So remember that these are just means. Anything could’ve happened. It’s a bit like flipping coins.
So in the instance of a game with two means - 1,5 and 1,8, you have to run the simulation and see what happens.
So team 1 will have its simulation of 1,5. You’ll have the distribution of outcomes for that xG. You’ll overlay with the distribution for team 2, which will be slightly different given its own composition. And then you’ll have the potential outcomes of the games. And you might see that more often than not it’s a draw versus a victory.
But the point is just taking the point on its own is uniformative because these are highly variable stats. It’s like taking a coin flip and saying it gives you 50% every time when the reality is you could get a head or a tail, which are vastly different outcomes in an individual event.
