We study stochastic games with a large number of players, where
players are coupled via their payoff functions. To deal with the
complexity of finding an equilibrium in large-scale stochastic games,
Weintraub et. al, introduced a notion of oblivious equilibrium
(OE). In OE, each player reacts to only the average behavior of other
players.  In this talk we focus on a special class of stochastic
games, where a player experiences strategic complementarities from
other players; formally the payoff of a player has increasing
differences between its own state and the aggregate distribution over
the states of other players. We find necessary conditions for the
existence of an oblivious equilibrium in such supermodular mean field
games. Furthermore, as a simple consequence of this existence theorem,
we show that from the viewpoint of a single agent, a near optimal
decision making policy is one that reacts only to the average behavior
of its environment.