Keep Houston weird

The Houston Rockets are playing unconventional basketball, but their young core is blasting off.

Amen Thompson grabbed three offensive rebounds — a Rockets specialty — in Houston’s win at Cleveland on Wednesday. Getty Images.

Through an action-packed month, the Houston Rockets are blistering hot. At 10–3, with the NBA’s best offense, they’ve validated their loudest optimists and, so far, silenced their biggest skeptics, with their biggest test since their opening-night OT loss to OKC set to come tonight against Denver.

The Rockets were the NBA’s “surprise team” last season, blowing by Vegas’s preseason win total of 43.5 and climbing to second in a crowded Western Conference before getting bounced in a bruising seven-game series against Golden State. By my estimation, they’ve played their rebuilding hand quite optimally: tanking when necessary, stacking prospects, signing veterans, and capitalizing on their players’ development at the right moments.

Zooming outward, this is really as much as you can ask for since James Harden decided to force his way to Brooklyn five years ago. While some teams that lose a major star spend years bouncing around in NBA purgatory, the Rockets have parlayed a series of smaller moves into a relatively quick path to contention. Houston even made a mediumish-risk, highish-reward splash this summer, trading Dillon Brooks and Jalen Green for Kevin Durant. Durant hasn’t been spectacular, but he’s been healthy and good, and along with Alperen Şengün and Amen Thompson, gives the Rockets three players who rank in the top 35 in the league in EPM.

The Rockets have also embraced some unconventional tactics. Last season, they leaned heavily on Steven Adams’ offensive rebounding to generate second-chance points — the team grabbed an astonishing 43.9 percent of offensive rebounds when Adams was on the floor. This year, they’ve doubled down on size, playing Sengun and Adams in more minutes together. They’ve even supplemented the double bigs with Durant and Jabari Smith, creating further supersized lineups. The result is a 40.6 percent offensive rebound rate so far, a multiple-standard-deviation outlier relative to the rest of the league, and on track to be the highest rate in NBA history.

It’s hard to believe that just five years ago, the Rockets were running full-on Micro Ball with Russell Westbrook and James Harden. But reinvention has basically been their trademark for two decades.

Some of Houston’s experimentation has been born out of necessity. They don’t exactly have a surplus of shooters, and losing Fred VanVleet to an ACL tear this summer only made that more apparent. On paper, this should be the kind of team built to withstand a rash of injuries — after all, their depth last season was so strong that even No. 3 pick Reed Sheppard couldn’t find consistent minutes. But that depth was concentrated in the frontcourt — the guard rotation is trickier.

Reed Sheppard, sneakily the weirdest player in the NBA

Sophomore guard Reed Sheppard started the year a little flat, but he’s been on a tear over his last ten games — putting up 21 points per 75 possessions on a scorching 69.2 percent true shooting and an even 50 percent from three. Still, the questions about his size and his ability to generate his own offense haven’t gone away. Sheppard is small, but he was not drafted to be a traditional lead guard — he’s much more in the mold of an off-ball player who can add the elite shooting that the Rockets badly needed.

This isn’t coming out of nowhere: Sheppard shot 52.1 percent from long range in his lone season at Kentucky, the sort of eye-popping number that the editor of this story (Nate) initially assumed must be a typo. Although Sheppard mostly came off the bench, it wasn’t a small sample with 144 attempts from three and remarkable consistency across catch-and-shoot attempts, off-the-dribble 3s, guarded and unguarded shots, you name it.

Does the shooting efficiency come at a defensive price? Not necessarily in this case. In fact, much of the intrigue around Sheppard’s draft profile was because of his stocks (steals and blocks) rate of 7.1 percent, abnormally high for someone of his size and (lack of) speed.

The broader question is whether steal or block rates are good measures of defensive impact. There was a point not so long ago when steals seemed like the most important stat. An old FiveThirtyEight article by Benjamin Morris, for instance, found that steals were worth the equivalent of 9.1 points in predicting player impact.

How can steals have such a large impact? Some of it is direct: they end the opposing team’s offensive possession and start a fastbreak. Fastbreak possessions are considerably more efficient. Furthermore, many steals occur on passes that would lead to open shots, or in close-to-the-basket situations where the offense’s expected value is high. So, a steal can be worth several points on its own.

But also, steals are a proxy for overall, harder-to-measure, defensive peskiness and activity. Most of the time, players with high steal rates are also good defenders as measured by on-court/off-court impact stats. So, yes, there’s a lot of empirical evidence that steals are super valuable — but practically, there are some issues with using steals as a measure of perimeter defense, the same way blocks are often used to signal the best rim protectors.

For example, in 2016, Stephen Curry led the league in steals — but how high did he actually rank in defensive impact within the Warriors’ rotation?

