Defense in the NBA Don’t Get No Respect

If you haven’t realized it yet (or are new to the site), my posts tend to come in waves.  I spend a lot of time on one topic and then veer away towards another.  With my latest post, I delved into the exciting and sometimes daunting topic of understanding advanced basketball statistics.  My next set of posts, which will lead to my most technical to date (I promise it won’t be scary!) will present the leaders and losers in some of these statistics category.  Eventually, this will lead into to a regression model that explains and/or predicts salaries in the NBA based on regular season statistics and other relevant variables.

But today we’re going to focus on the leaders in win shares from this past regular season.  As I mentioned in my previous post, win shares estimates the amount of a player’s team’s wins that are attributable to that player.  Win shares is comprised of two separate statistics: offensive win shares and defensive win shares.  These sub-statistics break win shares down into the offensive and defensive components, but as I’ll discuss below, it doesn’t necessarily do so equally.

Win Shares

Let’s start by presenting the players who earned the highest total win shares last season.

The two doomed pretenders to the vaunted MVP throne on this list, Jimmy Butler and Rudy Gobert, averaged 23.3/5.3/6.0/2.0 (points, assists, rebounds, turnovers) and 14.9/13.6/2.8 (points, rebounds, blocks) per 36 minutes.  Devoid of context, these are all really good numbers.Three of the five highest win shares recipients this past season belong to three of the four legitimate MVP candidates, with the other (LeBron) came in sixth.  James Harden led the NBA in win shares so for those who try to make the case for James Harden as MVP, this is great news.  For the non-Harden supporters, well, it’s easy to mention how imprecise these statistics are and how they are all within a standard deviation (or so) of each other.

What’s interesting is that Butler and the Stifle Tower arrived on this list from very different paths.  Of Butler’s 13.8 win shares, 10.0 are offensive win shares.  Conversely, Gobert’s win shares are more evenly split with 8.3 offensive win shares and 6.0 defensive win shares, which is the highest amount of defensive win shares earned by any player this past season.  8.3 offensive win shares is above average but 6.0 defensive win shares is startlingly impressive.

While these gaudy numbers speak to Butler’s and Gobert’s offensive and defensive abilities, respectively, they also speak to the wider variability in offensive win shares compared to defensive win shares.  Last season, 52.8% of win shares were offensive win shares and defensive win shares composed 47.2% of total win shares.

This is fairly equal but it’s extremely deceiving.  Players can earn higher offensive win shares (highest was 11.5) and much lower offensive win shares (minimum was -1.4).  Comparatively, defensive win shares ranges from 0.2 to 6.0.  Higher win share totals are typically driven by offensive win shares.  For example, if we only include players who earned 4 win shares or more, offensive win shares comprise 62.4% of total win shares.

What this suggests is that defensive ability is not being as accurately measured as offensive ability.  Greater variability in offensive win shares suggests a more precise calculation of offensive win shares (assuming a normal distribution).  Less variability in defensive win shares suggests that a lot of defenders (elite defenders like Gobert excluded), regardless of actual skill level, are being lumped together as average, which doesn’t show the true variability in defensive ability (assuming that there is variability in defense ability, which I think is the case).

Offensive Win Shares

Now that my tirade on the issues with compiling total win shares from offensive win shares and defensive win shares is over, let’s review the players with the highest offensive win shares.

The extra A Isaiah Thomas (Knicks killer Isiah Thomas spells his name a little differently) established himself as the NBA’s shortest superstar, is about to be paid handsomely, and is the most successful of the NBA’s version of Mr. Irrelevant.  LeBron is LeBron.  He’s good enough to do whatever he wants offensively.As I discussed above, Jimmy Butler had a great year offensively (but not good enough to command a ransom from the Celtics) that warranted him the third highest offensive win shares.  James Harden led the league in win shares and offensive win shares, mainly due to his three point shooting and ability to get to the line.  49% of Harden’s shots were threes and he took 57 free throws for every shot he took, both of which are just ridiculously high.

As a Timberwolves fan, Karl-Anthony Towns has me excited for the first time since Sam Cassell injured his hip doing something called the Big Balls dance.  To best LeBron in offensive win shares (albeit slightly) in his second season is extremely impressive, which is definitely due to KAT’s workouts with Sasquatch.  I’ve heard that Sasquatch is quite the workout warrior.

Defensive Win Shares

So what about the elite defensive players?  We know the Stifle Tower leads this category but who falls closely behind him?  The table below details the top 5 highest defensive win shares from the past regular season.

As previously mentioned, Rudy Gobert is a defensive god gifted from the heavens to injure his knee, break Chris Paul’s heart, and watch the Warriors, including #2 on this list, Draymond Green, crush his playoff dreams.  Draymond, despite all the weird things that come out of his mouth, is a weird 6’5 defensive hybrid capable of incredible defensive numbers (the craziest is 2.2 steals per 36 minutes).  He is one of the few players who I can compare to Dennis Rodman without people laughing hysterically to my face.

The other guys on this list?  The Brow is, um, pretty good at playing basketball and is capable of playing any position on the floor.  Hassan Whiteside is a great rim protector, averaging 2.2 blocks per 36 minutes.  Andre Drummond combines his ability as a rim protector (1.3 blocks per 36 minutes) with a surprisingly deft ability to steal (1.9 steals per 36 minutes).

The low totals for the highest defensive win shares compared to offensive win shares sticks out like a sore thumb here.  For superstars, total win shares give players a lot more credit for their offensive abilities than their defensive abilities. This presents a major flaw in win shares. The correlation between win shares and offensive win shares is 0.96 (almost a perfectly positive relationship). In contrast, the correlation between win shares and defensive win shares is 0.77, which is still high but not as high. As I’ve discussed above, this suggests that offensive win shares has a greater impact on total win shares than defensive win shares.

What is the point of all this?  It’s great to delve deeper into statistics to better understand them and to understand their limitations.  Total win shares credit MVP caliber players for their offensive abilities but don’t assign enough credit to their defensive abilities.  I like James Harden and would probably vote for him if I had an MVP vote (I know, it’s surprising I don’t), but his lead in total win shares doesn’t really sway me, given that it’s largely fueled by offensive win shares.

Leave a Reply

Your email address will not be published.