Although IT shouldn't have called out anyone in the media, Stevens does experiment too much with lineups , especially late in games and late in seasons. I think it's part of his process, to get enough of a statistical sample size in important game situations, to be able to statistically evaluate the effectiveness, or lack thereof, of various lineup combinations, so that he can be confident in what should work in the playoffs. The thing is, unless he explains that to players ahead of time, they aren't going to agree with those lineups. Additional problem is, if he does explain it to players ahead of time, he biases the data.