What does Adjusted Plus Minus say about the 2013 – 2014 Cavs?

January 14th, 2014 by Kevin Hetrick

varejao-001.jpg?w=610

Editor’s Note: Comments were accidentally turned off for this article, earlier.  They’ve been re-enabled.

As gets mentioned around here a tad, Nate, Tom and I, in addition to the esteemed Randall Cooper (@laughingcavs), started another website, gotbuckets, largely devoted to Adjusted Plus Minus (APM) and Regularized APM (RAPM).  For a brief discussion of APM, check here. Yesterday, the first update of 2-year APM to incorporate 2013 – 2014 play was posted.  With that, let’s discuss what the results may mean for the 2013 – 2014 Cavs.  Keep in mind that these results include all games from 02/14/2012 through 01/06/2014 (not exactly two years due to strike shortened season), so it’s much more than just this season.  Using reasonably large sample sizes is important for APM though.

Based on the update, Tristan Thompson rates as approximately the NBA’s 150th best player over this time frame, slightly below average on offense (average being approximately zero), but a positive defensive force.  Compared to the regression encompassing the 2011 – 2012 and 2012 – 2013 seasons, Tristan’s early play this season is slightly improved on offense, but regressed on defense.  That particular result will be discussed a bit more with Andrew Bynum, but as a snapshot of this entire 2-year timeframe, APM pegs Tristan pretty well.

Over the previous five seasons, the APM regression continues to give Anderson Varejao much credit for any Cavaliers success, especially as a defensive nuisance, where he rates among the league’s best.  For the four prior 2-season regressions, his overall APM never rated lower than 78th.  Based on the current update, he reaches a peak level of tenth.  That certainly overstates Andy’s importance, but his impact on the team’s fortunes does go under appreciated, especially outside of Cleveland.  Through 36 games this season, the Cavs had been 10.8 points per 100 possessions better with Wild Thing on the court.  Can the NBA put a guy averaging 8 points and 10 boards in the All-Star game?

With Andy, I wanted to raise the readers’ spirits before mentioning Kyrie Irving.  The APM regression probably declares him the most overrated player in the NBA.  His offense has been solid, ranking as the 50th best per possession player of guys with over 2000 minutes during this timeframe.  But his defense?  The result that gets churned out is hideous, in the bottom ten percent.  The end result…a below average player.  A variety of thoughts on this:

  • In small sample sizes, APM can be quirky.  Kyrie is a young guy that battled injuries each of the previous two seasons.  It can be deceiving to read too much into one or two years of APM. Once upon a time, Wayne Winston, one of the originators of APM, said he would not want Kevin Durant on his team, primarily due to his early career APM results.  It seemed silly then, and is obviously worse in hindsight.  With his elite offensive ability, Kyrie’s evaluation via APM will almost certainly improve (Durant now rates as elite, as would be expected…3rd on the gotbuckets MVP board).
  • Andrew Bynum used to be one of the NBA’s best paint defenders.  The spreadsheet that calculates APM has no idea that he is a shell of the player he was in 2011 – 2012, half of which season is still included in a 2-year regression.  But it does know that several Cavs have been playing with Bynum, the formerly amazing defensive enforcer.  To an unbiased regression, why are recent defenses featuring Bynum so poor?  Certainly most of the blame gets dispersed onto the other Cavs playing with Bynum.  (As time goes by, Bynum’s 2011 – 2012 performance will constitute less and less of the regression, and all that will be left is his largely stationary November and December of 2013.  As the regression only considers that Bynum, the Cavs that played with him will see a positive bounce).
  • From the 2011 – 12 and 2012 – 13 regression, Kyrie was a top-15 offensive player out of guys with over 2000 minutes.  Obviously after last year’s All-Star break and early this season, something was amiss with Kyrie’s shot.   If he keeps draining shots and flashing his recently improved penchant for distributing (prior to Sunday, over his previous 10 games, he averaged 7.3 assists and only 2.4 turnovers), he should resume a trajectory towards the league’s elite offensive players.
  • So there are some reasons to not over think Kyrie’s APM.  On the other hand, it would be dumb to ignore it.  Steph Curry is 4.5 standard deviations above Kyrie.  Damian Lillard is 3.5.  That is not a statistical fluke.  Certainly early in his career, Kyrie has been and still can be an abysmal defender.  It’s something he needs to keep improving on.   Also, the Cavs stink, and have for a few years.  His fancy, extended ball-handling displays probably don’t lift the performance of his teammates as well as they could.  Anyways, keep working on defense, and on offense, play like he did pre-All Star break 2013, and the recent ten games stretch.  Easy, right?

Dion Waiters rates as a reasonably solid player, largely thanks to APM’s assessment of his defensive play improving  from last year, when he provided the 14th worst defensive APM in the entire league.  Generally there appears to be a positive “chaos effect” that Dion imparts into the offense, moving the ball and attacking the basket, as the team’s offense performed 5.1 pp100p better with him on the court last year (best differential on team), and continues to show more success this year when he plays.  On the other hand, don’t automatically get over-excited about one APM set.  If this continues as a trend, then we’re onto something.

Jarrett Jack rates as ineffective on offense and defense, calculating overall as part of the bottom 25% of all NBA players.  Thanks to his brief time with the Cavs, his offensive value has plummeted, whereas in Golden State, he was above average.

The APM regression churns out that C.J. Miles has actually been the second best per-possession player on the Cavs during this 2-year timeframe.  I won’t go that far, but for 2 years and $4.5 million total, he’s been a really nice free agent signing for Cleveland.  Almost amazingly, through Friday the Wine & Gold outscored their opponents by 3.7 pp100p during Miles’ 646 minutes.  That is equivalent to a 52 win team.  Keep finding minutes for that man.

Who should those minutes come from?  Well, Alonzo Gee of course.  His APM has progressively worsened from 2011 to 2012 to 2013 to 2014.  It was a nice dream, thinking the Cavs unearthed a rotation player from the D-League, but that time has come and gone.

APM also continues to exhort  a loud, “Earl Clark stinks at offense!”, where the updated result indicates he is the NBA’s 19th worst player at bucket generating.  The real problem is that APM has never considered him better than average at D.  Add it up, and it’s not pretty.

But who cares, right?  Luol Deng!!  He continues to grade out as a guy who is generally average or better on offense, and well above average on defense.  Whereas no longer showing up as elite on D, still, minute weighted, APM likes him as a top 50 NBA player.  In the Eastern Conference, that may warrant an All-Star berth.

The early results of 2013 – 2014 have plummeted Andrew Bynum into the bottom 20% of players according to APM.  Based on the remnants of 2011 – 2012 still in the current regression, his defense rates as above average, but I can tell you, it is not.  A mediocre Cavalier defense was notably worse (3 pp100p) when the slow-moving seven-footer took the court.

Finally, Tyler Zeller. Based on limited minutes this year, he is still largely saddled with last season’s underwhelming performance.  APM thinks he stinks, and we are still biding our time waiting to find out if that is right.

That is it.  Matt Dellavedova and Anthony Bennett haven’t registered 750 minutes, so no 2-year APM for them yet.  Other than Varejao and Miles, the results aren’t necessarily pretty.  I may be the only one, but it will be exciting to see how Kyrie,  Tristan and Dion’s play continues to affect the scoreboard moving forward.