From 2015 - 2019 dancers competing at the Lindy Hope Championships were split into 4 levels of expertise. The level four dancers beat the level three dancers on average 83% of the time. If we were to automatically assume that level four dancers have a higher elo rating than level three dancers how much higher would that rating be. The equation for calculating elo is: $$ e = \frac{1} {1 + 10^{(b-a)/400}}, $$ where $e$, I think, is the expected amount of time of victory. $b$ is the rating of the level 4 dancer. $a$ is the rating of the level 4 dancer.
We know that $e$ must be $e= 0.83$ and $a=0$. But I cannot figure out how to solve for $b$. Any insight would be appreciated.
The solution is as follows
\begin{align}e &= \frac{1} {1 + 10^{(b-a)/400}}\\ e\left(1+10^{(b-a)/400}\right)&=1\tag{1}\\ 1+10^{(b-a)/400}&=\frac 1e\tag{2}\\ 10^{(b-a)/400}&=\frac1e -1\tag{3}\\ \frac{b-a}{400}&=\log_{10}\left(\frac 1e -1\right)\tag{4}\\ b-a&=400\log_{10}\left(\frac 1e -1\right)\tag{5}\\ b&=400\log_{10}\left(\frac 1e -1\right)+a\tag{6}\end{align}
where the steps are
You can then use Wolfram|Alpha to solve the equation by filling in your values for $a$ and $e$ to find that $$b=-275.452$$