I was tempted to make a similar game infrastructure design decision recently, namely, faking players in game highly dependant on reaching critical volume of players, in attempt to alleviate problem of players waiting too long for their next opponent.

I, of course, said 'no' to myself, and then tried to work around it, and [some technical difficulties ommited] we now have awesome feature presented as Special Encounter where players play against NPCs, but still weighed similarly to plays against real players.

Being plain evil is easy way out, with higher risk assigned to those decisions. But somewhere in the problem-solution space, there's an ethical solution begging to be found, frequently better performant than the obvious evil one.

Unrelated, how long did it took you to fold a company? It's a lengthy and painful process here, even if you're profitable and without debts to anyone.