A "Random" Triangle

Geometry Level 4

Two side lengths of a triangle are randomly chosen uniformly from ( 0 , 1 ) . (0,1). Then, the third side length is randomly chosen uniformly from the interval of side lengths that form a valid triangle.

Is this triangle more likely to be acute or obtuse?


Bonus: Find the exact probabilities.

Obtuse Equally likely Acute

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Aug 21, 2017

If the first two sides are x , y x,y , chosen uniformly and independently from ( 0 , 1 ) (0,1) , then the third side z z must satisfy the inequalities x y + z , y x + z , z x + y x \le y+z, y \le x+z, z \le x+y , and hence we must have x y z x + y |x-y| \le z \le x+y Since the cosines of the angles X , Y , Z X,Y,Z are cos X = y 2 + z 2 x 2 2 y z cos Y = x 2 + z 2 y 2 2 x z cos Z = x 2 + y 2 z 2 2 x y \cos X \; = \; \frac{y^2+z^2-x^2}{2yz} \hspace{1cm} \cos Y \; = \; \frac{x^2 + z^2 - y^2}{2xz} \hspace{1cm} \cos Z \; = \; \frac{x^2 + y^2 - z^2}{2xy} the triangle will be acute provided that x 2 y 2 + z 2 , y 2 x 2 + z 2 , z 2 x 2 + y 2 x^2 \le y^2 + z^2, y^2 \le x^2 + z^2, z^2 \le x^2 + y^2 , and so we must have x 2 y 2 z 2 x 2 + y 2 |x^2 - y^2| \; \le \; z^2 \; \le \; x^2 + y^2 Since x 2 y 2 ( x y ) 2 |x^2 - y^2| \ge (x-y)^2 and x 2 + y 2 ( x + y ) 2 x^2 + y^2 \le (x+y)^2 , we deduce that the probability that the triangle is acute, given that the first two sides are x , y x,y , is x 2 + y 2 x 2 y 2 x + y x y \frac{\sqrt{x^2+y^2} - \sqrt{|x^2 - y^2|}}{x+y - |x-y|} and hence the probability of an acute-angled triangle is therefore (putting y = x u y = xu ) in the last line): p = 0 1 d x 0 1 d y x 2 + y 2 x 2 y 2 x + y x y = 2 0 1 d x 0 x d y x 2 + y 2 x 2 y 2 2 y = 2 0 1 d x 0 1 d u 1 + u 2 1 u 2 2 u × x = 1 2 0 1 1 + u 2 1 u 2 u d u \begin{aligned} p & = \; \int_0^1\,dx \int_0^1\,dy \frac{\sqrt{x^2+y^2} - \sqrt{|x^2 - y^2|}}{x+y - |x-y|} \; = \; 2\int_0^1\,dx \int_0^x\,dy \frac{\sqrt{x^2 + y^2} - \sqrt{x^2 - y^2}}{2y} \\ & = \; 2\int_0^1\,dx \int_0^1\,du \frac{\sqrt{1+u^2} - \sqrt{1-u^2}}{2u} \times x \; =\; \tfrac12\int_0^1 \frac{\sqrt{1+u^2} - \sqrt{1-u^2}}{u}\,du \end{aligned} Thus (it is simplest to check this integral by differentiating back again) p = 1 2 [ 1 u 2 + 1 + u 2 + ln ( 1 + 1 u 2 ) ln ( 1 + 1 + u 2 ) ] 0 1 = 1 2 ( 2 ln ( 1 + 2 ) ) = 0.26642 p \; = \; \tfrac12\Big[-\sqrt{1-u^2} + \sqrt{1+u^2} + \ln\big(1 + \sqrt{1-u^2}\big) - \ln\big(1 + \sqrt{1+u^2}\big)\Big]_0^1 \; = \; \tfrac12\big(\sqrt{2} - \ln(1 + \sqrt{2})\big) \; = \; 0.26642 making an obtuse-angled triangle more likely.

A quick Monte Carlo test in Excel, looking at 1000 randomly generated triangles at a time, consistently gives a proportion of roughly 26% of acute-angled triangles.

How is the usage of y = x u y=xu justified?

Atomsky Jahid - 3 years, 9 months ago

Log in to reply

I am not quite sure how to reply - because it works? It is a valid substitution for any x > 0 x > 0 , and the case x = 0 x=0 is of measure zero, and so can be ignored. Using it cuts a double integral very quickly down to a single integral which can be evaluated.

Mark Hennings - 3 years, 9 months ago

Log in to reply

The fact that d y = x d u dy=xdu is due to the independence of x x and y y . Am I right?

Atomsky Jahid - 3 years, 9 months ago

Log in to reply

@Atomsky Jahid I have written the integral as an iterated integral. I am performing the integral with respect to y y , and then integrating the result with respect to x x . While performing the y y integral, x x can be regarded as a constant, so a substitution like y = x u y = xu is fine.

Mark Hennings - 3 years, 9 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...