conjecture Right, so this was something I stumbled upon whilst looking at forms of the golden ratio (if you replace x with one in the diagram above then you get phi) and then I wondered about what if you messed with the value of x, did a massive amount of trial and error (lost in a notebook somewhere) then came up with the equation in the right, which worked exactly for every value I tested (about up to 100)

Now I have found stack exchange, in wondering if there is a way to prove this, I plugged in some values into wolfram alpha and geogebra, which seems to show it works, but I want a proof, not just trial and error answers. graphs!


1 Answer 1


This is what's called a continued fraction. Let's denote your continued fraction as $f(x)$. Notice that if we subtract $x$ from $f(x)$ and take its reciprocal, we get $f(x)$ back. Therefore, $$\frac{1}{f(x)-x} = f(x)$$ or $$f(x)^2-xf(x)-1 = 0$$ Solving for $f(x)$ using the quadratic formula, we get $$f(x) = \frac{x\pm\sqrt{x^2+4}}{2}$$ It's not hard to see that if $x$ is positive, then $f(x)$ must be positive (as the reciprocal of a positive number is always positive, as is the sum of two positive numbers). Therefore, for $x > 0$, we have $$f(x) = \frac{x+\sqrt{x^2+4}}{2}$$ This is the same as $\frac{x}{2}+\sqrt{\left(\frac{x}{2}\right)^2+1}$ (where we have just distributed the $\frac{1}{2}$). Notice that $f(1) = \frac{1+\sqrt{5}}{2} = \phi$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.