The 120-degree wedge and the area-loss lemma #
sideForm i is the linear form whose level set = 1 is the hexagon side [vertex i, vertex (next i)].
wedge i is the closed 120-degree wedge at vertex i bounded by the two sides through it; it
contains
the hexagon. The area-loss lemma bounds the part of an equilateral triangle outside the wedge in
terms of the lengths of its intersections with the two rays.
Human proof: notes/waste_cycle_reduction.md §1 (geometric) and
notes/one_rise_area_reduction_2026_09_06.md §2 (normalised linear forms). In skew units the
physical statement |T \ W| ≥ (√3/4)·min(a², b²) reads volume ≥ min(a², b²)/2.
The closed 120-degree wedge at vertex i bounded by the two incident sides.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Toolkit: membership in a filled triangle #
Half-plane description of a positively oriented triangle: the three edge inequalities put a point into the filled triangle.
The polygon in ray coordinates #
Write points of the wedge at vertex i as vertex i + α • u + β • w with u, w the unit
directions
of the two rays. The three edge functionals of an equilateral triangle become
σₖ + cₖ α + c_{k+1} β (the coefficient along the second ray is the cyclic shift of the first),
with σ₀ + σ₁ + σ₂ = s², c₀ + c₁ + c₂ = 0, c₀² + c₁² + c₂² = 2 s². Everything below is pure
algebra in these coordinates; Point = ℝ × ℝ is reused for ray coordinates, coordDet for areas.
Area of a triangle in ray coordinates.
Equations
- ConwaySoifer.SevenTriangles.triArea q₁ q₂ q₃ = |ConwaySoifer.SevenTriangles.coordDet (q₂ - q₁) (q₃ - q₁)| / 2
Instances For
The polygon is covered by three (possibly degenerate) triangles of total area at most A.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Both rays exit through the same side σ + c α + d β = 0: the polygon is the triangle
O, A, B.
Quadrilateral case: the first ray exits through side 0, the second through side 2, and the
common vertex V₀ of these sides lies strictly inside the wedge. Two triangles O A V₀, O V₀ B.
Pentagonal case: the first ray exits through side 0, the second through side 1, and the
common vertex V₁ of these sides is outside the wedge; the other two vertices V₀, V₂ are inside.
Three triangles O A V₀, O V₀ B, V₀ V₂ B. Here c₁ = -c₀ - c₂ and σ₂ = D + c₀ a + c₂ b.
Algebraic core of the strengthened bound: with s² = X² + X Z + Z² and
(X + Z) a + X b < s (a + b), the quadratic form (X+Z) b² + 2 Z a b - X a² is nonnegative.
The polynomial inequalities behind the two bounds #
Quadrilateral, weak bound: D · min(a², b²) ≤ D² - X (2 X a b + (X - c₁) a² + c₁ b²) with
D = X² - X c₁ + c₁².
Quadrilateral, strong bound (s < a + b).
Pentagon, strong bound (s < a + b), with D ≥ C a + Z b and D > X a + C b.
The quadrilateral area bound in the form produced by coverBound_quad.
The pentagon area bound in the form produced by coverBound_pent.
The case tree when the first ray exits through side 0.
The case tree for an arbitrary exit side of the first ray, by cyclic relabelling.
Extension of a contained length to the full length: the first side hit along the ray.
The abstract wedge lemma: from contained lengths a, b > 0 to a cover of the polygon by
three triangles with the required total area.
Concrete filled triangles: half-plane description #
Every filled equilateral triangle is a positively oriented A, A + e, A + rotate60 e.
Ray coordinates at a hexagon vertex #
Unit direction of the ray from vertex i towards vertex (previous i).
Equations
Instances For
Unit direction of the ray from vertex i towards vertex (next i).
Equations
Instances For
Coordinate along rayW i (vanishes on the side [vertex (previous i), vertex i]).
Equations
Instances For
The three edge functionals of the triangle A, A+e, A+rotate60 e in ray coordinates at
vertex i: the coefficient along the second ray is the cyclic shift of the one along the first.
The affine ray map and areas #
Subadditivity: a cover of T ∩ W by three images of ray-coordinate triangles.
A cover bound in ray coordinates yields the area-loss bound.
Weak area-loss bound. a is a length along the ray towards vertex (previous i), b along
the
ray towards vertex (next i); both rays are unit length in sqDist. Any contained lengths may be
used (the bound is monotone), not only the full intersection lengths.
Strong area-loss bound, valid when the two contained lengths exceed the side.
Loss outside the hexagon dominates loss outside the wedge.