0

This would be perhaps a good brain teaser at the elementary levels of Geometry, and I am trying to get my head around it.

How can we space exactly six points anywhere along the boundary or inside of a square, so that each pair of points are equally spaced away from each other?

I started by placing the four points on the four corners, and fifth one in the middle, but obviously that was to no avail. I thought to space them along two-thirds distance away along each edge, starting with the first at a vertex, but then realized that because of crossover along the edges, I am again not getting the same distance.

After reading the answers below, I realize that the problem in my head was not written in the intended manner. What I had on my mind was

If six points are placed either along the boundary or inside of a unit square, such that the minimum distance of each of these points from the remaining five points has the same value, then what is the numerical value of this distance?

I realize that my previous post was quite misleading. Does this make better sense, and is there a way out for it?

  • It has nothing to do with the square. Suppose you can find six distinct points in the plane with all pairwise distances the same. Then that figure is bounded, and so could be placed in a sufficiently largee square. You could have all six points inside the square, which you sid was legal. – coffeemath Oct 13 '21 at 04:40
  • Consider the incircle of the square. – Hussain-Alqatari Oct 13 '21 at 06:33
  • This is one of Hugo Steinhaus's famous problems or its obvious variation. – Wlod AA Oct 14 '21 at 03:35

2 Answers2

1

Alternative approach to prove impossibility.

Assumed that all 6 points are required to be distinct from each other.

Any $3$ points that are equidistant from each other must form an equilateral triangle.

Therefore, without loss of generality, points $P_1, P_2, P_3$ form an equilateral triangle.

However, since $P_4$ is also to be this same distance from both $P_1$ and $P_2$, points $P_1, P_2, P_4$ must also form an equilateral triangle.

Similarly, points $P_1, P_2, P_5$ must also form an equilateral triangle.

This is impossible because:

given points $P_1, P_2$, in order for a third point to form an equilateral triangle with $P_1$ and $P_2$ it must be on the perpendicular bisector of $P_1, P_2$, with an altitude $= \frac{\sqrt{3}}{2} \times $ the length of $\overline {P_1, P_2}$.

Therefore, there are only two distinct vertices possible that can be used to form an equilateral triangle with $P_1, P_2$.

Therefore you can not have $P_1, P_2, P_3$ and $P_1, P_2, P_4$ and $P_1, P_2, P_5$ all equilateral triangles.

Edit
The impossibility could have been proven, without consideration of point $P_5$, by noting that if $\overline{P_1P_2}$ has length $= 1$ unit, then, per the above analysis, $\overline{P_3P_4}$ must have a lenth of $\sqrt{3}$ units, which is not equal to $1$ unit.

user2661923
  • 35,619
  • 3
  • 17
  • 39
0

The only possible case where this can happen is if the points are all a distance zero from each other. Otherwise, the set of points would, without loss of generality, be a unit-embedding of $K_6$ (the connected graph on $6$ vertices) into the plane $\mathbb{R}^2$, which is impossible, since the smallest Euclidean dimension that $K_6$ can be embedded into is $5$. See here for a more general and detailed proof of this claim.

C Squared
  • 3,648
  • 1
  • 9
  • 32
  • The only reason that I don't upvote your answer is because I am totally ignorant in this area. Therefore, I don't understand your answer. However, I do strongly suspect that your answer is superior to mine. – user2661923 Oct 13 '21 at 22:11