0

I am looking for an elegant and quick proof for the formula: $ \tan(A) + \tan(60^\circ+A) - \tan(60^\circ-A) = 3\tan(3A)$ I have looked through some proofs through pure trigonometry but was hoping was something involving complex numbers or geometry.

  • 1
    Just curious: are you sure these are equal? a quick look at their graphs shows they aren't. Also, the equation is true for $A=\pi n$ and $A=\pi n-\pi/2$ (according to WolframAlpha)

    Check here: https://www.wolframalpha.com/input/?i=tan%28x%29%2Btan%2860%2Bx%29-tan%2860-x%29%3D3tan%283x%29

    – Sergio May 23 '20 at 15:44
  • @Sergio He might be using radians, if you replace $60$ with $\pi/3$, they are equal. – ZKG May 23 '20 at 15:50
  • @ZKG: no, there is a typo, $-$ instead of $+$. –  May 23 '20 at 15:51
  • Sergio I should point out that 60 is 60 degrees. Also we shouldn't really talk about the equation at A=πn−π/2 as these points are not in the domain of the function TanA. Also I am pretty sure it's -. – Thenard Rinmann May 23 '20 at 15:51
  • @ThenardRinmann If that's the case, you are using inconsistent notation. Why isn't 3 replaced with $\pi/60$ radians? You're using degrees on the LHS but radians on the RHS. – ZKG May 23 '20 at 15:53
  • I didn't think of degrees, weird using these, thanks everyone – Sergio May 23 '20 at 15:53
  • @ZKG that is what confused me! – Sergio May 23 '20 at 15:54
  • @ZKG: Don't be silly. $3$ is not the argument of a trigonometric function. – TonyK May 23 '20 at 15:54
  • @TonyK I'm sorry, I don't understand what you mean. – ZKG May 23 '20 at 15:55
  • @ZKG: In this context, $3$ doesn't mean $3^\circ$ or $3$ radians. It's just $3$. What would it even mean to write $(\pi/60$ radians$);\tan(3A)$? – TonyK May 23 '20 at 15:57
  • sorry for the inconvenience, I hope the new notation makes it clear that the angles are in degrees. – Thenard Rinmann May 23 '20 at 15:58
  • @TonyK Fair point. But, without the $^{\circ}$ symbol, I was asking why we should have interpreted the numbers on the LHS as being equivalent to $60^{\circ}$ but the numbers on the RHS are not $3^{\circ}$. – ZKG May 23 '20 at 16:00
  • @TonyK I was thinking it would be a legitimate interpretation to see the RHS as $3\tan((\pi/60)A)$. Though, you make a good point, and it probably should have been clear through context. – ZKG May 23 '20 at 16:06
  • @ZKG: Oh, I see. But that $3$ is just a multiplier of the angle $A$, not an angle itself. So I think you are wrong there too. – TonyK May 23 '20 at 16:59
  • https://math.stackexchange.com/questions/455070/proving-a-fact-tan6-circ-tan42-circ-tan12-circ-tan24-cir/455573#455573 – lab bhattacharjee May 23 '20 at 18:26
  • @labbhattacharjee: That's a completely different expression! – TonyK May 23 '20 at 22:20

3 Answers3

1

By the addition formula,

$$\text{LHS}:t+t_++t_-=t+\frac{t+\sqrt3}{1-\sqrt3t}+\frac{t-\sqrt3}{1+\sqrt3t}=3\frac{3t-t^3}{1-3t^2}$$

and

$$\text{RHS}:t_3=\frac{t+t_2}{1-t\,t_2}=\frac{t+\dfrac{2t}{1-t^2}}{1-t\dfrac{2t}{1-t^2}}=\frac{3t-t^3}{1-3t^2}.$$

I can't think of a simple geometric proof, as there would be an angle trisection.


Note that

$$1+it_3\propto(1+it)^3=1+3it-3t^2-it^3\implies t_3=\frac{3t-t^3}{1-3t^2}.$$

  • You are letting $t=\tan(A)$, is this correct? If so, I think you should explain where those identities are coming from. I assume it's a sum trig identity and a triple angle identity. – ZKG May 23 '20 at 16:11
  • @ZKG: I have slightly improved. –  May 23 '20 at 16:14
1

I'm confident that in terms of text this is not the simplest way but i consider it elegant because it uses very few first principles and can also be easily automated. So if you ever have a trig identity and don't know where to go, this will ALWAYS work as a faithful faithful last resort.

