The Houston Rockets were an offensive juggernaut last season.
Scoring 108.6 points per 100 possessions last season, the Rockets held the fourth-best mark in the league. They ranked third in effective field-goal percentage, first in free-throw rate and seventh in offensive rebound percentage.
Other than a high turnover rate, it was difficult to find fault with anything they did on offense—yet somehow they left us wanting more.
Both are phenomenal offensive players, each with well-defined strengths and weaknesses. Although their games grew more intertwined through the season, it still felt like they were often stepping on each other's feet or dramatically clearing out of the other's way.
The most obvious disconnect involves Howard's post-up game.
According to mySynergySports (subscription required), Howard scored just 0.77 points per possession on post-ups last year, good for 128th in the league.
However, the Rockets offense accommodated 785 possessions finished by Howard in the post during the regular season and playoffs, or just over 10 per game. That worked to about about 8.1 percent of their total offensive possessions.
Remember that we're just talking about possessions where Howard either attempted a shot, turned the ball over or was fouled while posting up. Other good things often happened when Howard posted up, such as manipulating the defense to create openings for other players.
Still, using 8.1 percent of your total offensive possessions on an approach with such a low level of efficiency seems counterproductive.
That impression becomes even stronger when you consider Howard's prowess in the pick-and-roll. When he finished as the screener in the pick-and-roll, Howard scored an average of 1.31 points per possession. That was the third-best mark in the league last season.
Put that together with Harden's elite abilities with the ball in his hands, and the recipe for turbo-charging this offense seems to clearly be more Harden-Howard pick-and-rolls.
When they are working in the pick-and-roll together, both players are at their best—both are in motion. Howard isn't clogging up driving lanes for Harden, and Harden isn't standing on the perimeter with his hands on his shorts, waiting for the ball to come back out to the perimeter.
If you need an example of just how devastating this combination can be, check out the video below:
Harden's ability to penetrate demands aggressive pursuit by his defender and a significant hedge by Howard's.
Once Howard catches the ball, his strength and agility are far too much for the back line of the defense. If the Miami Heat play Harden any less aggressively, he's charging into the teeth of the defense and getting to the basket or drawing a foul.
Simply put, the Howard-Harden pick-and-roll makes the best use of both players' offensive abilities.
Position shooters like Patrick Beverley and Trevor Ariza around the one-two punch, and these pick-and-rolls become even more deadly.
Unfortunately, just saying the Rockets should run more Howard-Harden pick-and-rolls isn't as simple as it sounds. Howard's preference for being stationed in the post is well-known.
He has often restated this preference as way to help get his teams back on track, which he did to Jonathan Feigen of the Houston Chronicle in the middle of their playoff series with the Portland Trail Blazers last season:
We have to play inside out, play their bigs and make it a long night for those guys. I have to demand the ball, get it and go to work.
We have to go right back at him. You have to make him play defense and make him use his energy on defense. Make him have to run around and guard.
His teammates seemed to be on the same page as well. Chandler Parsons echoed roughly the same sentiments to Feigen.
"We have to dump the ball down for him. He has to make plays. He’s had a lot of success against (Robin) Lopez. We have to keep feeding him the ball. That doesn’t mean we have to slow down. We want to get out in transition and run," he said.
The idea that Howard, his teammates and, at some level, the coaching staff seem to be exploring is the way Howard's post-ups can distort the defense. The Rockets certainly got better at this throughout the season, finding new and different ways to build motion around him.
Drew Garrison of SBNation broke down some of this improvement in early March:
Howard pulls defenders in like a tractor beam. Harden and Parsons have space to operate and turned heads to zip by. Parsons has been the greatest beneficiary, slicing through defenses and getting to the rim throughout the season. Houston's philosophy is simple: points in the paint, open threes and free-throws ...
The offense has been altered to give him [Howard] his low-post touches, but isn't dependent on him sinking a high percentage of his attempts. The threat of his post scoring is enough.
The last sentence of that quote is key. It is the threat of his post scoring that bends the defense.
Take the Howard post-up below, for example. You can see how the offensive motion stops as soon as Howard catches the ball. Harden's man, Dwyane Wade, is cheating in for a soft double on Howard.
But by the time Harden really gets open, Howard has already spun baseline, beginning his move:
The thing is, the way the offense becomes static really limits the options here. There may be the possibility for a cut on the weak side, but essentially this offensive possession is winnowed down to Howard and Harden.
There is a moment when a kick-out from Howard leads to Harden with the ball in his hands and plenty of space to attack. That sort of scenario is generally much likelier to lead to a positive outcome than Howard spinning baseline for a jump hook.
As Garrison noted, the opening for Harden created by the soft double-team only occurs because the Heat know Howard is active and likely to shoot when he catches the ball on the block.
If Howard stops attempting so many shots in the low post, opponents stop doubling and the hyper-efficient shots for teammates that come off of those possessions evaporate.
Although it isn't quite as cut-and-dry as the simple difference between Howard's 0.77 points per possession on post-ups and 1.31 points per possession as the screener in the pick-and-roll, the Rockets offense seems to get better outcomes when Howard is in motion at the offensive end.
The Rockets are one of the most statistically savvy teams in the NBA and are almost certainly aware of this.
However, they need to live with a certain amount of Howard post-ups—both to keep him happy and keep the defense working honestly against all of the different offensive scenarios the Rockets throw at them.
Ultimately, the Rockets need both pieces of their offense—the post-ups and the pick-and-rolls. The key is the balance. As you can see from the graph below, that's where the Rockets can make some improvements in their offensive efficiency:
The graph shows the ratio of post-up possessions to pick-and-roll screener possessions for Howard across the last five seasons.
Last year with the Rockets, Howard finished 7.5 post-up possessions for every possession he finished as the screener in the pick-and-roll. Although that was lower than his peak in Orlando, it was significantly higher than his two previous seasons—his last with the Magic and his one season with the Los Angeles Lakers.
If the Houston Rockets want to push the boundaries of their offensive efficiency even further, the task is not enormous.
They don't have to redesign everything or try and talk Howard into abandoning his interior possessions. Houston just needs to work on subtly shifting the balance. It may seem like a scary proposition, especially considering how frustrated Howard was with his role in Los Angeles, but the coaching staff wouldn't be asking him to do something he hasn't done before.
The Rockets are a process-oriented organization with the patience and foresight to see how small changes at the margins can have a big impact when stretched across an entire season.
If anyone can gently nudge Dwight Howard towards a more advantageous offensive distribution while still maintaing the integrity of the system and his commitment to its principles, it's Houston.
Statistical support for this story from NBA.com/stats.