Documentation

LeanPool.BrillNoetherGraphs.Bananas.CrossOneOff.CrossOneOffPeriodSeparation

Period separation for the corrected cross-one-off block #

The existing torsion dichotomy records only g ≤ k, but its nonzero-rise argument actually proves the strict inequality g < k. In the long-second- strand regime the zero-rise exception is impossible. Since then crossOneOffCutoff g n ≤ g+1, this supplies exactly the separation needed by the corrected finite-row inversion count.

theorem Bananas.crossOneOff_torsionOrder_gt_genus_of_second_long {g k : ℕ} (B : Banana g) (alpha beta : Fin (g + 1)) (hg : 3 ≤ g) (hab : alpha ≠ beta) (hAlpha : 1 < B.length alpha) (hBetaLong : g + 1 ≤ B.length beta) (hTO : IsTorsionOrder (mark (Utilities.Certificate.SubdivisionGraph.Spec.graph B) (strandVertex B alpha ⟨1, ⋯⟩) (strandVertex B beta ⟨B.length beta - 1, ⋯⟩)) k) :
g < k

A cross-one-off exact torsion order is strictly larger than the genus when the second strand has length at least g+1.

theorem Bananas.crossOneOff_cutoff_le_torsionOrder {g k : ℕ} (B : Banana g) (alpha beta : Fin (g + 1)) (hg : 3 ≤ g) (hab : alpha ≠ beta) (hAlpha : 1 < B.length alpha) (hBetaLong : g + 1 ≤ B.length beta) (hTO : IsTorsionOrder (mark (Utilities.Certificate.SubdivisionGraph.Spec.graph B) (strandVertex B alpha ⟨1, ⋯⟩) (strandVertex B beta ⟨B.length beta - 1, ⋯⟩)) k) :

The exact torsion order separates the whole corrected row cutoff from the affine period.

theorem Bananas.crossOneOff_kGeneral_cutoff_le_period {g k : ℕ} (B : Banana g) (alpha beta : Fin (g + 1)) (hg : 3 ≤ g) (hab : alpha ≠ beta) (hAlpha : 1 < B.length alpha) (hBetaLong : g + 1 ≤ B.length beta) (hK : KGeneralTransmission (mark (Utilities.Certificate.SubdivisionGraph.Spec.graph B) (strandVertex B alpha ⟨1, ⋯⟩) (strandVertex B beta ⟨B.length beta - 1, ⋯⟩)) k) :

In particular, a k-general cross-one-off marking supplies the missing period-separation premise in corrected Corollary 4.31.