nae_angle_contraction
For 0 ≤ θ ≤ π, arccos(1−(4/9)(1−cos θ)²) ≤ 15θ/16.
Statement
theorem nae_angle_contraction (θ : ℝ) (hθ0 : 0 ≤ θ) (hθ : θ ≤ Real.pi) :
Real.arccos (1 - (4 / 9 : ℝ) * (1 - Real.cos θ) ^ 2) ≤
(15 / 16 : ℝ) * θ- Source
- arXiv:1211.0721, Section 3.3, p-computation rule analysis
- Verified
- 14 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 1
- nae_sine_contractionFor 0 ≤ t ≤ π/2, (√8/3) sin²(t) ≤ sin(15t/16).
- Used by 1
- nae_p_schedule_potentialFor every triple (d, θ, q) reachable by the specified p-computation rule schedule, 0 ≤ θ ≤ π, q ≥ 0, and θ ≤ πq(15/32)^d.
The proof, explained
The proof reduces the arccosine contraction inequality to an algebraic inequality between sines by expressing cosines via the half-angle formula 1 - cos(2x) = 2 sin²(x). Applying the companion lemma nae_sine_contraction at θ / 2 gives (√8 / 3) sin²(θ / 2) ≤ sin((15 / 16)(θ / 2)), which upon squaring yields (8 / 9) sin⁴(θ / 2) ≤ sin²((15 / 16)(θ / 2)). Translating this back to cosines yields cos((15 / 16) θ) ≤ 1 - (4 / 9)(1 - cos θ)², and taking the arccosine of both sides reverses the inequality to produce the desired contraction.
The sine contraction (√8 / 3) sin²(θ / 2) ≤ sin((15 / 16) * (θ / 2)) holds, and both sides are non-negative.
Since 0 ≤ θ ≤ π, the half-angle satisfies 0 ≤ θ / 2 ≤ π / 2, so the auxiliary lemma nae_sine_contraction applies at t = θ / 2. Furthermore, (15 / 16) * (θ / 2) lies in [0, π], which ensures sin((15 / 16) * (θ / 2)) ≥ 0 by Real.sin_nonneg_of_nonneg_of_le_pi and Real.pi_pos, while the left-hand side is non-negative by positivity.
uses
nae_sine_contraction,Real.sin_nonneg_of_nonneg_of_le_pi,Real.pi_poshave hs := nae_sine_contraction (θ / 2) (by linarith) (by linarith) have hs0 : 0 ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) := Real.sin_nonneg_of_nonneg_of_le_pi (by positivity) (by nlinarith [Real.pi_pos]) have hc0 : 0 ≤ Real.sqrt 8 / 3 * Real.sin (θ / 2) ^ 2 := by positivity
(√8 / 3)² = 8 / 9.
Distributing the square over the quotient with div_pow gives (√8)² / 3², and because 8 ≥ 0, Real.sq_sqrt simplifies (√8)² to 8, yielding 8 / 9.
uses
div_pow,Real.sq_sqrthave hsq : (Real.sqrt 8 / 3) ^ 2 = (8 / 9 : ℝ) := by rw [div_pow, Real.sq_sqrt (by norm_num : (0 : ℝ) ≤ 8)] norm_num
(8 / 9) sin⁴(θ / 2) ≤ sin²((15 / 16) * (θ / 2)).
Since both sides of the sine contraction inequality are non-negative, squaring both sides preserves the inequality (mul_self_le_mul_self). Rewriting (√8 / 3 * sin²(θ / 2))² as ((√8 / 3)² * sin⁴(θ / 2)) and applying the previous identity (√8 / 3)² = 8 / 9 alongside the definition of squaring (pow_two) gives the desired bound.
uses
mul_self_le_mul_self,pow_twohave hsquared := mul_self_le_mul_self hc0 hs have hsquared' : (8 / 9 : ℝ) * Real.sin (θ / 2) ^ 4 ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by calc (8 / 9 : ℝ) * Real.sin (θ / 2) ^ 4 = (Real.sqrt 8 / 3 * Real.sin (θ / 2) ^ 2) ^ 2 := by rw [← hsq]; ring _ ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by simpa [pow_two] using hsquared
The half-angle identities cos θ = 1 - 2 sin²(θ / 2) and cos((15 / 16) θ) = 1 - 2 sin²((15 / 16) * (θ / 2)) hold.
From the double-angle formula cos(2x) = 2 cos²(x) - 1 (Real.cos_two_mul) and the Pythagorean identity sin²(x) + cos²(x) = 1 (Real.sin_sq_add_cos_sq), one derives cos(2x) = 1 - 2 sin²(x). Applying this identity with x = θ / 2 and x = (15 / 16) * (θ / 2) produces the two expressions.
uses
Real.cos_two_mul,Real.sin_sq_add_cos_sqhave hcos1 : Real.cos θ = 1 - 2 * Real.sin (θ / 2) ^ 2 := by have h := Real.cos_two_mul (θ / 2) have hid : 2 * (θ / 2) = θ := by ring rw [hid] at h nlinarith [Real.sin_sq_add_cos_sq (θ / 2)] have hcos2 : Real.cos ((15 / 16 : ℝ) * θ) = 1 - 2 * Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by have h := Real.cos_two_mul ((15 / 16 : ℝ) * (θ / 2)) have hid : 2 * ((15 / 16 : ℝ) * (θ / 2)) = (15 / 16 : ℝ) * θ := by ring rw [hid] at h nlinarith [Real.sin_sq_add_cos_sq ((15 / 16 : ℝ) * (θ / 2))]
cos((15 / 16) θ) ≤ 1 - (4 / 9) (1 - cos θ)².
