Does coaching matter in basketball?

Professor Berri's, the architect of the Win Score metric and the author of the fabulous blog "The Wages of Wins Journal" is coming out with a book Stumbling on Wins. I'm waiting for it with the same anticipation those Star Wars geeks waited for those awful movies that made up the "prequel trilogy".

But I might have some trouble swallowing the book's central thesis. Berri and Schmidt will argue that coaching does not matter much in the sport of basketball. But I'm not sure that's true. At the risk of commiting an act of heresy, I'm going to try to rebut that assertion.

Whenever I'm trying to figure out whether something may have an influence on something else, I learned to use an Einstein style thought experiment that imagines the opposite. Lets do that in this case. Lets say your author was appointed Coach of the Milwaukee Bucks because the Senator wanted to sabotage the team, like the owner in the movie The Natural, or Major League, I guess.

Couldn't I do that through misguided assignments and poor distribution of playing time? Couldn't I scuttle the team's chances by sitting all the best players, playing all the worst, and then deploying those scrubs in disadvantageous positions? Obviously, I could. If I played all of the team's "replacement level" players, I could reduce the team's Win Probability to replacement level: 25%. So coaching can have an effect.

So for the "coaching doesn't matter" thesis to be true, one has to assume a certain level of technical competence amongst NBA coaches. But I'm not sure even that exists. (Now, obviously, I'm in dangerous waters here, because I'm critiquing a thesis I have never read. But, the Slate article seems to lay out what I am outlining.)

Take the case of Kevin Durant. Some birdbrain in the Sonics/Thunder organization decided last season to play him at shooting guard. That misdeployment costs the team dearly. As a shooting guard, Durant's career MWS48 is -3.9... way below replacement level. However, if you move him to small forward, his career MWS48 is +2.3... way ABOVE replacement level. Finally, after an early season COACHING change, it appears Durant is becoming the productive player we all envisioned him to be out of college. So, in his case, coaching seems to have mattered, and I've run across dozens of examples of the same-type misdeployments (hell, the Timberwolves have about 5 such cases by themselves -- Kevin Love is far better at center than power forward, yet he is usually deployed at power forward -- Al Jefferson is far better at power forward than center, yet he's routinely played at center -- the injured Corey Brewer is far better at shooting guard than at small forward, yet he normally plays small forward)

And if coaching does not matter, how do you explain the effect Scott Skiles has had on the Bucks. Lets examine the players he inherited from last season... the same cast of characters who couldn't defend a junior high team.