The Bradley Effect: Looking at 5-Man Units

Posted by Hayes Davenport on Apr 2, 2012 2839432 commentshttp%3A%2F%2Fcelticshub.com%2F2012%2F04%2F02%2Fthe-bradley-effect-looking-at-5-man-units%2FThe+Bradley+Effect%3A+Looking+at+5-Man+Units2012-04-02+17%3A30%3A09Hayes+Davenporthttp%3A%2F%2Fcelticshub.com%2F%3Fp%3D28394

How fun is it to actually have a starter controversy to deal with? I don’t remember the last time it wasn’t a foregone conclusion who would start a game for this team. It was probably Eric Williams vs. Walter McCarty or something. What a horrible time to be alive.

And this Bradley-Allen discussion is as interesting as it is unfamiliar. So far we’ve mostly heard arguments for Bradley surrounding the need for offense on the bench, the necessity of a starter who can check the other team’s best player, and the name-checking of guys like Thabo Sefolosha and Jason Terry. But I’m mostly interested in which unit is better. This is how the debate should be settled: produce evidence for which unit is better, and give that unit more minutes than the other ones. There are hundreds of Trapper Keepers worth of data available on unit value. Unzipping a few of them could lend a lot to this discussion.

Because if you look at the numbers, the unit featuring Avery Bradley instead of Ray Allen has been much, much better. Like, coconuts better.

Lineup analysis is mostly based on the Offensive and Defensive Efficiency, or how many points the team scores or allows per 100 possessions when that lineup’s in the game. Here are those figures, first for the Rondo-Allen-Pierce-Bass-Garnett lineup, then the Rondo-Bradley-Pierce-Bass-Garnett lineup (via BasketballValue).

Ray’s lineup (269 minutes): 103.89 offensive efficiency, 101.78 defensive efficiency
Avery’s lineup (78 minutes): 114.57 offensive efficiency, 89.86 defensive efficiency

That gives Ray’s lineup a differential of 2.11, and Avery’s a screaming 24.70. Keep in mind that these figures DO NOT INCLUDE LAST NIGHT’S GAME, the most dominant game of the Bradley era, so Bradley’s differential is actually better than that. Plenty more after the jump.

Boston’s average efficiencies (including last night) are 98.6 on offense and 95.8 on defense. Notice that the offensive rating of Ray’s lineup is slightly higher than the average, as you’d hope. But if the offensive rating for that Bradley number, if we could extrapolate it to a full season, would make for easily one of the top three highest-scoring starting lineups in basketball. Miami’s starters put up a 109.93, which isn’t even close.

Same goes for Bradley’s defensive rating of 89.86. Other than Philly’s starters and Chicago’s bench-with-Deng unit, that lineup would give up fewer points than any other high-rotation lineup in basketball. And again, the Bradley unit numbers are actually better than those after yesterday.

There are plenty of other factors flying around in these figures, the small sample size foremost among them. Bradley’s had the good fortune of entering the starting lineup just as Kevin Garnett starting using his powers of concentration to herd the cells in his bone marrow and return his body to its high school powers. Rajon Rondo has also been playing basketball with a renewed commitment to playing basketball.

But wouldn’t it be great to find out how much is noise? With a lineup this hot, a coach should send it on the floor until it stops crushing other teams, probably, right? Which is why I don’t get this as a debate. If Ray Allen feels like coming off the bench would be an affront to his dignity, and preserving Ray’s dignity is Boston’s primary organizing principle, than the team should stop trying to win basketball games and start writing acrostic poems about the value of Ray’s friendship. But if they want to win games, they should keep playing a super-dominant lineup until it stops being super-dominant.