Weitzenböck's Inequality An Elegant Dialogue Between Side Lengths and Area
For any triangle, the sum of squared side lengths is never less than $4\sqrt{3}$ times the area. The equilateral triangle attains equality exactly—this is no coincidence, but a perfect geometric inevitability.
Part One
Discovering the Inequality: Finding Patterns in Concrete Numbers
① The Problem: How Much Area Can Three Sides "Support"?
Suppose you have three sticks of lengths $a$, $b$, and $c$. Connect them end-to-end to form a triangle.
The area $S$ of this triangle is completely determined by the three sides (Heron's formula), but today we ask a different question:
💡 Core Question
Between $a^2 + b^2 + c^2$ (the sum of squared side lengths) and the area $S$, does there exist an inequality that holds for every triangle?
To answer this, we start with the most "fair" triangle—the equilateral triangle.
🤔 Why start with the equilateral triangle?
The equilateral triangle is the "most symmetric" triangle. Any pattern that holds for all triangles must hold for the equilateral triangle—and it is often the boundary case. Let's compute it first and treat it as our "baseline."
② Exploration: The Equilateral Triangle—Birth of the Baseline
Let the side length of the equilateral triangle be $t$.
1
Sum of squared sides:$a^2 + b^2 + c^2 = t^2 + t^2 + t^2 = 3t^2$
2
Area: Equilateral triangle area formula $S = \frac{\sqrt{3}}{4}t^2$
💡 A Pattern Emerges
Equilateral triangle: equality; $3$-$4$-$5$ triangle: left side is larger. Try a few more—the left side is never smaller than the right!
Verify with several different triangles (all side lengths form valid triangles):
Sides $(a,b,c)$
$a^2+b^2+c^2$
$4\sqrt{3}S$
Comparison
$(1,1,1)$
$3$
$3$
Equal ✓
$(3,4,5)$
$50$
$41.57$
Larger
$(5,5,6)$
$86$
$72.75$
Larger
$(2,3,4)$
$29$
$19.74$
Larger
$(7,8,9)$
$194$
$159.35$
Larger
✅ Weitzenböck's Inequality (Weitzenböck, 1919)
For any triangle with sides $a,b,c$ and area $S$, we have: $a^2 + b^2 + c^2 \geq 4\sqrt{3}S$
Equality holds if and only if the triangle is equilateral.
⚠️ The Question Now
The numbers in the table are merely "evidence," not "proof." We need a rigorous geometric proof that holds for all triangles. Next, we accomplish this by constructing equilateral triangles on each side.
Part Two
Geometric Proof: "Weighing" the Original Triangle with Equilateral Triangle Areas
④ Discovery Two: Translating into the Language of "Equilateral Triangle Areas"
Note a key fact:
💡 Key Fact
$\frac{\sqrt{3}}{4}a^2$ is precisely the area of an equilateral triangle with side length $a$. Denote it as $S_a$.
Multiply both sides of the original inequality by $\frac{\sqrt{3}}{4}$:
The inequality transforms into a purely geometric statement:
💡 Geometric Translation
On each of the three sides, construct an equilateral triangle outward. The sum of their areas is at least 3 times the area of the original triangle.
Three Equilateral Triangles: Construct equilateral triangles outward on BC, CA, and AB Blue Sa + Purple Sb + Cyan Sc ≥ 3 × Orange S
Now our task becomes: prove $S_a + S_b + S_c \geq 3S$. This still doesn't look obvious—what relationship do the three large "outer" equilateral triangles have with the small triangle inside?
⚠️ The Problem
We need to find a way to connect the original triangle $S$ with the outer equilateral triangles. The answer lies hidden in a magical point.
⑤ Discovery Three: The Fermat Point—The Magic of $120^\circ$
First, assume that every interior angle of the triangle is less than $120^\circ$ (the obtuse case will be handled separately later).
💡 Fermat Point (Fermat–Torricelli Point)
Inside the triangle there exists a unique point $F$ such that the lines connecting $F$ to the three vertices are pairwise at $120^\circ$:
$\angle AFB = \angle BFC = \angle CFA = 120^\circ$
The Fermat point divides triangle $ABC$ into three small triangles: $\triangle BFC$, $\triangle CFA$, $\triangle AFB$. Their areas sum exactly to $S$.
Fermat Point $F$: Lines to the three vertices are pairwise at $120^\circ$, dividing the triangle into three parts
Now focus on one sub-triangle—$\triangle BFC$—and see how it relates to the equilateral triangle $S_a$.
⑥ Core Insight: Comparing $S_a$ with $3 \cdot \text{Area}(\triangle BFC)$
In $\triangle BFC$:
Side $BC = a$ (a side of the original triangle)
$\angle BFC = 120^\circ$ (by definition of the Fermat point)
$BF$ and $CF$ are the distances from $F$ to $B$ and $C$
Focus on $\triangle BFC$: Compare the blue equilateral triangle $S_a$ with the orange sub-triangle Area(BFC)
Starting from $\triangle BFC$, proceed along two paths:
A
Sub-triangle Area (using $BF, CF$ and included angle $120^\circ$)
$\text{Area}(\triangle BFC) = \frac{1}{2} \cdot BF \cdot CF \cdot \sin 120^\circ = \frac{\sqrt{3}}{4} \cdot BF \cdot CF$
B
Equilateral Triangle $S_a$ (expanding $a^2$ via the Law of Cosines)
In $\triangle BFC$, by the Law of Cosines ($\cos 120^\circ = -\frac{1}{2}$):
$a^2 = BF^2 + CF^2 - 2 \cdot BF \cdot CF \cdot \cos 120^\circ = BF^2 + CF^2 + BF \cdot CF$
So $S_a = \frac{\sqrt{3}}{4}a^2 = \frac{\sqrt{3}}{4}(BF^2 + CF^2 + BF \cdot CF)$
💡 A Major Pivot (Key Insight)
Don't directly compare $S_a$ and $\text{Area}(\triangle BFC)$—they have different dimensions. What we need to compare is $S_a$ with $3 \cdot \text{Area}(\triangle BFC)$.
