The Square Cross-over Part (a)

Three points are chosen at random along the square's outline. These points are then combined to form a triangle (or a straight line if all 3 points are chosen on the same side). What is the probability that the center of the square will be contained within a triangle formed by these 3 points?

1 3 \frac13 2 3 \frac23 1 4 \frac14 3 7 \frac37 1 2 \frac12 3 4 \frac34

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.

6 solutions

Vladimir Smith
Dec 4, 2015

In the diagram above, we have chosen our first two points: x x and y y . We have also drawn the two limits to triangles – we have drawn the right most, and left most possible triangles with these values of x x and y y . Thus we have shown the total distance in which our third point could be in. This distance is A B AB . By rotational symmetry, we can prove that X Y XY is equal to A B AB . Thus we have that for any given values of x x and y y where x > y x>y , the total distance in which a third point can be chosen is the shortest distance between x x and y y along the perimeter of the square.

In this diagram, we have placed the first point of our triangle - point x x , randomly on the perimeter of the square and drawn in point k k , which is the point directly opposite of point x x . Let the perimeter of the square be equal to p p .

Notice that the closer we put point y y to k k , the shortest distance between x x and y y along the perimeter of the square increases linearly to a maximum of p 2 \frac { p }{ 2 } . This is the same no matter if we let point y y approach point k k clockwise or counter-clockwise from point c c . The minimum distance from x x to y y possible is obviously 0 0 .

Thus the mean shortest distance from x x to y y along the perimeter of the square, is the average of 0 0 and p 2 \frac { p }{ 2 } , which is equal to p 4 \frac { p }{ 4 } . Dividing this by p p we get our answer: p 4 p = 1 4 \frac { p }{ 4p } =\frac { 1 }{ 4 }

after reading your proof, I was wondering how could this be a level 2 problem!! Thanks for a clear solution...

Vikram Kumar Vuyyuru - 5 years, 6 months ago

what does " linearity of expectation" mean?

avn bha - 5 years, 6 months ago

Log in to reply

@avn bha See here: Linearity of Expectation

Eli Ross Staff - 5 years, 6 months ago

I am afraid that I am missing the obvious and if I post this, it will be out there forever, but here goes: If x = . 9 p x= .9p and y = . 1 p y= .1p , the range for the third point is . 8 p .8p ? I don't think using x y x-y is correct when it is greater than 1 2 \frac{1}{2}

Addendum: I did it in a far less elegant way and got the same probability so, as I feared, I am likely missing some symmetry aspect in your solution.

David Ritterskamp - 5 years, 6 months ago

Log in to reply

Good job on noticing, and thank you for pointing out this mistake. I have now edited the solution so that it considers when x y x-y is greater than 1 2 \frac { 1 }{ 2 }

Vladimir Smith - 5 years, 6 months ago

Oh my god... I answered 3/4 because all the time I had been working out the probability of the triangle NOT containing the centre :'(

Davy Ker - 4 years, 11 months ago
Eli Ross Staff
Dec 3, 2015

This is a very nice problem; thanks for sharing it!

If you work this out for a circle, you'll get the same answer. In fact, if you do this for any regular polygon with an even number of sides, you'll get the same answer. Can anyone show this?

I know the solution for the square version only.

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

You should post that solution, and then see if anyone is able to generalize it to other shapes!

This is one of the best parts of being part of a problem-solving community :)

Eli Ross Staff - 5 years, 6 months ago

Log in to reply

The solution is quite long and it might take me some time to type it out.

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

@Yusuf All-Husaini No problem! I'm sure someone will post one soon.

Eli Ross Staff - 5 years, 6 months ago

Log in to reply

@Eli Ross BTW are you sure about the circle having the same probability as the square?

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

@Yusuf All-Husaini Yes. Have you tried working this out?

Eli Ross Staff - 5 years, 6 months ago

Log in to reply

@Eli Ross Yes I am working on it right now and what I find is that in order for a triangle to contain the center in the circle, 2 of the points must be within 180 degrees of each other from the center and the opposite side to the third point must lie between these 2 points along the circumference. I found that the probability of this happening to be 1/2.

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

@Yusuf All-Husaini Consider the first point, and draw the diameter that passes through it.

The other two points must lie on opposite sides of this diameter, which happens with probability 1 2 . \frac{1}{2}.

However, there are further restrictions to make sure the center is contained. Specifically, the line connecting the "other two points" must be on the "opposite" side of the center from the first point. What is the probability of this happening?

Eli Ross Staff - 5 years, 6 months ago

Log in to reply

@Eli Ross Ah I didn't notice that way of looking at it. Thanks.

BTW are you absolutely sure this applies to every regular polygon?

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

@Yusuf All-Husaini It applies to regular polygons with an even number of sides.

When there are an odd number of sides, the "diameter" doesn't create the same symmetry. For example, the problem with the triangle is much trickier. See here: https://brilliant.org/problems/triangles-are-probably-fun/

Eli Ross Staff - 5 years, 6 months ago

Log in to reply

@Eli Ross I was talking about problems where points are chosen along the outline of the shape only.

yusuf all-husaini - 5 years, 6 months ago

Log in to reply

@Yusuf All-Husaini Ah yes, I didn't read that problem carefully enough. Still, I am only sure of this for even n n -gons.

Eli Ross Staff - 5 years, 6 months ago
Jonn Jonsen
Dec 9, 2015

For each point the two remaining points must reside on opposite half-perimeters determined by the diameter defined by the first point. This condition is always true for at least one of three random points. The probability of the condition being true for the remaining two random points is 1/2 each giving a total probability of 1/4.

Your method is wrong but with a slight modification you'll get the correct answer. You are over counting as well under counting here. See the image in the link below: ( sorry i don't know how to upload one)

http://drawisland.com/upload/14499259287769.png

You are over-counting by counting triangles like ABC satisfying your criteria and you are under-counting by considering points A and X to be the same.

You can resolve this by considering triangles like XBC instead of ABC when such a situation arises.

Eyan Noronha - 5 years, 6 months ago

Log in to reply

You are right, thanks. I took the position of the third point as independent of the second point, which it is not under the condition I used.

Jonn Jonsen - 5 years, 6 months ago

For any two points on the square (side s) of minimal distance x from each other, with distance measured along the outline of the square. Due to symmetry, I find the range for the third point to be also x. The probability for containing the center would be x/4s if the third point is randomly chosen along the outline of the square. We integrate x/4s over the whole possible values of x = 0 to 2s. = (integral [from x = 0 to 2s] of (x/4s) dx) / 2s = 1/4

Can you explain why "Due to symmetry, I find the range for the third point to be also x"?

In fact, you have made a key observation that allows you to deduce the answer as 1/4, without having to do any integration.

Calvin Lin Staff - 5 years, 6 months ago
Darren Tan
Dec 18, 2015

I got the answer by using the following: 1 - ( (0.5)^3 * 3! ) = 0.25

Steve Brown
Dec 12, 2015

I got this right by relying on intuitive maths and Vladimir Smith's solution kind of summed up my gut instinct.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...