"Change in rank" I don't really think is that great of a metric because the ranks not uniformly divided. That's to say that for one person, a pulse of 50k keys may increase their rank by 10, whereas for someone else that same pulse might not increase their rank at all.
I think ideally, the best metric to use would be the keys/clicks/miles within the specified timeframe. So at the beginning of the event a snapshot is taken of every user participating, and the same at the end. The difference between the figures is the net gain in keys/clicks/miles which would be used to rank the "competitors".