Let's say I have two models to predict the out come of a game.
Model A uses the player specific statistics (speed, reaction times etc.) and predicts a 75% chance of a win for my team.
Model B uses the team statistics (recent performance, historical performance against the other team etc.) and predicts a 75% chance of a win for my team.
Can I combine these two predictions and be more than 75% confident that my team will win? My gut tells me No, but there's more to this.
If I ask a friend if they think I'm making a good bet (for example) and they agree then it'll make me more comfortable with the bet. If I ask 1000 friends, then I'll be significantly more happy that I've made the right call (if they all agree). That makes me think that, intuitively at least, my prediction confidence increases the more predictions I get (independent predictions ofcourse).
Another way of looking at this is that each model only has access to a subset of the variables. If I trained my model with all the variables it should be a better predictor (assuming enough dataset).
So what's the deal? Can I combine the models predictions? If so how? Is there an equation or methodology I can use?