Optimal Running Path Around a Triangular House
One-line statement: Given a triangular house and a runner who starts at a point, must touch each of the three walls (sides), and return to start, the optimal (shortest) running path is found using reflection of the starting point across the walls. Key result: When the runner must touch all three sides in order and return, the minimal closed path is a straight line in the unfolded (reflected) plane; for the special case of touching each side once the inscribed path is governed by reflection and, for the triangle-inscribed-triangle problem, by the orthic triangle. Where it applies: Optimization, geometric reflection ("billiard") problems, the calculus of shortest paths, recreational and competition mathematics. Who/when: Rooted in Heron of Alexandria (c. 60 AD) on reflection; Fagnano's inscribed-triangle problem (1775); modern billiard dynamics.
A path that bounces off a wall obeys the law of reflection (angle in = angle out) precisely when it is shortest. Equivalently, reflecting the start point across a wall turns "go to the wall then continue" into "go straight to the mirror image." Chain these reflections and the optimal multi-touch path becomes a single straight segment in an unfolded copy of the plane.
Formal Statement
Let triangle represent the house, with side lengths , , . A runner located at point must touch a sequence of walls and return to . We measure the total path length and seek its minimum. Two canonical versions exist:
- (V1) Fixed start, return to start, touch one wall: Touch line once and come back; minimize -type sums.
- (V2) Inscribed triangle / full lap: Choose one point on each side, , , , and run the closed loop . Minimize the perimeter
Derivation hint: This is just the perimeter of inscribed triangle ; the optimization is over the three free points constrained to the sides.
Given a point and a line , the reflection satisfies: is the perpendicular bisector of segment . The key distance-preserving property is, for any point on ,
Derivation hint: lies on the perpendicular bisector of , and any point on a perpendicular bisector is equidistant from the two endpoints.
For version (V1), the runner starting at , touching line at point , and continuing to a target , minimizes total distance when
achieved exactly when lies on the straight segment , where is the reflection of across . The optimal obeys the law of reflection: the incoming and outgoing rays make equal angles with . Derivation hint: Replace with using equation [2]; then by the triangle inequality, with equality iff on segment .
For an acute triangle , version (V2)'s minimum-perimeter inscribed triangle is the orthic triangle , whose vertices are the feet of the three altitudes. Its perimeter is
equivalently expressible as , and the orthic triangle's sides satisfy a reflection (equal-angle) condition at each touch point. Derivation hint: Each side of the orthic triangle reflects off the host triangle's wall at equal angles — the same "billiard" condition forced by [3]; Fejér's double-reflection unfolding shows this is the global minimum.
Why It Holds / Derivation
Setup. Runner at , target , both on the same side of wall line . We must pick the touch point minimizing .
Step 1. Reflect across to get . By equation [2], for every candidate on ,
Step 2. Substitute:
Step 3. and lie on opposite sides of (reflection flips the side). The triangle inequality gives
with equality iff lies on the straight segment .
Step 4. Since segment crosses exactly once, there is a unique optimal touch point , and the minimum length is , matching equation [3].
Step 5 (equal-angle condition). Because equals (reflection preserves angles) and are collinear, the angle of incidence equals the angle of reflection.
∎Step 1 — fix one vertex. Temporarily fix on side . We optimize the broken path .
Step 2 — double reflection. Reflect across line to get , and across line to get . By equation [2], the path length equals , which for fixed -freedom straightens to — a straight segment.
Step 3 — the key length identity. Both reflections fix distance to vertex , so , and the angle (twice the angle at , independent of ). Hence
Derivation hint: is isosceles with apex angle and equal legs ; drop the apex bisector to get the base .
Step 4 — minimize over . Since is constant, is smallest when is smallest, i.e. when . So is the foot of the altitude from .
Step 5 — symmetry closes the loop. Repeating for and forces to be the other two altitude feet. The simultaneous minimizer is exactly the orthic triangle, and at each side the equal-angle (reflection) law from [3] holds.
