In a former life I had to sample optical transmission vs. wavelength in thin films to determine optical thickness in real time, relying upon the periodic nature of constructive interference.
In a technique looking analogous to the focus V-curve, I found the transmission minima by using a quadratic equation - fitting a quadratic function to 3 samples and then differentiating it and solving for zero.
At this point in time I think you could apply it to focusing algorithms too, with just three points straddling the focus. I’m going to build up a spreadsheet and test my theory against the AF figures from my logs. If it looks promising and I get good correlation between V-curve vs quadratic fit, I might ask for some other user’s focus data. It is possible that it may not be as accurate as an extended V-curve but it might be useful for quick coarse focus using just three points.