7
$\begingroup$

With $\phi=\frac{\sqrt5+1}2$ the golden ratio, we have $$\log_2 \phi = 0.6942\ldots\\ \,\log_e 2 = 0.6931\ldots$$

Equivalently,

$$2^{\log 2} =1.6168\ldots\\ \quad\;\phi = 1.6180\ldots$$

Is it a simple coincidence or is there a deeper reason for this?

To give an idea of the sort of thing I'm looking for, the identity $\frac{1/\sqrt2}2 = \sinh \frac{\log 2}{2}$ is a very good explanation of the observation that $1/\sqrt2 \approx \log 2$. It also allows estimating the error term by expanding the series for $\sinh$.

Edit: Expanding on infinitylord's remark, this equivalent to the observation

$$0.49915\ldots = \sinh\left((\log 2)^2\right)\approx 1/2$$

$\endgroup$
  • $\begingroup$ It's depend on the problem, which we need to solve. $\endgroup$ – Michael Rozenberg Jun 12 '17 at 15:54
  • $\begingroup$ I think that I would to see them a lot closer before I got very interested. $\endgroup$ – badjohn Jun 12 '17 at 16:01
  • 2
    $\begingroup$ @MichaelRozenberg: It's a soft question, but to make it a little bit more formal I could ask: "Is there any algebraic equality that could help derive the bound $|(\log 2)^2 - \log \phi| < 1/1000$ or even $1/500$ without resorting to direct calculation of both terms?" $\endgroup$ – Milor Jun 12 '17 at 16:03
  • 1
    $\begingroup$ As an alternative view, $\log(\phi) - \log(2)^2 = \text{csch}(2)^{-1} - 4 (\coth(3)^{-1})^2 $ $\endgroup$ – infinitylord Jun 12 '17 at 16:16
  • 1
    $\begingroup$ It is easy to find a lot of very accurate coincidences, for example : $$\phi=\frac{1+\sqrt{5}}{2}\simeq 1.618033988..$$ $$\cos(\sqrt{2}\:e^{-2})-\cos(\sqrt[3]{2}\:e^\pi) \simeq 1.618033988..$$ $$\sqrt{\cosh(\gamma)+\cos(\gamma)}-\frac{\gamma^3}{\sin(5)}\simeq 1.618033988..$$ $$\cosh\left(G^2\sinh(1)\right)+\cos\left(\frac{\sqrt[3]{\pi}}{\cos(3)}\right)\simeq 1.618033988..$$ $\gamma=$Euler-Masheroni constant $\quad;\quad G=$Catalan's constant. From the paper : fr.scribd.com/doc/14161596/Mathematiques-experimentales , where it is shown how to compute such false equalities. $\endgroup$ – JJacquelin Jun 12 '17 at 17:33
3
$\begingroup$

Take a few well-known constants such as $\gamma,1,\phi,2,e,3,\pi$, a few functions $\log,\exp,\sin,\cos$ and the six basic operations $+,-,\times,\div,a^b,\sqrt[b]a$. Form as many simple expressions as you can by combining them. For example, there are more than $6000$ expressions of the form $f(x)\text{ op }g(y)$ many of which fall in a small range, say $[0.1,10]$. Hence you can expect a few pseudo-coincidences.

Below a plot of the values of $2717$ such expressions with a value in the above range.

enter image description here

This exhaustive search reveals the interesting

$$\frac{e^\gamma}{e^e}\approx\frac{e}{e^\pi}$$ $$0.117529\approx0.117467$$

which is nothing but $$\gamma-e+\pi\approx 1.$$ (Actually $1.00052649003$.)

$\endgroup$
  • $\begingroup$ pretty neat! how did you draw that? $\endgroup$ – Glougloubarbaki Jul 31 '17 at 19:18
  • $\begingroup$ @Glougloubarbaki Excel. $\endgroup$ – Yves Daoust Jul 31 '17 at 20:26
  • $\begingroup$ +1 I've made some minor edits for brevity. Can you give a few more examples from your search? $\endgroup$ – Tito Piezas III Aug 1 '17 at 12:49
  • $\begingroup$ @TitoPiezasIII: brevity ? Who cares ? $\endgroup$ – Yves Daoust Aug 1 '17 at 12:58

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.