I made a cardboard triangular tile out of 6 other triangles. Here are the triplets shown in order of CAB, BAC, CAB, BAC, CAB, BAC. The slashes show sides-C joined and the vertical lines show sides-B joined. I found the arcsines of the sine values and my best calculations show that the exterior acute angles add up to $\pi$ radians. The problem is, unless this is proven, I cannot ’know’ that the peaks all meet at a central point and that the figure is 2D and not a pyramid. What approach can I take to prove my conjecture?

/ 565 403 396 | 396 1053 1125 / 1125 675 900 | 900 1925 2125 / 2125 2107 276 | 276 493 565 \

enter image description here

  • 3
    $\begingroup$ That the sum of angles in a triangle is equal to $\pi$ is equivalent to the parallel postulate, one of the axioms of Euclidean Geometry and which need not necessarily be true for example in Hyperbolic or Spherical Geometries. See https://en.wikipedia.org/wiki/Sum_of_angles_of_a_triangle. Showing that a particular sum of numbers is approximately equal to $\pi$ can be accomplished using basic arithmetic. Showing that those numbers adds up to exactly $\pi$ is impossible in your case since the numbers are inexact approximations. $\endgroup$ – JMoravitz Jun 8 '18 at 20:33

Like @dxiv, I considered the central angles. Unlike @dxiv, I reached a positive verdict:

Your pieces should fit.

We know your pieces are perfect right triangles that fit together in pairs as follows:

enter image description here

To determine whether they fit together in one whole, we need only determine whether $\alpha+\beta+\gamma = 360^\circ$. Well,

$$\begin{align} \sin(\alpha+\beta+\gamma) &= \sin\alpha\cos\beta\cos\gamma + \cos\alpha\sin\beta\cos\gamma +\cos\alpha\cos\beta\sin\gamma - \sin\alpha\sin\beta\sin\gamma \\ \cos(\alpha+\beta+\gamma) &= \cos\alpha\cos\beta\cos\gamma - \cos\alpha\sin\beta\sin\gamma - \sin\alpha\cos\beta\sin\gamma - \sin\alpha\sin\beta\cos\gamma \end{align}$$


$$\cos\alpha = \frac{-a^2+q^2+r^2}{2qr} \qquad \cos\beta = \frac{-b^2+r^2+p^2}{2rp} \qquad \cos\gamma = \frac{-c^2+p^2+q^2}{2pq}$$ $$\sin\alpha = \frac{2|\triangle OBC|}{qr} = \frac{ad}{qr} \qquad \sin\beta = \frac{be}{rp} \qquad \sin\gamma = \frac{cf}{pq}$$

Substituting ... $$a \to 493 + 2107 \qquad b \to 675 + 1925 \qquad c \to 1053 + 403$$ $$d \to 276 \qquad e \to 900 \qquad f \to 396$$ $$p \to 1125 \qquad q \to 565 \qquad r \to 2125$$ ... Mathematica calculates (via precise integer arithmetic) ... $$\sin(\alpha+\beta+\gamma) = 0 \;\;\text{(exactly)} \qquad\qquad \cos(\alpha+\beta+\gamma) = 1\;\;\text{(exactly)}$$ ... so that (subject to the accuracy of your measurements) $\alpha+\beta+\gamma$ is definitely an integer multiple of $360^\circ$. By inspection, the integer multiplier is $1$. $\square$

| cite | improve this answer | |
  • $\begingroup$ What drawing software do you use? I could use it for a paper I'm writing on this subject. $\endgroup$ – poetasis Jun 9 '18 at 18:00
  • 1
    $\begingroup$ @poetasis: I use GeoGebra for my drawings. $\endgroup$ – Blue Jun 9 '18 at 21:44
  • $\begingroup$ @Blue Thanks for the correction. Could you please enter the expression linked in my now deleted answer in the standalone Mathematica and see what it returns. I am puzzled that WA would screw up on such a straightforward calculation. $\endgroup$ – dxiv Jun 10 '18 at 5:27
  • $\begingroup$ @dxiv: Hmmm ... As it turns out, you seem to have made a simple typo: $2125/276$ should be $2107/276$. $\endgroup$ – Blue Jun 10 '18 at 5:37
  • 1
    $\begingroup$ @Blue Silly typo, indeed. Fixed now, thanks for catching that. $\endgroup$ – dxiv Jun 10 '18 at 19:22

[ EDIT ] Thanks @Blue for catching the original typo. Corrected and reposted as CW.

According to WA the sum of the center angles equals $\,2 \pi\,$, which proves that the six triangles with the given sides can indeed be arranged as pictured around a common vertex in the plane:

$$ \arctan \frac{403}{396} + \arctan \frac{1053}{396} + \arctan \frac{675}{900} + \arctan \frac{1925}{900} + \arctan \frac{2107}{276} + \arctan \frac{493}{276} = 2 \pi $$

This is an exact calculation which can be verified by hand using the $\tan$ of sum of angles formula.

  • $\begingroup$ Yes, it looks like the peaks won't quit meet at a point. (We'll have a Mercedes star in the middle.) I've generated many of these programmatically. Still, if I find a set for which WA says the internal arctangents add up to 2$\pi$, wouldn't it still be just a better approximation? Thanks. $\endgroup$ – poetasis Jun 8 '18 at 22:26
  • $\begingroup$ @poetasis The above is an exact calculation, not an approximation. It can be verified by hand using the sum of angles $\tan$ formula. $\endgroup$ – dxiv Jun 8 '18 at 22:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.