Friday, January 4, 2013

Steve Novak's Defense - By The Numbers

Steve Novak's defensive abilities have recently been the subject of some debate.  Synergy says he's good, and 82games says he's not.  Our numbers say they're both right, partly.

We can measure defensive ability across many different categories, including:

a) how well he keeps opposing drivers in front of him (keep in front% or KIF%)
b) how well he contests, alters, or blocks shots (Contest+%)
c) opposing players' field goal percentage when he is defending (FG% Against)
d) how well he double teams (Double Team Effectiveness Rate or DBL)
e) how well he helps on defense (Help Effectiveness Rate or HELP)

A) Keep in Front 

In our sample (which includes the final quarter of last season and a little less than half the games played this season), Novak's KIF% is 56%.  This includes all situations, from screens to closeouts to isos.  This puts him below the median for the Knicks, a little better than the league average, and just about 10 percentage points higher than Kobe Bryant (with apologies to morning radio hosts who still think Kobe is an adequate defender). 

B) Contest+

Our data has already shown the importance of getting a hand up on shooters.  At 38%, Novak's Contest+% is average for the Knicks but toward the bottom of the league.  

C) FG% Against 

Novak at 42.11% is average in this metric for both the Knicks and the league.

D) Double Team Effectiveness

DBL measures all attempted double teams as well as outcomes deemed effective (no points, no assist, etc.) and ineffective (assist, pass to open shot, crucial pass, foul, points).  At only 6% DBL, Novak is the worst double team defender on the Knicks and watching the film is a painful experience.  Novak's inability to double quickly allows offenses to easily find open shooters.  Furthermore, Novak allows opponents to split his double team once every 5 attempts. This precipitates a complete defensive breakdown.

E) Help Effectiveness 

HELP measures all attempts to help a teammate after a No Keep in Front and again takes into account effective (no points, keep in front, no assist, etc.) and ineffective outcomes.  At 12.5% HELP, Novak is again the worst Knick.


What these metrics tell us is that if one were to look only at Steve Novak's on-ball defense, he's definitely adequate and even above average in some important categories such as KIF%.  But the full picture shows that because of his inability to effectively help or double team, he becomes a liability on the defensive end.  Because his weaknesses lead to points scored on his teammates, defensive ratings that only account for points against will not accurately measure his overall defensive value.


  1. I like this - but I would posit that Kobe is still an actual better defender than Novak - how often is Novak paired off against an all-star caliber offensive player (especially vis-a-vis Kobe)? Given that his help/double values are so low (I suspect guys who generate a double team are in the 80th percentile "offensively", so when Novak goes there, it's against a significantly better player than the guys he's Keeping in Front of)

    1. Your hypothesis could well be right. That was an off-hand comment about only 1 metric (KIF%) and we did not provide a systematic comparison of the two players across all relevant defensive metrics.