Documentation

LeanPool.MarkoffModP.BGS.Markoff.ExplicitNumericCertificates

Numerical certificates for the explicit strong-approximation cutoff #

This file packages the closed real-power and natural-power calculations used by the explicit middle game, endgame, cage, and small-order argument.

The certificates for the first explicit record remain available for comparison, but the canonical declarations below use the ninth-moment cutoff.

theorem BGS.Markoff.Legacy.explicit_middleGame_corvajaZannier_sizeBounds {p currentOrder : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hLower : ↑p ^ (7 / 32) < ↑currentOrder) (hUpper : ↑currentOrder < ↑p ^ (25 / 32)) :

At the explicit cutoff, the all-divisors Corvaja--Zannier estimate obeys both finite size conditions throughout the range p^(7/32) < currentOrder < p^(25/32).

theorem BGS.Markoff.Legacy.explicit_endgamePrimitiveTrace_explicitInequality {p orbitExponent coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (horbit : ↑orbitExponent ≤ 2 * ↑p ^ (9 / 32)) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

The primitive-inclusion--exclusion error is strictly smaller than its main term for an exponent at most 2*p^(9/32) and every coefficient at most 34.

theorem BGS.Markoff.Legacy.explicit_endgamePrimitiveTrace_explicitInequality_of_card_sub_one {p orbitExponent orbitOrder coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hmul : orbitExponent * orbitOrder = p - 1) (horder : ↑p ^ (23 / 32) ≤ ↑orbitOrder) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

Split-torus form of the explicit primitive-trace certificate.

theorem BGS.Markoff.Legacy.explicit_endgamePrimitiveTrace_explicitInequality_of_card_add_one {p orbitExponent orbitOrder coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hmul : orbitExponent * orbitOrder = p + 1) (horder : ↑p ^ (23 / 32) ≤ ↑orbitOrder) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

Nonsplit-torus form of the explicit primitive-trace certificate.

The trace-zero order bound 4 lies below the explicit endgame threshold.

The middle-game lower threshold also dominates the order-four exceptional case.

theorem BGS.Markoff.Legacy.explicit_cageWitness_explicitInequality {p coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hcoefficient : coefficient ≤ 100522) :
↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

The coefficient 100522 cage error is dominated by p at the explicit cutoff.

theorem BGS.Markoff.Legacy.explicit_lowOrder_contradiction {p d : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hpLe : p ≤ 32 * d ^ 4) (hd : ↑d ≤ ↑p ^ (7 / 32)) :

The elementary low-order count cannot contain a nonempty orbit whose cardinality is divisible by p once d ≤ p^(7/32).

theorem BGS.Markoff.Legacy.explicit_lowOrder_contradiction_512 {p d : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hpLe : p ≤ 512 * d ^ 4) (hd : ↑d ≤ ↑p ^ (7 / 32)) :

Variant of explicit_lowOrder_contradiction for the d + 1 cutoff used when maximality turns a strict order bound into a finite-set inclusion.

theorem BGS.Markoff.explicit_middleGame_corvajaZannier_linear_of_lt_fiveSixths {p currentOrder : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hUpper : ↑currentOrder < ↑p ^ (5 / 6)) :

Below the five-sixths threshold, the coefficient-48 divisor term times the current order is strictly smaller than p.

theorem BGS.Markoff.explicit_middleGame_corvajaZannier_linearBound {p currentOrder : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hUpper : ↑currentOrder < ↑p ^ (5 / 6)) :

Assembly-facing alias for the five-sixths middle-game certificate.

theorem BGS.Markoff.explicit_endgamePrimitiveTrace_explicitInequality {p orbitExponent coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (horbit : ↑orbitExponent ≤ 2 * ↑p ^ (1 / 6)) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

Primitive-trace inclusion--exclusion at exponent at most 2*p^(1/6).

theorem BGS.Markoff.explicit_endgamePrimitiveTrace_explicitInequality_of_card_sub_one {p orbitExponent orbitOrder coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hmul : orbitExponent * orbitOrder = p - 1) (horder : ↑p ^ (5 / 6) ≤ ↑orbitOrder) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

Split-torus form of the improved primitive-trace certificate.

theorem BGS.Markoff.explicit_endgamePrimitiveTrace_explicitInequality_of_card_add_one {p orbitExponent orbitOrder coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hmul : orbitExponent * orbitOrder = p + 1) (horder : ↑p ^ (5 / 6) ≤ ↑orbitOrder) (hcoefficient : coefficient ≤ 34) :
↑orbitExponent * ↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

Nonsplit-torus form of the improved primitive-trace certificate.

The trace-zero order bound lies below the improved endgame threshold.

theorem BGS.Markoff.explicit_cageWitness_explicitInequality {p coefficient : ℕ} (hp : explicitStrongApproximationCutoff ≤ p) (hcoefficient : coefficient ≤ 100522) :
↑(p - 1).divisors.card ^ 2 * (↑coefficient * √↑p) < ↑p

The coefficient-100522 cage error is dominated by p.

A divisor-sensitive small-order set large enough to contain a nonempty p-divisible orbit forces the Corvaja--Zannier cube below the maximal order.

Assembly-facing alias for the divisor-sensitive cube certificate.