You could gather empirical data in controlled settings and make a model that relates the things you can measure easily (speed and downward force) to the things

You could gather empirical data in controlled settings and make a model that relates the things you can measure easily (speed and downward force) to the things you care about (rock distance and control). I don't know for certain, but I'd guess that's what's been done here.