Four-point error K implies 2K-thinness of geodesic triangles
ProvedRayBundle.MetricSegment.four_point_thingeodesicshyperbolic-geometrymetric-geometry
Let be a metric space and . Suppose all satisfy
Let , , and be specified arc-length geodesic segments. Every point on has a point on one of the other two segments with
Applying the result to each side gives a -thin geodesic triangle. The theorem concerns any three supplied segments; it does not assume global geodesic existence. With the four-point convention whose error is , this estimate gives -thinness.
Preamble
import Definitions.Def_RayBundle_MetricPathGeometry
Formal statement
theorem RayBundle.MetricSegment.four_point_thin {X : Type*} [MetricSpace X] (K : ℝ) (hK : 0 ≤ K)
(hfour : ∀ a b c d : X, dist a b + dist c d ≤
max (dist a c + dist b d) (dist a d + dist b c) + K)
{a b c : X} (ab : RayBundle.MetricSegment a b) (ac : RayBundle.MetricSegment a c) (bc : RayBundle.MetricSegment b c)
(t : Set.Icc (0 : ℝ) (dist a b)) :
∃ q, (q ∈ Set.range ac.map ∨ q ∈ Set.range bc.map) ∧ dist (ab.map t) q ≤ 2 * K := by sorry
Source
Source declaration: RayBundle/Thm_Cayley_MetricSegment_four_point_thin.lean, theorem RayBundle.MetricSegment.four_point_thin. Supporting quantitative hyperbolicity lemma proved in this development using two four-point comparisons. Thin-triangle and unit-edge graph conventions are described in Nicholas Touikan, On geodesic ray bundles in hyperbolic groups (2018), Section 2, https://arxiv.org/abs/1706.01979 . The constant here is the explicitly proved bound, not a numbered theorem attribution in that paper.