Substituting the half-angle formulas for cos θ and cos((15 / 16) θ) rewrites the claim as 1 - 2 sin²((15 / 16) * (θ / 2)) ≤ 1 - (16 / 9) sin⁴(θ / 2), which is algebraically equivalent to the squared sine contraction inequality established earlier.
have hcomp : Real.cos ((15 / 16 : ℝ) * θ) ≤ 1 - (4 / 9 : ℝ) * (1 - Real.cos θ) ^ 2 := by rw [hcos1, hcos2] nlinarith [hsquared']
arccos(1 - (4 / 9) (1 - cos θ)²) ≤ (15 / 16) θ.
Applying the arccosine function reverses the inequality because arccosine is order-reversing (Real.arccos_le_arccos). Since 0 ≤ (15 / 16) θ ≤ π, arccos(cos((15 / 16) θ)) simplifies directly to (15 / 16) θ by Real.arccos_cos (using Real.pi_pos), finishing the proof.
uses
Real.arccos_le_arccos,Real.arccos_cos,Real.pi_poshave hanti := Real.arccos_le_arccos hcomp rw [Real.arccos_cos (by positivity) (by nlinarith [Real.pi_pos])] at hanti exact hanti
nae_sine_contraction- For 0 ≤ t ≤ π/2, (√8 / 3) sin²(t) ≤ sin(15t / 16).
Real.sin_nonneg_of_nonneg_of_le_pi- The sine function is non-negative on [0, π].
Real.pi_pos- The constant π is strictly positive.
div_pow- For any real numbers a and b and natural number n, (a / b)ⁿ = aⁿ / bⁿ.
Real.sq_sqrt- For any non-negative real number x, (√x)² = x.
mul_self_le_mul_self- If 0 ≤ a ≤ b in an ordered semiring, then a² ≤ b².
pow_two- For any element a in a monoid, a² = a * a.
Real.cos_two_mul- The double-angle cosine identity: cos(2x) = 2 cos²(x) - 1.
Real.sin_sq_add_cos_sq- The Pythagorean trigonometric identity: sin²(x) + cos²(x) = 1.
Real.arccos_le_arccos- The arccosine function is decreasing: if x ≤ y, then arccos(y) ≤ arccos(x).
Real.arccos_cos- For any x ∈ [0, π], arccos(cos(x)) = x.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
theorem nae_angle_contraction (θ : ℝ) (hθ0 : 0 ≤ θ) (hθ : θ ≤ Real.pi) : Real.arccos (1 - (4 / 9 : ℝ) * (1 - Real.cos θ) ^ 2) ≤ (15 / 16 : ℝ) * θ := by have hs := nae_sine_contraction (θ / 2) (by linarith) (by linarith) have hs0 : 0 ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) := Real.sin_nonneg_of_nonneg_of_le_pi (by positivity) (by nlinarith [Real.pi_pos]) have hc0 : 0 ≤ Real.sqrt 8 / 3 * Real.sin (θ / 2) ^ 2 := by positivity have hsq : (Real.sqrt 8 / 3) ^ 2 = (8 / 9 : ℝ) := by rw [div_pow, Real.sq_sqrt (by norm_num : (0 : ℝ) ≤ 8)] norm_num have hsquared := mul_self_le_mul_self hc0 hs have hsquared' : (8 / 9 : ℝ) * Real.sin (θ / 2) ^ 4 ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by calc (8 / 9 : ℝ) * Real.sin (θ / 2) ^ 4 = (Real.sqrt 8 / 3 * Real.sin (θ / 2) ^ 2) ^ 2 := by rw [← hsq]; ring _ ≤ Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by simpa [pow_two] using hsquared have hcos1 : Real.cos θ = 1 - 2 * Real.sin (θ / 2) ^ 2 := by have h := Real.cos_two_mul (θ / 2) have hid : 2 * (θ / 2) = θ := by ring rw [hid] at h nlinarith [Real.sin_sq_add_cos_sq (θ / 2)] have hcos2 : Real.cos ((15 / 16 : ℝ) * θ) = 1 - 2 * Real.sin ((15 / 16 : ℝ) * (θ / 2)) ^ 2 := by have h := Real.cos_two_mul ((15 / 16 : ℝ) * (θ / 2)) have hid : 2 * ((15 / 16 : ℝ) * (θ / 2)) = (15 / 16 : ℝ) * θ := by ring rw [hid] at h nlinarith [Real.sin_sq_add_cos_sq ((15 / 16 : ℝ) * (θ / 2))] have hcomp : Real.cos ((15 / 16 : ℝ) * θ) ≤ 1 - (4 / 9 : ℝ) * (1 - Real.cos θ) ^ 2 := by rw [hcos1, hcos2] nlinarith [hsquared'] have hanti := Real.arccos_le_arccos hcomp rw [Real.arccos_cos (by positivity) (by nlinarith [Real.pi_pos])] at hanti exact hanti