Just for fun, here's a very naive model for winner-take-all tournaments with 10, 100 or 500 players. We assume that all the other players are of equal strength with one another, and we expect to win 51% of the time against a single opponent. We also assume there's no rake, no friction and no air resistance. We then calculate our expected value from the tournament for every field size.
EV for 10 players: -1 BI + 0.51/(0.51+9*0.49)*10 BI = 0.037 BI
EV for 100 players: -1 BI + 0.51/(0.51+99*0.49)*100 BI = 0.040 BI
EV for 500 players: -1 BI + 0.51/(0.51+499*0.49)*500 BI = 0.041 BI
Same calculations if we expect to win 55% of the time against a single opponent:
EV for 10 players: -1 BI + 0.55/(0.55+9*0.45)*10 BI = 0.197 BI
EV for 100 players: -1 BI + 0.55/(0.55+99*0.45)*100 BI = 0.220 BI
EV for 500 players: -1 BI + 0.55/(0.55+499*0.45)*500 BI = 0.222 BI
So at least under these circumstances, it would be more profitable over the long term to play two 10-player tournaments than one 500-player tournament over the same time frame. Of course, it's a separate question whether the above assumptions have any relevance to reality