Steph certainly wasn’t as impactful as Draymond Green or Andrew Bogut — the Warriors’ defensive anchors — and probably not as valuable on that end as Klay Thompson or Andre Iguodala, who regularly took the tougher backcourt assignments. Curry was typically matched up with off-ball shooters, which freed him to roam passing lanes and rack up steals, even if he wasn’t necessarily the player creating the most defensive “value”.

According to EPM, Sheppard is at −0.3 points per 100 possessions defended so far on the season. EPM is measured relative to league-average — so Reed’s been average, basically. That might feel a little disappointing when you consider his prodigious steal rate (3.3 steals per 100 possessions). But average is more than acceptable for a short, relatively unathletic 21-year-old guard, a profile that typically implies a (massive) defensive liability.

Sheppard is filling big shoes. Why did analytics geeks, including me, think VanVleet would be so hard to replace? It’s a bit of a weird question. FVV last made an All-Star team in 2023.

The easiest answer is that he did some very important “small” things for this very unconventional team. VanVleet has been a positive defender for his entire career. His ability to navigate screens and play up to the ball handler without losing any space has made him one of the best point-of-attack defenders¹ in the NBA for a decent stretch of his career, and he’s mostly aged gracefully in that regard.

Sheppard, on the other hand, is playing primarily off the ball defensively because he’s neither very strong nor laterally quick. Still, although he’s limited as a point-of-attack defender and exploitable due to his size, he does have good hands and can play within the gaps while passes are attempted to shooters.

I wonder, though: can Sheppard ever advance from average to being a good or even great defender? It probably depends on his system. It’s telling that the Rockets have already logged 221 possessions of zone coverage through 13 games — the most in the league by a wide margin. (This is becoming a theme: if you look up a given statistical category, the Rockets are probably near the very top or the very bottom.) The zone lets Houston lean more heavily on its double-big lineup. Without VanVleet at the point of attack, they’ve shifted toward a more balanced defensive scheme instead of letting opponents hunt individual matchups. Zones have traditionally been rare in the NBA because they concede more open threes, but they also offer a team-based defensive approach that helps Sheppard play directly to his strengths as a turnover creator.

The curious case of Amen Thompson

The Rockets also boast an electrifying talent in Thompson, who, by my estimation, is the most athletic player in the NBA.

This is not my most controversial take: Thompson is probably the only player in NBA history who can putback dunk his own miss from a standstill and sprint at a top speed of 23 miles per hour. P3, the sports performance firm, categorized him as a “Specimen” — a label reserved for the NBA’s elite athletes. Still, even within that group, Thompson has posted outlier numbers. His athleticism isn’t just raw or flashy; it’s functional. He uses it to sky for rebounds and cover ground without the ball in his hands. That last part —his ability to consume space — is arguably the defining trait of Amen’s game.

We talked a bit in my last piece about “gravity”. So let’s return to the question of why spacing has become so important in the NBA. Many casual fans claim that it’s just because 3-pointers are “worth” more than two-pointers, but let’s think a little bit about that. Is a team’s main objective on every possession to shoot a three? What is the most efficient shot in the game?

Unsurprisingly, rim attempts are still the most valuable shots in basketball. The more important trend, however, is a spike in rim efficiency over the last 10 years.

Raise your hand if you’ve been in this situation. You’re peacefully trying to watch a basketball game, and that one guy — maybe a couple beers down at this point — goes on a tirade about how the NBA is getting softer:

“All players do is shoot threes now! None of them want to get to the rim! Back in the 90s, players played defense!” — Random Guy In Bar²

A simple data dive dispels most of these narratives. We see that rim attempts have been fairly stable since 2001. It’s really just long midrange jumpers that have declined.

It turns out rim efficiency and 3-point attempts are highly correlated, and for good reason: the function of spacing isn’t just that 3 > 2. Rather, spacing makes offensive interactions, such as the pick-and-roll, less congested. In theory, the more functional 3-point shooting a team has, the more space there is to operate, and the offense as a whole becomes more efficient.

These trends also shed light on why many fans think defense is a thing of the past. Offense seems to be “winning” because reliable 3-point shooting now stretches defenses to their breaking point. In terms of pure talent, defenses are probably better than they’ve ever been — it’s just that modern rules and offensive spacing have made it harder for even elite defenders to bend the game in their favor.

I admit this was a long detour, but it leaves us with an essential question for a player of Thompson’s archetype. What happens when a non-big has no hope of becoming a good shooter?

In the mid-2010s, Tom Haberstroh, searching for a way to quantify this, created “respect rating”. Respect rating combined two measures, gravity and distraction, to capture the degree of defensive attention a player commands. In essence, it quantified floor spacing. High scores usually belonged to shooters who stretched out defenses, while most low scores went to those who didn’t draw extra coverage.

The results, for the most part, were not surprising. Stephen Curry, Kyle Korver, Klay Thompson, and other shooters were consistently at the top. However, one name popped up that didn’t belong with the rest: Dwyane Wade.

