ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam i a) * σ * b : Bf, Gfam ((scenMartB SA L hdisj π μ xw g zb hM).idx 0 ω₀) b))).re.negMulLog

      Single cut-step row budget (Bob-side mirror of scenMartA_cut_le).

      theorem CommutingRepetition.TracialStrategy.scenMartA_colBudget_mkALabel {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hcol : R.ColEntropyBudget) (htotF : ∀ (s : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)), a : Af, Ffam s a = S.setEffectA D s.1 μ s.2.1 s.2.2.1 s.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (yw : Fin nY) (g : { j : Fin n // jL }X) (za : DA) (hM : c : L, v : X, μ v (yw c) 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (J : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)) (ω₀ : LX) :
      s : Fin L.card, ω : LX, (scenMartA SA L hdisj π μ yw g za hM).law ω * R.branch σ (mkALabel D (insert (↑(π s)) (SA ordPrefix π s)) ((wordSplit L X).symm (ω, g)) yw za) J - R.branch σ (mkALabel D (SA ordPrefix π s) ((wordSplit L X).symm (ω, g)) yw za) J ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam (mkALabel D SA ((wordSplit L X).symm (ω₀, g)) yw za) a) * σ * b : Bf, Gfam J b))).re.negMulLog

      Column budget in canonical-label form (node 1.2.9 assembly bridge): scenMartA_colBudget with the martingale's .idx steps expanded via scenMartA_idx/ordPrefix_succ into the mkALabel canonical labels that the alignCostA integrand (through the aLabel = mkALabel rfl bridge and the reveal-datum wiring revealMartA_cut(Succ)_eq_effectiveH(Bar)) actually names. The total, over all reveal steps s, of the law-weighted squared branch increment ‖φ_{insert (π s) prefix} − φ_{prefix}‖² against a fixed Bob index J is bounded by H₁ of the initial pairing (label at the background SA). Node 1.2.9 identifies each alignCostA cut-datum (at cut kY) with the step s = kY (prefix = SA ∪ ordPrefix π kY = C_X, insert (π kY) prefix = insert i C_X); summing the cut-data over the uniform interior cut kY : Fin L.card reconstructs this ∑ s telescoping sum, because the Bob index bLabel (insert i C_Y) = bLabel (D ∪ L ∪ alicePrefixX) is kY-independent (insert i C_Y collapses off the live coordinate).

      theorem CommutingRepetition.TracialStrategy.scenMartB_rowBudget_mkBLabel {n : } {X Y A B : Type} [Fintype X] [Fintype Y] [Fintype A] [Fintype B] [DecidableEq X] [DecidableEq Y] [DecidableEq A] [DecidableEq B] [Nonempty X] [Nonempty Y] [Nonempty A] [Nonempty B] (S : TracialStrategy (Fin nX) (Fin nY) (Fin nA) (Fin nB)) {D : Finset (Fin n)} {Af Bf : Type} [Fintype Af] [Fintype Bf] (μ : XY) ( : ∀ (x : X) (y : Y), 0 μ x y) {Ffam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)AfS.M.A} {Gfam : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)BfS.M.A} (R : ResolverArena S.M Ffam Gfam) (hrow : R.RowEntropyBudget) (htotG : ∀ (t : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nB)), b : Bf, Gfam t b = S.setEffectB D t.1 μ t.2.1 t.2.2.1 t.2.2.2) (SA L : Finset (Fin n)) (hdisj : Disjoint SA L) (π : Fin L.card L) (xw : Fin nX) (g : { j : Fin n // jL }Y) (zb : DB) (hM : c : L, v : Y, μ (xw c) v 0) (σ : S.M.A) ( : S.M.τ (star σ * σ) = 1) (I : Finset (Fin n) × (Fin nX) × (Fin nY) × (Fin nA)) (ω₀ : LY) :
      s : Fin L.card, ω : LY, (scenMartB SA L hdisj π μ xw g zb hM).law ω * R.branch σ I (mkBLabel D (insert (↑(π s)) (SA ordPrefix π s)) xw ((wordSplit L Y).symm (ω, g)) zb) - R.branch σ I (mkBLabel D (SA ordPrefix π s) xw ((wordSplit L Y).symm (ω, g)) zb) ^ 2 (S.M.τ (star σ * ((∑ a : Af, Ffam I a) * σ * b : Bf, Gfam (mkBLabel D SA xw ((wordSplit L Y).symm (ω₀, g)) zb) b))).re.negMulLog

      Row budget in canonical-label form (Bob-side mirror of scenMartA_colBudget_mkALabel), for the alignCostB cut assembly.