I'd focus on coins only. There was a link on HN about some pilot study by some students, where they tried to emulate or used an actual mechanical hand or someth

I'd focus on coins only. There was a link on HN about some pilot study by some students, where they tried to emulate or used an actual mechanical hand or something similar flipping coins. How? (though not in my protocol) For example, attaching a magnetic levitation system under the surface with the flipper unaware of it, and upon switching the current on through the system, the coin levitates probably on its edge. And it is a problem for the flippers and inference makers for not being able to arrange for such option.