James Harden vs. double-teams: Who's winning?

In a way, it all started because of Trae Young. Eight days before the Denver Nuggets tweaked the blueprint for guarding James Harden by double-teaming him everywhere, Young lit up Denver for 42 points in a rare road win for the Atlanta Hawks.

"After that game, I said never again will one player beat us," Nuggets coach Michael Malone told ESPN.

Malone rectified that against Harden on Nov. 20. Denver doubled Harden all over the floor, sometimes early in the shot clock and sometimes late -- random, last-second blitzes from odd angles. The Nuggets held Harden to 27 points on 16 shots and beat the Houston Rockets 105-95 -- a major reversal after Houston had won 10 of 11 prior matchups.

"It definitely worked," Malone said. "Will it work next time? We'll see."

Since that game, Harden has been doubled earlier in possessions and more aggressively than perhaps any player ever over any extended stretch. The LA Clippers trapped Harden even though they sport some of the league's best perimeter defenders. The Toronto Raptors and Dallas Mavericks dialed it up to 11; they had second defenders ready before Harden crossed the half-court line.

The most sophisticated tracking systems do not have language to evaluate what teams are doing to Harden. Second Spectrum, the gold standard, cannot directly tell you how many points per possession Houston scores when opponents trap Harden.

You can get that figure for Harden pick-and-rolls. That's a discrete play type these systems identify and log. There is no play type for "guy walking near half-court when two enemies bum-rush him." It's not an isolation. There is no screen. It's just a dude strolling with a basketball.

The only solution is to watch every Houston possession and tabulate what happens after the trap. The gurus at ESPN Stats & Information and I did that, starting with that Nuggets game and running through Sunday's action.

The results suggest the gambit isn't working, but evidence is murky for any sample size this small, and there are already hints of how the scheme might irritate the Rockets. Houston has scored 1.093 points per chance on shots that follow -- after one or more passes -- Harden dishing out of a double-team as he crosses midcourt, a mark that would lead all half-court offenses both for the season and since Nov. 20.

That dovetails with public data. Houston has scored 114.3 points per 100 possessions since Nov. 20 -- fourth overall, per NBA.com data. That number jumps to 116.6 with Harden on the floor. Houston in that span has averaged a mammoth 1.14 points per possession anytime opponents trap Harden on a pick-and-roll, according to Second Spectrum data.

"It doesn't bother us," Houston coach Mike D'Antoni told ESPN. "It's like playing against a zone. The more James sees it, the easier it's going to be. We are getting wide-open shots."

The main goal is obvious: Get the ball out of Harden's hands and make his supporting cast beat you.

Harden's shots and points are down since Nov. 20 in Denver, even though that nine-game stretch includes both 60- and 50-point outbursts.

Harden is a control freak. He arranges the chess pieces the way he likes. Trapping robs him of control. It steers his teammates into playmaking roles outside their normal skill sets -- a recipe for turnovers:

Houston's turnovers are up slightly since the trapping started, and opponents are punishing them more in transition. Houston's defense ranks just 18th in points allowed per possession over those nine games. They are 4-5.

The double-teams are a direct outgrowth of Houston exchanging Chris Paul for Russell Westbrook. Last season, Houston surrounded Harden with Clint Capela and three capable shooters. Trapping Harden meant a near instant catch-and-shoot 3.

Now, most of Houston's lineups feature two non-shooters -- Westbrook and Capela. It's easier to defend 3-on-4 below the trap if two of those four won't launch 3s. Perhaps three defenders can chase those four non-Harden Rockets until the defense resets itself.

Zoom further out, and the double-teaming hints at tensions surrounding the very nature of the sport. As Kirk Goldsberry outlined in his must-read book "Sprawlball," the overhaul in offensive strategy during the past decade stems from a simple fact: In a game of finite possessions, three is so much more than two that teams should do almost anything to chase that extra point.

Before 2017, teams pursued 3s using traditional tools: pick-and-roll, drive-and-kick, shooters fanning to the wing in transition. In 2017-18, Harden became the first player ever to consistently generate 3-pointers one-on-one -- no screen, no prelude -- with what has become his patented move: the step-back trey. Last season, he attempted almost 700 step-back 3s -- more than any other team.

Few conceived it was possible to build an efficient offense around isolation 3s. Harden proved he could make enough, and draw enough three-shot fouls, to do it. If it wasn't Harden, it was going to be someone else. The math made it inevitable that some player, sometime, would stretch the boundaries of the sport this way.

Teams are resorting to traps because they feel they have no alternative -- that the risk of fouling Harden is too high to defend him with one player. Some who worry the game has veered too far toward 3-pointers argue the only fix is to change the rules somehow. (This is the subject of Goldsberry's book.)

In a bizarre circle-of-life twist, trapping has reimposed stylistic normalcy upon the Rockets. This is the antithesis of one-on-one math ball, even if the result is a 3:

Houston's supporting cast is punishing the traps. Even if Westbrook refuses to shoot, he uses the open space to slice into the lane, draw help and kick the ball to someone else.

PJ Tucker has hit 44% from deep and 51% from the corners. Ben McLemore is reborn. Austin Rivers is 14-of-40 on 3s since the Nuggets game, and he has been a steadying two-way presence.

Going into Monday's game against the Sacramento Kings, other Rockets had hit an astonishing 15 of 21 3s directly off passes Harden tossed out of traps, per ESPN Stats & Info data. (This does not include conventional traps on pick-and-rolls -- just all-out two-against-one traps.) On those shots, the closest defenders were more than seven feet away. Four of Houston's 10 best games in terms of shot quality -- their expected effective field-goal percentage based on the location of each shot, the shooter and the proximity of the nearest defender -- have come since Nov. 20 in Denver, per Second Spectrum.