The result we want to show is

$$ \tan(A) + \tan \left(\frac{\pi}{3}+A \right) - \tan \left( \frac{\pi}{3} - A \right) = 3 \tan \left( 3A\right)$$

We can utilize Euler's formula. So we have that $\tan(x) = \frac{\sin(x)}{\cos(x)} $ and because of the formula:

$$\sin(x) = \frac{e^{ix} - e^{-ix}}{2i}$$

$$\cos(x) = \frac{e^{ix} + e^{-ix}}{2}$$

So then it naturally follows that

$$ \tan(x) = \frac{e^{ix} - e^{-ix}}{i(e^{ix} + e^{-ix})} $$

Our identity then is to verify that $$ \tan(A) + \tan \left(\frac{\pi}{3}+A \right) - \tan \left( \frac{\pi}{3} - A \right) = 3 \tan \left( 3A\right)$$

But instead it will look like:

$$ \frac{e^{iA} - e^{-iA}}{i(e^{iA} + e^{-iA})} + \frac{e^{i(\frac{\pi}{3}+A)} - e^{-i(\frac{\pi}{3}+A)}}{i(e^{i(\frac{\pi}{3}+A)} + e^{-i(\frac{\pi}{3}+A)})} -\frac{e^{i(\frac{\pi}{3}-A)} - e^{-i(\frac{\pi}{3}-A)}}{i(e^{i(\frac{\pi}{3}-A)} + e^{-i(\frac{\pi}{3}-A)})} = 3\frac{e^{i3A} - e^{-i3A}}{i(e^{i3A} + e^{-3iA})} $$

We can make some simplifications [example the i on the denominators on all the fractions cancel out]

$$ \frac{e^{iA} - e^{-iA}}{e^{iA} + e^{-iA}} + \frac{e^{i(\frac{\pi}{3}+A)} - e^{-i(\frac{\pi}{3}+A)}}{e^{i(\frac{\pi}{3}+A)} + e^{-i(\frac{\pi}{3}+A)}} -\frac{e^{i(\frac{\pi}{3}-A)} - e^{-i(\frac{\pi}{3}-A)}}{e^{i(\frac{\pi}{3}-A)} + e^{-(\frac{\pi}{3}-A)}} = 3\frac{e^{i3A} - e^{-i3A}}{e^{i3A} + e^{-3iA}} $$

And now all that's left is to combine these 3 fractions. As a quick sanity check we verify that if we combine the 3 denominators everything matches, that is we want to see that

$$ (e^{iA} + e^{-iA})(e^{i(\frac{\pi}{3}+A)} + e^{-i(\frac{\pi}{3}+A)})(e^{i(\frac{\pi}{3}-A)} + e^{-(\frac{\pi}{3}-A)}) = e^{i3A} + e^{-3iA}$$

We do some good old FOIL on the right 2 products on the left hand side

$$ (e^{iA} + e^{-iA})(e^{2\frac{i\pi}{3}}+e^{2iA} + e^{-2iA} + e^{- 2\frac{i \pi}{3}})$$

And from here we evaluate the $e^{\frac{2\pi i}{3}}$ terms. Review this if unfamiliar with roots of unity and we have

$$ (e^{iA} + e^{-iA})(\frac{-1-i\sqrt{3}}{2}+e^{2iA} + e^{-2iA} + \frac{-1+i\sqrt{3}}{2}) $$

Which yields:

$$(e^{iA} + e^{-iA})(e^{2iA} + e^{-2iA} - 1 ) = e^{3iA} + e^{iA}+e^{-iA} + e^{-3iA} - e^{iA} - e^{-iA} = e^{3iA} - e^{-3iA} $$

As desired!

So now all of that is the sanity check... to do the numerator is considerably more work, but maybe with about 15-20 minutes of continuous algebra its doable.

What's nice, is ALL we used here were: complex numbers, multiplying exponents/simplifying fractions, and one formula.

This strategy in general can be utilized to verify ANY trig identity you want. So this strategy requires the least intelligence (there's almost no thinking involved just repeatedly simplifying and multiplying) but probably the most endurance.

0

Like Proof$\#2$ of Proving a fact: $\tan(6^{\circ}) \tan(42^{\circ})= \tan(12^{\circ})\tan(24^{\circ})$

$$\tan\left(y+r\cdot60^\circ\right); r=0,1,2$$ are the roots of $$\tan^3y-3\tan3y\tan^2y-3\tan y+\tan3y=0$$

Using Vieta's formula $$\tan y+\tan(60^\circ+y)+\tan(120^\circ+y)=\dfrac{3\tan3y}1$$

But $\tan(y+120^\circ)=\tan(180^\circ-(60^\circ-y))=-\tan(60^\circ-y)$

I believe this is how the problem came into being.

We can similarly prove the following :

$$\tan y\tan(60^\circ+y)\tan(120^\circ+y)=-\dfrac{\tan3y}1$$

$$\tan y\tan(60^\circ+y)+\tan(60^\circ+y)\tan(120^\circ+y)+\tan(120^\circ+y)\tan y=-\dfrac31$$