Well, the best thing to do would be to work outOriginally Posted by c00ky
for several values of x between 0 and and see what happens when you compare that value to cos(x) when x is in radians.
What you will find is that this function only works well when x is "small." I don't know how well the approximation works for x near, say , but for x less than, say, 1 it's probably pretty decent.
For the record, as a Taylor series about x = 0 we may approximate cos(x) as
so this approximation should be valid.