Post

A neat trick with inverse trigonometric functions

A neat trick with inverse trigonometric functions

Update 2023-01-25: So I first encountered this in a STEP problem 00-S2-Q4 on STEP Database. Today I found something that is relevant and it’s Machin’s formula and its generalisations.

Update 2022-04-28: Michael Penn just did a video on solving a differential equation with a similar technique, go check it out here.

Bullcrap, don’t read

So I was practicing STEP while my brain was malfunctioning (at this point it might be the norm, so I guess functioning as normal?) and struggled to combine 2 arctangent expressions. Being the pussy I am, I checked the hints and wondered why I have never come across this before.

The ‘Trick’

Suppose I want to simplify arctana+arctanb, for a,b in the domain, one should consider tan(arctana+arctanb) and apply the compound angle formula. We obtain

arctana+arctanb=arctan(a+b1ab).

Unfortunately, this suffers from problems of the range of ab. The above formula works only for ab<1 due to signs and periods and stuff. For ab>1, we have the following cases:

For a,b>0, we have

arctana+arctanb=arctan(a+b1ab)+π

For a,b<0, we have

arctana+arctanb=arctan(a+b1ab)π

The proof is left as an exercise to the reader or you may consult these links on MathStackExchange, which explain things much clearer than I do.

Do note that

tan(arctana+arctanb)=a+b1ab

works for all a,b though.

Naturally, we may repeat this with similar functions and produce the following results:

sin(arcsina+arcsinb)=a1b2+b1a2 for a,b[1,1] sin(arcsina+arccosb)=ab+(1a2)(1b2) for a,b[1,1]

Similar results may be derived for other trigonometric functions.

We now turn our attention to hyperbolic trigonometric functions. Actually, no, we are not going to do that since that is just logarithms.

This post is licensed under CC BY 4.0 by the author.