Their shooting around Harden should only improve. Danuel House Jr. missed three recent games, but he is back and shooting 43% from deep. Eric Gordon has hit 38% of his catch-and-shoot 3s as a Rocket. Coaches wonder if Gordon's return from knee surgery will mark the tipping point at which trapping becomes untenable.

"When we get him back, I don't see how it works," D'Antoni said.

"Them not having Gordon really helped," Malone said. "One less shooter."

Houston and Harden already are growing comfortable gashing the traps. Westbrook often flashes to the top of the arc, takes a bounce pass from Harden and attacks 4-on-3 up the gut. On Harden pick-and-rolls, Houston might station Westbrook on the wing as a release valve -- a classic D'Antoni technique known as "shorting" the pick-and-roll:

The Rockets vary Westbrook's positioning or have Harden attack from places where it is hard to trap -- the sidelines and even the corners.

They also know the best way to defang double-teams is to rush the ball up the floor before the defense can spring one. Westbrook has injected that sort of pace. His presence chips away at one trap-busting method -- long-range shooting -- but creates an entirely new one.

The Rockets also are smart about having Capela screech to a halt in semi-transition and screen for Harden near midcourt. Capela's defender -- the guy who is supposed to trap -- is often running toward the paint, unaware Capela and Harden have called an audible. If Harden gets downhill, it's over.

Interestingly, the Raptors were ready for all of this on Thursday in Toronto -- and it didn't really matter. Instead of doubling Harden with whatever defender happened to be nearby, they designated Fred VanVleet as Harden's second shadow. The Capela midcourt screens no longer worked, because Capela's guy wasn't supposed to double. It was VanVleet's job, and he was ready.

The Kings tried something similar Monday, only they sometimes had two extra defenders sandwich Harden -- and linger in something like a three-man trap long after Capela slipped back toward the rim:

Doubling Harden with smaller players has the downstream benefit of leaving larger, rangier defenders to protect the rim and fly at shooters.

When Harden tried to outrace the traps against Toronto, VanVleet was lurking:

Panicked defenses have swarmed Stephen Curry in transition like that for years. Other Golden State Warriors get dunks because two and three defenders surround Curry 30 feet from the rim. That is happening with Harden now.

Houston is also third in offensive rebounding rate since the Nuggets went all-in on trapping. Scrambling, zone-style defenses have a harder time locating opponents to box out. D'Antoni is urging guys to hit the glass.

"'We have been emphasizing it," D'Antoni said. "It's a constant."

The Rockets can stick Harden off the ball and let Westbrook initiate. Teams aren't going to double Harden away from the ball. Westbrook can kick to Harden and let him attack a jumbled defense. Either iteration of the Westbrook-Harden pick-and-roll -- rarely used now -- would raise complications for trapping defenses.

The traps might allow Harden to preserve energy. It's more taxing to drive into the teeth of the defense over and over than to give up the ball at half-court. Toward the end of the first quarter in Toronto, D'Antoni asked Harden when he wanted to rest. Harden replied he was still fresh because he hadn't done much beyond passing out of doubles, D'Antoni said.

And yet: The trapping probably isn't going away. Traps might expose open 3s for role players, but Houston is going to take a ton of 3s regardless.

"That's the way I looked at it," Toronto coach Nick Nurse told ESPN. "They are going to take a million anyway."

Maybe it's easier to swallow McLemore and Tucker beating you than to watch Harden get 50.

Doubling done well can minimize Harden's free throw attempts. Houston's opponents should face less foul trouble. Forcing the Rockets to pass more also bumps up the risk of live-ball turnovers. One seldom-discussed benefit of iso ball -- including Harden's brand -- is that it typically results in a low turnover rate.

Steals don't just represent empty trips for the Rockets. They are the most profitable way for opponents to start an offensive possession. Snare enough, and the math behind trapping might tilt toward those opponents.

Trapping also could take a mental toll on Harden. He wants to shoot. Even if trapping is the ultimate compliment, you can sense Harden doesn't respect it. Disdain drips from his fingertips when he finally acquiesces to the trap and passes. They are basketball eye rolls.

Harden already has evinced some frustration when his outlets don't go where he wants:

This was the very next Houston possession:

That's not a bad shot. It was very much an I'm shooting this no matter what because I'm sick of passing shot. There have been a few of those. Late in the Raptors game, Harden found himself open on the wing, caught a pass from Rivers, drained a rare catch-and-shoot 3, and flashed the prayer hands gesture at Rivers: Thank you for satisfying my craving. Harden was overeager against Sacramento Monday whenever he spotted single coverage, and careened his way into eight turnovers.

It is worth playing those mind games. There will be a quarter or a half when it works: When role players miss, Harden gets fidgety, and Houston goes haywire.

"He's gonna get frustrated," D'Antoni said. "If we go long periods where guys aren't making shots, he might try to do it all himself. And you couldn't really blame him."

Playoff teams -- the teams Houston will see when it matters -- are naturally more stocked with long, smart defenders who can survive 3-on-4 below the traps. They also will be better at disguising traps and finding the middle ground between pressuring Harden and overcommitting with an obviousness he diagnoses a mile away.

No one strategy is going to contain Harden. He is too good. The Rockets have just enough shooting, and they can always juice it by removing Capela and slotting Tucker at center. The best teams are going to toggle between lots of looks.

Those teams should keep the traps in their arsenal, regardless of the early returns.