Why did defenders pay attention to Wade so much, even if he was a pretty famous example of a poor 3-point shooter? You would think, based on some of the narratives about the modern NBA, that non-shooters can’t function outside of a few offensive actions. In D-Wade’s case, it was his cutting. He had a knack for executing his bursts to the basket at the right moments — sometimes unseen and sometimes just too fast — and that required his defender to stay attached to him.

Basketball is a game of attention, and while shooting is a great way to garner it, it’s not a universal requirement. For example, in this play, the Clippers run a pick-and-roll action for Ivica Zubac. Christian Braun camps himself in the paint while Kris Dunn sits at the 3-point line (Dunn is a 32.7 percent lifetime 3-point shooter, despite a diet of mostly wide-open shots). This means the Nuggets have a clear advantage, right?

Well, it turns out the Clippers can scheme around this — Dunn times the cut perfectly, and even if he doesn’t receive the ball, the Clippers are left with an advantage. But Dunn is an okay cutter at best. What happens when the cutter is bigger, faster, and leaps over everyone? When that cutter is Amen — 44-inch vertical, 7-foot wingspan, elite contact finisher — the geometry becomes even more warped. Thompson isn’t D-Wade (yet), but the impact metrics have him as a significant net positive on offense (not to mention a DPOY contender) largely because of that cutting.

Şengün is pressing all the right buttons

Part of Thompson’s upside depends on the personnel around him, particularly Şengün. I’ll be up front: I came into this season as a Şengün skeptic. Some of this was perhaps grounded in aesthetic bias: I’m not opposed to post players so much as doubtful of their ultimate ceiling.

Furthermore, the “Baby Jokić” comparisons weren’t exactly positive to me. Nikola Jokić is a player who fundamentally shouldn’t work — but he’s synergized skills that stretch the imagination. When most fans think about Jokić, they’re probably reminded of the ridiculous post-up scoring and passing. But is that the only way to describe his uniqueness? Geometrically, Jokić also provides value from his ability to occupy areas other than the low block. It’s not just that he can score from the post, it’s that he allows possessions to flow downstream without disrupting the tempo of his teammates’ cuts.

This ability has only manifested in a few players in NBA history — Jokić, Chris Webber, and Vlade Divac are among them. Despite being a good post player, Şengün had struggled with the rest of that skill set. Last season, he didn’t grade out as efficient across most play types, especially those that depended on athleticism. As a roll man and cutter, he was one of the worst centers in the NBA:

Granted, the sum was somewhat greater than the parts: the Rockets had a +7.5 net rating with Şengün on the floor last season. But this year, that’s exploded to a staggering +13.4 with him on the court.

Is this early-season noise or the result of real development? One notable finding is that Şengün’s share of possessions as the roll man in PNRs has declined significantly — from 18.2 percent in the 2024-25 season to only 6.4 percent now. While some of this is the result of their offseason moves³, it’s possible the Rockets are pivoting entirely towards a Şengün-ball offense.

And defenses have reacted with more attention. Last season, the defense committed to or double-teamed Şengün on 65.1 percent of his post-ups⁴ — a solid rate, but nothing in comparison to players like Jokić or Giannis Antetokounmpo, who saw rates above 75 percent. This season, Sengun’s double-or-commit rates have skyrocketed to 80 percent in an admittedly small sample size. The efficiency of his postups and isolations has increased as well, suggesting real development.

What’s the final verdict?

In some ways, the story in Houston isn’t that complicated: young players typically improve, whether incrementally or through bigger leaps. As we’ve covered, Şengün, Thompson and Sheppard are unusual players — but they’re increasingly making the most out of their roles. The odds for that sort of positive development are much more favorable for younger players, but it also speaks to the good work of the front office and coach Ime Udoka.

There are also more red flags here than for a typical 10-3 team. It’s early in the season, and the league will find counters to the Rockets’ unusual tactics — if not now, then in the playoffs. Nobody on the roster is a safe bet to make one of the three All-NBA teams. VanVleet’s injury is obviously not ideal, and Houston might be best served by taking a more egalitarian approach to ball-handling this year. Letting the offense take shape organically while players continue to grow into their roles feels like the right move for a roster still figuring itself out.

Even so, this is one of the league’s best teams. I don’t think they’re equipped to beat Oklahoma City just yet — to be fair, who is? But they’re moving toward the inner circle of contenders, and there is the shape of even better things to come — a leap that should have H-Town buzzing.

1. Point-of-attack defenders aim to stifle the ball-handler and, in turn, prevent the offense from flowing. Jrue Holiday, Mikal Bridges, and Jaden McDaniels are examples.

2. Or a studio analyst for a very famous basketball halftime show.

3. The departed Jalen Green spent roughly 42 percent of his possessions last season as a pick-and-roll handler.

4. Post-up possessions and PPP include passes and shots made out of post-ups.