I have struggled with getting EKOS to autofocus with my Esatto and C11.  

I discovered that I needed a much larger initial step size than I had thought.  In order to move my HFR values 3x the initial number, I needed to use a step size of 30,000.  The Esatto has 420,000 steps, so 30K is not has large as it seems.  I tested it with the Polynomial method and it achieved perfect focus in just 7 steps.  Repeatedly.  I was never able to get the Linear method to work, as it still stops at the wrong spot after calculating the correct one.  But the Polynomial works great.

So the advice that solved this was simply to manually achieve perfect focus, measure the HFR, and then see how many steps it takes to get an HFR values 3 or 4 times higher.
With the Esatto, that turned out to be 30,000.