October 15, 2018
Houston seem to be no noticeably worse when their All-Star center is out injured. Or are they?

The Houston Rockets are 22-10 (.688) when All-Star center Dwight Howard plays, but 14-7 (.667) when he doesn’t, thereby winning just 2% less of their games without him.

You may have heard that statistic cited elsewhere, as a measure of Howard’s apparent unimportance. Ostensibly, Dwight’s absence appears to not have had a huge impact on the Rockets’ performance – something that sounds pretty insane in isolation. On wins alone, Houston really have not dropped off much without Howard. And it is certainly not because of the play of his replacement, Joey Dorsey.

But the reality of the situation is vastly more complicated.

Firstly, based on margin of victory alone, the Rockets look far worse when Dwight has not suited up:

dwight_MOV

In games without Dwight, the Rockets are hovering just above average, outscoring opponents by 0.9 points. With lots of close wins no doubt being corralled in by James Harden’s MVP-level offense, the Rockets are winning games, sure. But they are not necessarily winning the, in a convincing fashion – based on that winning martin, one would expect such a team to only win roughly 53% of their future games.

Furthermore, when we look at the plus-minus data at Basketball-Reference.com, we can see that Houston is outscoring opponents by 7.2 points per 100 possessions when Dwight is on the floor, and only scraping by at +0.9 on all other possessions. That’s a difference of 6.3, a respectable number that makes Dwight actually look pretty good standing from the broad perspective of plus-minus.

But this, too, is unfortunately only part of the picture. It is fair to say that Dwight is not nearly as effective as he was.

[newsbox style=”nb1″ display=”category” cat=”539″ title=”More Player Breakdowns” number_of_posts=”2″ show_more=”no” nb_excerpt=”0″]

Anyone who has stuck their noses deep into advanced basketball statistics can tell you that Dwight Howard has not been the same on offense since 2011, and his decline this season represents one of the worst outputs of his career. Howard sports a -2.7 in offensive box plus-minus – an estimate of total offensive impact per 100 possessions from Basketball-Reference.com. He’s also posting his worst PER (18.2) since his rookie season in 2005 (17.2), down from 21.1 last year and down from 26.1 at his peak. Finally, his defense appears to have dropped, posting his worst defensive box plus-minus to date, just +1.4.

For a closer look at his decline, I pulled Dwight’s game logs from NBA.com, and calculated his Game Score for every game of his career. Game Score is a simple measurement of productivity on a scale similar to Points scored; as a point of reference, 40 is great, 10 is average.

dwight_gamescore

His peak 30-game average of 22.7 in the spring of 2011 has dropped all the way to 12.5. Just looking at the box score, Rockets-era Dwight has regressed to almost Laker-era Dwight levels, which is a pretty unwelcome thing to hear.

In reality, Dwight’s contributions to the Rockets have probably been better than his box score stats represent, as is the case with most defensive bigs. To blend both box-score stats and plus-minus stats, look no further than ESPN’s Real Plus Minus (RPM), the best publicly available metric for predicting NBA games. Here, Howard’s defensive RPM is sitting at +1.24 this year. Combine that with his offensive RPM, and his total contribution to his team’s Net Rating (that is, his total RPM) is about +0.34

Throwing in a player rated at +0.34 to an average lineup only nets them an extra 3 points per 1,000 possessions. Which is to say that by Dwight’s current level of play, he doesn’t add much to an average lineup. But there is no such thing as an average lineup. Who has been playing in Dwight’s absence the most this season?

To measure which players are replacing Dwight (or any player) in a lineup, I’ve built a stat that measures two things: what percentage of possessions a player is on the floor WITH Dwight, and what percentage without, via nbawowy.com. In the following chart, Without% minus With% = Replacement%. That is, how much more does a player get to play when Dwight is off the court than when he is on?

dwight_replacementpercent

As you can see, Dwight’s replacements are not better than Dwight, nor when he’s absent are those who he shares a proportionally higher amount of minutes with. In fact, they’re significantly worse for the most part. Dorsey, Black, K-Pop and J-Smoove are all hovering around or below replacement level, and the gap between the current fill-in Dorsey (-1.85) and Howard (+0.34) this season is enough to show us that Dwight is still an integral part of this team.

Combining box score and plus minus stats, we can see that although Dwight may be in decline, his presence is still something the Rockets need.

[newsbox style=”nb1″ display=”tag” tag=”Nathan” title=”More from Nathan Walker” number_of_posts=”2″ show_more=”no” nb_excerpt=”0″]
RELATED

Nathan Walker

I was born into a basketball loving family at the age of 0 in North Carolina. My mom a Carolina fan, my dad a State fan. Carolina was way better than State from the time I learned English until now, so I naturally sided with the Heels. In college I learned how to use Excel and discovered kenpom.com and I haven't closed Excel since.

View all posts

2 comments

Subscribe on YouTube

The Podcast

Subscribe on YouTube