Over the last week or so, there has been a fantastic discussion of Kevin Durant, specifically in the context of +/-. Numerous cases have been made from TrueHoop to Basketball Reference . However, one point that has not been properly made yet may hold the truth of the matter.
Kevin Durant played an absolute load of minutes last year- 39 MPG in the 74 games he appeared in. While I find +/- to be an incredibly useful stat most of the time, it runs into problems when you look at players with particularly high minutes played (38+ as opposed to guys in the high 20's or super-low 30's) because such heavy minutes changes the situations where they do and do not play.
Unfortunately, I am not privy to a minutes graph or substitution pattern chart for Kevin Durant from the 2008-2009 season. However, what makes this situation different from other seemingly parallel seasons is that the Thunder of last year were a simply terrible team, especially at the level of their starters. It would follow that they were "knocked out" of a portion of games on their schedule. Since Durant was (or at least was perceived to be) an extremely important part of any potential success for the team and played a ton of minutes, it would follow that he would get very heavy minutes during the portion of the game when the outcome was still in jeopardy.
Additionally, it would follow that he would largely come out of the game after the game had been effectively clinched by the other team to avoid risk of injury- it would essentially parallel a line shift in hockey, except that Durant was the only player on that substitution pattern.
As someone who actually watched a fair amount of Oklahoma City?s games last year, memory tells me that this disparity between KD and the starters was especially true for Russell Westbrook, as he was essentially learning the point guard position and needed all of the exposure to game situations he could get, regardless of whether victory was still in play. At that juncture, the Thunder would likely play better in terms of +/-, since the imminently victorious opponent would no longer have either the incentive to play their best players or the incentive to play as hard (for those top guys who stayed in- I assume the lower rotation guys would want to try and get noticed and whatnot).
This also makes sense considering Oklahoma City was exactly even against their opponents when Durant was off the court- this team was not even with the average opponent on talent in any reasonable circumstance, especially with KD off the court.
Interestingly, it also appears that many of the writers out there are forgetting the unmitigated disaster that was Kevin Durant at shooting guard; he played 8% of the team?s minutes at the spot and posted a more than disastrous -17.3 Net48. This time was not only the nadir of Durant's net productivity; it also marked the period before Scotty Brooks took over and turned the Thunder from a terrible team to a below average team. Again, this does not wipe away the flaws in Durant's game, but it certainly helps explain why his +/- was so terrible.
Thanks to the fact that Durant played far heavier minutes than the rest of his Thunder teammates (only Jeff Green played more than 33 MPG), it would follow that no other player on the squad would follow his usage pattern. Of course, Durant's defensive limitations and all of the other compelling factors that writers have made play a factor too, but I would love to see if a front-loaded minutes distribution on a terrible team was the larger contributor to the apparent anomaly.