i have no clue how it works. i think noone knows the formulas so noone will know. it will be all pure speculation. Most agree that many different stats influence these team ratings.
Myself and quite a few other ones agree I believe, the numbers for team ratings and for player ratings forget them. they don't make sense at all. many teams have a lot worse ratings and still beat the way better team (when you look at the team ratings only), but than when you look at the game stats it gets very clear... the "weaker team" had more rebounds, better shot %, less pf and so many other things you can read from the game stats. they will never lie...
ps: I don't have proof. but a player that gets a 12 in rating in a div IV, doubt he will get the same 12 in rating in division I. that would mean that these ratings are quite useless...