A regular hexagon does not look like a circle.
But if we keep adding sides—12, 24, 48, 96—the shape begins to look almost round.
That raises a deeper question than it first seems to:
What is actually changing when a polygon gets more sides?
This is not only a geometry question. It is one of the first clear examples of a bigger mathematical habit: replace a hard shape with a simpler one, then improve the approximation step by step.
In the previous article, we saw Archimedes trap π between inner and outer polygon estimates. Here we isolate the mechanism inside that method: why does increasing the number of sides make the approximation tighter?
Quick Answer
A polygon is a shape made of straight sides. A regular polygon is a polygon whose sides and angles are all equal.
When a regular polygon gets more sides, each side becomes shorter. Shorter straight pieces can follow a curved shape more closely than longer straight pieces can.
So the polygon does not suddenly turn into a circle. Instead, its straight edge becomes a better and better approximation of the circle.
More sides do not create a curve. They make the straight-line model fit the curve more closely.
Start with One Word: Approximation
An approximation is something that is not exact, but is close enough to help us understand or calculate a harder thing.
That is the main idea in this article.
A circle has a smooth curved boundary. A polygon has straight sides. So a polygon is not a circle. But it can still help us understand one.
That is exactly why Archimedes used polygons when he studied π.
Why a Six-Sided Polygon Misses the Circle
Imagine a circle with a regular hexagon drawn inside it. The corners touch the circle, but each side cuts straight across a curved arc.
The total distance around the polygon is called its perimeter. Because each side is a shortcut across a little piece of the curve, the hexagon's perimeter is shorter than the circle's circumference—the distance around the circle.
Now imagine a regular polygon drawn outside the circle so that each side touches it. This outer polygon goes around the circle from the outside, so its perimeter is too large.
Figure 1. One polygon gives a lower estimate. The other gives an upper estimate.
This gives us a powerful structure:
inside perimeter < circle circumference < outside perimeter
Even before we know the exact answer, we already know where the answer must live.
Why More Sides Help
Suppose we double the number of sides.
6 → 12 → 24 → 48 → 96 → 192
Now each side covers a smaller part of the circle.
A long straight segment cannot follow a curve very well. A short straight segment can follow a small piece of the curve much more closely. That is why the shape begins to look rounder.
Figure 2. The polygon stays made of straight sides, but the mismatch becomes smaller.
What Is the “Mismatch”?
The important word here is mismatch: the difference between the straight-sided model and the curved shape we are trying to understand.
For the inside polygon, the mismatch shows up because the perimeter is too short. For the outside polygon, it shows up because the perimeter is too long.
The difference between those two estimates is the gap.
gap = upper estimate − lower estimate
As the number of sides increases, the lower estimate rises, the upper estimate falls, and the gap shrinks.
This is the clearest sign that the approximation is improving.
A Modern Numerical Reconstruction
Archimedes did not write modern trigonometric formulas. But we can use modern tools to reconstruct the same geometric idea.
For a circle of radius 1, the full polygon perimeters approach the circle circumference 2π. To keep the numbers focused on π itself, we can divide those perimeters by 2. One convenient modern form is:
lower bound for π = n sin(π/n)
upper bound for π = n tan(π/n)
Here, n is the number of sides. The function sin is sine, and tan is tangent.
You do not need to master the formulas yet. The important idea is simpler: the inner polygon gives a value below π, the outer polygon gives a value above π, and increasing n squeezes those two bounds closer together.
| Sides | Lower | Upper | Gap |
|---|---|---|---|
| 6 | 3.000000 | 3.464102 | 0.464102 |
| 12 | 3.105829 | 3.215390 | 0.109562 |
| 24 | 3.132629 | 3.159660 | 0.027031 |
| 48 | 3.139350 | 3.146086 | 0.006736 |
| 96 | 3.141032 | 3.142715 | 0.001683 |
| 192 | 3.141452 | 3.141873 | 0.000421 |
The pattern matters more than the decimals.
The lower estimate climbs upward. The upper estimate moves downward. The gap becomes much smaller.
That is what “closer to a circle” really means in a mathematical sense.
Closer Is Not Yet the Same Question as Equal
At 96 sides, the polygon may look almost circular. At 192 sides, the mismatch is smaller still.
But notice what we have actually established in this article: the approximation improves and the gap shrinks.
That is not quite the same question as asking whether the polygon ever becomes exactly identical to a circle.
Keep those two questions separate:
- How does the error shrink as we refine the polygon?
- Does getting arbitrarily close eventually mean becoming exactly the same object?
The first question belongs here. The second is the next conceptual step.
Try It — Watch the Gap Shrink
If you want to test the pattern yourself, Python can repeat the calculation for several values of n.
Before you run it, make one prediction: as n grows, which two columns should move toward each other, and which column should move toward zero?
import math
for n in [6, 12, 24, 48, 96, 192, 384]:
lower = n * math.sin(math.pi / n)
upper = n * math.tan(math.pi / n)
gap = upper - lower
print(n, lower, upper, gap)The for line tells Python to repeat the same calculation for each number in the list.
If you add more values such as 768 or 1536, the same pattern continues: the lower and upper estimates squeeze closer together.
Why This Article Matters
This article is not only about polygons and circles.
It introduces a larger way of thinking:
hard shape → simpler model → controlled improvement → smaller error
That pattern appears everywhere: in geometry, numerical methods, physics, engineering, and computer science.
A concrete modern descendant appears in CAD, computer graphics, and numerical simulation: a curved boundary can be represented by many short segments or small elements. The details are more advanced, but the reusable habit is familiar — refine the model, then check whether the mismatch becomes smaller.
We often cannot reach the difficult thing all at once. But we can surround it, measure the error, and refine our model.
Words to Keep
polygon
A closed shape made of straight sides.
regular polygon
A polygon whose sides and angles are all equal.
circumference
The distance around a circle.
perimeter
The total distance around a polygon.
approximation
A value or model that is not exact but is close enough to help.
gap
The difference between the upper and lower estimates.
limit
A later mathematical idea for describing what a process approaches as it continues indefinitely.
One Sentence to Keep
More sides make the straight-line model fit the circle more closely: the lower and upper bounds squeeze together and the measurable gap shrinks.
What Should We Ask Next?
We now know how to make the approximation better.
But one question is still deliberately unanswered:
If the mismatch can become as small as we like, does the polygon ever actually become a circle?
That is the next step, because mathematics must distinguish very carefully between approaching something and being that thing.
Previous: How Did Archimedes Calculate Pi Without a Calculator?
Next: Can a Polygon Ever Actually Become a Circle?
Sources & Further Reading
- Archimedes, Measurement of a Circle, in The Works of Archimedes — edited and translated by Thomas L. Heath, Cambridge University Press.
- David Weisbart, “Modernizing Archimedes’ Construction of π” — Mathematics, 8(12), 2204, 2020.
- “A History of Pi” — MacTutor History of Mathematics, University of St Andrews.
Historical note: the decimal values, trigonometric formulas, and Python code in this article are modern tools used to explore the geometric idea. They are not presented as Archimedes’ original notation or calculation method.