∎The orthic-triangle result holds only for acute triangles. If is right or obtuse, the minimum "inscribed triangle" degenerates: the infimum perimeter approaches (twice the shortest altitude) but is not attained by a genuine triangle. Always check the triangle is acute before invoking equation [4].
Caption: The actual bounce (gold) has the same length as the straight dashed segment , where is reflected across the wall. Equal incidence/reflection angles emerge automatically.
Heron first proved that light reflecting off a mirror takes the shortest path, establishing the equal-angle law via a geometric (reflection) argument — the exact tool that solves the running-around-the-house problem. His insight predated the modern variational principle by over 1500 years and reframed an optics fact as a pure minimization, the conceptual seed of every "billiard path" result here.
Building Intuition
Imagine standing at inside a mirrored room; you throw a ball that must bounce off one wall and reach a friend at . The ball "wants" the straightest route. Through the mirror, the friend's image (or your image ) appears behind the wall, and the straight line of sight to that image marks where the ball must strike. Running around a triangular house is the same trick chained three times: each wall is a mirror, and "unfolding" the triangle by repeated reflection turns the zigzag lap into one straight ruler-line. The optimal touch points are simply where that ruler-line crosses each wall.
Worked Examples
Given: Runner at , water tap (target) at , both above the wall along the -axis (). Find the optimal touch point and minimal distance.
Step 1 — Reflect : .
Step 2 — Minimal length:
Step 3 — Touch point (where segment meets ). Parametrize: , . Set . Then .
Given: An acute triangular house with angles , , and sides , , (consistent with the Law of Sines). Find the minimal lap length (orthic-triangle perimeter) using equation [4].
Step 1 — Compute each term :
Step 2 — Sum:
Sanity check: Each wall is touched at the foot of an altitude, and at each touch the incoming/outgoing run angles are equal — the billiard condition.
Common Misconceptions
"The shortest lap touches the midpoints of the sides (the medial triangle)." False. The medial triangle has perimeter equal to half the house's perimeter and does not satisfy the equal-angle reflection law; it is generally longer than the orthic triangle. The minimizer is the orthic triangle (altitude feet), not the medial triangle (side midpoints).
Take with , , . Here , so equation [4] would yield a negative contribution from the term — a sign that no genuine minimal inscribed triangle exists. Geometrically, the altitude foot from the obtuse vertex lands outside the triangle, so the orthic "triangle" is not inscribed. The infimum perimeter slides toward (twice the shortest altitude) but is never realized; the optimal "path" degenerates to a back-and-forth along that altitude.
Connections
This problem is the elementary face of Billiard Dynamics: a frictionless ball in a triangular table follows exactly the equal-angle reflections derived here, and the orthic triangle is the unique closed billiard orbit of period 3 in an acute triangle. It rests on the Triangle Inequality (the engine of every shortest-path argument) and on Reflections and Isometries. The minimization-by-unfolding is the discrete cousin of Fermat's Principle of Least Time in optics and a baby version of the Calculus of Variations. The altitude feet tie it to The Orthocenter and Altitudes.
Real-World Applications
Historical Development
Subtleties & Limits
The acute case is fully solved, but the general triangular billiard problem is famously hard: it is unknown whether every obtuse (and in particular every irrational-angle) triangle admits a closed billiard orbit at all. While right and obtuse triangles defeat the orthic-triangle minimizer, the existence of any periodic running path that touches the walls and returns is open for large classes of obtuse triangles — a striking gap given how elementary the acute case appears. The "shortest lap" thus transitions from a clean high-school exercise to an unresolved frontier of dynamical systems the moment the largest angle exceeds .
See Also
Billiard Dynamics · The Orthocenter and Altitudes · Reflections and Isometries · Fermat's Principle of Least Time · Triangle Inequality · Calculus of Variations · Geometric Optimization