Dear Uncle Colin,

@CmonMattTHINK unearthed the challenge to prove that:

tan(311π)+4sin(211π)=11. Wolfram Alpha says it’s true, but I can barely get started on the proof and I’m worried no-one will like me.

Grr, Really Obnoxious Trigonometry Has Evidently Nice Denouement – Is Everything Common Knowledge?

Hi, GROTHENDIECK – don’t worry, there is practically no correlation between one’s ability to do trig proofs and one’s ability to make friends.

This particular problem is a Big Blue Meany. It took me a couple of days to get anywhere near a solution, too, and even afterwards I’m not 100% happy with it. (There’s a paper on it here, though, with some alternatives if you don’t like mine.)

My working approach is to turn everything into complex form, letting z=e111π for convenience:

tan(311π)z3z3i(z3+z3)

4sin(211π)2z2z2i

Doesn’t look much nicer, does it? OK, let’s add the fractions to get :

z3z3i(z3+z3)+2z2z2i=z3z3+2(z2z2)((z3+z3)i(z3+z3)

It gets nicer, honestly!

z3z3+2(z5+zz1z5)i(z3+z3)

This is the same as

2z5+z3+2z2z1z32z5i(z3+z3)

Square and square root:

±4z10+4z87z6+4z4+4z218+4z2+4z47z6+4z8+4z10(z6+2+z6)

No, wait, come back! Remember the first rule of roots of unity club? No? Well, you should! The eleventh roots of unity, that is to say z10+z8+z6+z4+z2+1+z2+z4+z6+z8+z10=0. Turns out, we’ve got nearly all of those – although we’ll need to add and subtract 4z6, 4z6 and 4 to make it work properly:

±[4z10+4z8+4z6+4z4+4z2+4+4z2+4z4+4z6+4z8+4z10]11z62211z6(z6+2+z6)

The big bracket is zero:

±11z62211z6(z6+2+z6)

… which is clearly ±11. However, since both of our original trig function are in the first quadrant, our answer must be the positive root.

Incidentally, GROTHENDIECK? Asking people to solve that sort of puzzle is a sure-fire way to lose friends.

-- Uncle Colin