Bugz Podder
Posted on Tuesday, 14 October, 2003 - 02:53 am:

triangle ABC has circumradius 1
any point M on the circumcircle, MA*MB*MC< =2
show it is equilateral
Bilen Ahmet
Posted on Tuesday, 14 October, 2003 - 02:50 pm:

this can probably be done by using the expression for a general point on a circumcircle, and multiplying the generalised expressions for MA,MB,MC, perhaps using vectors.

To prove it, you should end with a resultant equation that shows for all other triangles the product is greater than 2.

You may also be able to use RAA.
David Loeffler
Posted on Tuesday, 14 October, 2003 - 04:17 pm:

> To prove it, you should end with a resultant equation that shows for all other triangles the product is greater than 2.

I don't think it can be that simple, since if M = A, then the expression is zero, and thus < 2. The question is whether or not there is some point on the circumcircle for which MA*MB*MC > 2. You can attack this either by guessing a suitable point and showing that if A is not equilateral it has the required property, or by finding explicitly the point at which MA*MB*MC is maximised.

David
Rachel Burns
Posted on Tuesday, 14 October, 2003 - 05:04 pm:

You could try using the fact that if M lies between A and C say on the circumcentre that MB= MA+MC ...
David Loeffler
Posted on Tuesday, 14 October, 2003 - 05:09 pm:

I think if you assume a>bc and take M to be the midpoint of the arc between B and C on the opposite side from A, then if Δ ABC is not equilateral, then it all works.

In fact, if I am not mistaken in my calculations, for this M we have MA×MB×MC=8 sin2 A/2cos(B-C)/2=4sinA/2(cosB+cosC).

Since A>BC, we have A>π/3, so sinA/2>1/2; thus it suffices to show that cosB+cosC>1.

This looks like it ought to be easy, but I haven't found a sensible proof. If we use the formula for cosine in terms of the side lengths we must show

b( a2 + c2 - b2 )+c( a2 + b2 - c2 )>2abc

Expanding everything out we must show

a2 b+b c2 + a2 c+ b2 c>2abc+ b3 + c3

We know that a2 b> b3 and b c2 > c3 so it suffices to show that

a2 c+ b2 c>2abc

which is just c(a-b )2 >0 and is consequently true.

So we are done. (Phew!)

I'll leave you the fun of checking that MA * MB * MC is never > 2 for an equilateral triangle (use Ptolemy's theorem).

David

David Loeffler
Posted on Tuesday, 14 October, 2003 - 05:11 pm:

Rachel, what you say is only true for an equilateral triangle (for a general triangle the corresponding statement is that a MA + c MC = b MB, which is a restatement of Ptolemy's theorem).

David
Bugz Podder
Posted on Tuesday, 14 October, 2003 - 11:43 pm:

A brute force solution, although solves the problem, is uninteresting...

but then again, i do not see any elegant way of attacking this problem
David Loeffler
Posted on Tuesday, 14 October, 2003 - 11:51 pm:

Are you criticising my solution?

No, I agree, it is not particularly inspiring. If I get round to it I might have another think and see if I can work out the real reason why it is true - finding the point Mmax which maximises the product in the statement of the problem could be illuminating.

David
Bugz Podder
Posted on Wednesday, 15 October, 2003 - 04:07 am:

at least you got a solution ;)