Most mobile learning applications support individual users, although experience with similar conventional learning games indicates that teams may be more appropriate. This paper reports on tests of the MobileGame to see whether individual users, teams of two, teams of three or teams of four are more successful. The test was conducted with over 100 natural users. The significantly increased activity level and team-building show that a preference for teams of two rather than individual players would be justified. There is little significant evidence to prefer teams of two to teams of three. However, the data shows that teams of four are suboptimal: this team size decreases fun and immersion as well as (maybe) learning. There is no evidence that these negative effects are balanced by improved team-building. The relatively high success of teams of two leads to a need for more research on dyadic users not only for mobile learning games, but also for other areas such as tourism, health, museum visitors, and entertainment.