I don't ever remember see moving parts on a trolley frog pan. But it seemed instead that the frog pan, and intersecting trolley wires were positioned well beyond the point of intersection of the turnout. By the time the trolley wheel got to it, lateral forces were already being applied between the trolley wheel and the wire, thus when in the pan, it natually when off on the diverging wire (if the car took the diverging path). When going straight, lateral forces would not be there, thus not pulling the wheel to the side. And then the pan's shape would capture the trolley wheel before it left the pan, guiding it to center on the straight wire.

This was only observed from the ground on the SF Muni, I may have not have actually been able to see what was really going on. Maybe there is an "Expert" somewhere that would set me straight - or confirm my observations?