✅ First Inequality
$S_a \;\geq\; 3 \cdot \text{Area}(\triangle BFC)$, with equality if and only if $BF = CF$.
🤔 Why is this the key?
$\frac{\sqrt{3}}{4}(BF-CF)^2$ is a square term—it is always non-negative, and the source of non-negativity is so clean: the perfect square formula. This means the excess of $S_a$ over $3\cdot\text{Area}(\triangle BFC)$ is exactly $\frac{\sqrt{3}}{4}$ times the square of the "degree of asymmetry" between $BF$ and $CF$. The closer $BF$ and $CF$ are, the smaller the gap; when $BF=CF$, the gap vanishes.
⑧ Summing the Three—Mission Accomplished
Apply the exact same derivation to the other two sub-triangles:
$S_b \;\geq\; 3 \cdot \text{Area}(\triangle CFA)$, with equality when $CF = AF$
$S_c \;\geq\; 3 \cdot \text{Area}(\triangle AFB)$, with equality when $AF = BF$
Summing the Three: Each inequality contributes $3\times$ the sub-triangle area, summing to $3S$
Adding the three inequalities, the right side becomes $3\cdot[\text{Area}(BFC)+\text{Area}(CFA)+\text{Area}(AFB)] = 3S$:
✅ Proof Complete
Starting from the Fermat point, we compared three independent "sub-triangle vs. equilateral triangle" pairs and summed them up—neatly arriving at Weitzenböck's inequality.
⑨ The Equality Condition—Why Only the Equilateral Triangle Attains Equality?
Look back at the equality conditions of the three inequalities:
$S_c = 3\cdot\text{Area}(\triangle AFB) \iff AF = BF$
The three inequalities attain equality simultaneously if and only if:
$BF = CF = AF$
The distances from $F$ to the three vertices are equal—this means $F$ is both the circumcenter (equidistant from vertices) and the Fermat point (all angles $120^\circ$). For a triangle's circumcenter and Fermat point to coincide, there is only one possibility:
✅ Conclusion
Weitzenböck's inequality attains equality $\iff$ the triangle is equilateral.
This is perfectly consistent with our "baseline" from the beginning—the equilateral triangle made a perfect "prediction."
There is also aesthetic beauty in the figure: when the triangle is equilateral, the Fermat point = center = centroid = incenter = circumcenter—everything converges to one point. The three "asymmetry" measures—$(BF-CF)^2$, $(CF-AF)^2$, $(AF-BF)^2$—all vanish.
Part Three
Extensions and Review: Exploration, Generalization, Full Map
⑩ Interactive Calculator: Verify the Inequality Yourself
Adjust the sliders to vary the three side lengths and observe the gap between $a^2+b^2+c^2$ and $4\sqrt{3}S$. The gap is always non-negative—try to make it zero.
⚠️ Note on Obtuse Triangles
The proof above assumes every angle is less than $120^\circ$. When one angle $\geq 120^\circ$, the Fermat point degenerates to that obtuse vertex. The corresponding comparison degenerates to a trivial inequality, while the other two comparisons become even "looser"—the sum inequality still holds strictly. Thus Weitzenböck's inequality holds for all triangles without exception.
⑪ Generalization Bonus: The Hadwiger–Finsler Inequality
Weitzenböck's inequality has a stronger version—the Hadwiger–Finsler Inequality (1937):
🤔 Why Is It "Stronger"?
The right side of Hadwiger–Finsler has three extra squared terms $(a-b)^2+(b-c)^2+(c-a)^2$ compared to Weitzenböck. These square terms are always $\geq 0$. Thus Hadwiger–Finsler $\implies$ Weitzenböck. The equality condition is the same—equilateral triangle.
Looking back at our proof, the square term $\frac{\sqrt{3}}{4}(BF-CF)^2$ already hints at Hadwiger–Finsler: if we keep all three square terms rather than simply "proving non-negativity," we can derive the stronger form.
Weitzenböck's inequality also appears in the following areas:
Optimization: Given a fixed perimeter, the equilateral triangle maximizes area (the triangle case of the isoperimetric inequality)
Rigid Body Mechanics: Lower bound on stiffness of triangular frameworks
Combinatorial Geometry: Energy minimization in discrete point sets
💡 Bonus
Weitzenböck's inequality can be seen as a discrete version of the two-dimensional isoperimetric inequality: with a fixed "sum of squared side lengths" (analogous to boundary "energy"), the equilateral triangle maximizes area—just as the circle maximizes area for a fixed perimeter. The equilateral triangle is "the circle of the polygon world."
⑫ Full Map: From Equilateral Triangle to General Proof
Review the entire derivation chain. Though the inequality appears simple, its proof weaves together multiple geometric concepts:
① Equilateral Triangle Discover the equality case: $a^2+b^2+c^2 = 4\sqrt{3}S$
② Non-Equilateral Verification Compute specific cases: inequality holds strictly
③ Translate: $S_a + S_b + S_c \geq 3S$ $\frac{\sqrt{3}}{4}a^2$ = area of equilateral triangle with side $a$
④ Fermat Point $F$ ($120^\circ$ connections) Divide into three sub-triangles $\triangle BFC, \triangle CFA, \triangle AFB$