> She was pretty much completely immobile from the neck down, and couldn't even see our hands properly from her wheelchair. She could only see the arc of the ball, but that was sufficient information for her to tell us how we could improve. "Pull your elbow in". "Focus on the left hand, the right will follow".
I've both been a coach (paintball/martial arts) and been coached (golf) and it really is wild how good your brain can become at seeing the outcome or just a piece of the process and then working backwards to a root cause.
I sometimes make the analogy "in particle physics, you don't actually see the collision. You see the after effects and then figure out what happened by going backwards to what must have occurred."
My favorite version of diagnosing a root cause from an outcome was "Car Talk" on NPR, where somebody would call into the Tappet Brothers's show and imitate the weird sound that the car was making, and then the brothers would diagnose a very specific car problem based on the secondhand impersonation of a weird noise. I have no idea how accurate their diagnoses actually were, but it always seemed like a tremendously impressive trick.
>> I sometimes make the analogy "in particle physics, you don't actually see the collision. You see the after effects and then figure out what happened by going backwards to what must have occurred."
I keep coming back to that. Nobody has ever directly seen direct the force carrying particles, only their effects (indirect evidence). The models make excellent predictions, but I still feel like they're "wrong" in some sense.
This is the only fact we know for certain in physics; we know for sure that our existing models are incorrect. But they sure are great at lots of stuff and correctly predict lots of phenomenon so until someone can come up with a better model they're likely to continue to be used widely!