I have two sets of measurements that measure one thing. I would like to graph the relative heights of the measurements, however one set is nearly consistently higher than the other one. I would like to know how to determine the constant I need to substract from the higher dataset, that minimizes the sum of the differences between each data point pair. So I tried WolframAlpha.com and set up an equation like this: $$ \begin{multline} \frac{d}{dx}(|(13840-x)-8192|+|(9588-x)-6666|+|(5713-x)-3994|+ \\ + |(12625-x)-8170|+|(22024-x)-4742|+|(10439-x)-8522|+ \\ + |(12725-x)-12525|)=0 \end{multline} $$
But WolframAlpha doesn't solve it since the computation times out. Is there a more elegant way of doing this?
well what about
in Wolframalpha