This story appears in ESPN The Magazine's May 29 issue. Subscribe today!
THE SPURS HAVE snuffed out Houston's season, but at least we can say this about the Rockets: They were true to themselves until the bitter end. Houston rode out one of the most thoroughgoing analytics experiments in the history of sports, in which the Rockets took 46.2 percent of their shots from 3-point territory this season, the highest proportion in NBA history. Daryl Morey crafted that strategy with a deeper playoff run in mind; too bad his mad bombers won't get another chance to face Golden State. "We want to win the title, and obviously that's probably going through the Warriors at some point," the Rockets' GM told SiriusXM NBA Radio in February. "We absolutely figure the only way we're going to beat them is with a barrage of 3-pointers."
Morey and coach Mike D'Antoni were right to dial up the bombs. Launching 3s made Houston's scoring not just more prolific but also more variable -- a valuable trait against top opponents. They might be gone, but other teams -- especially the Cavaliers, who ranked second in the league in percentage of field goal attempts from 3-point range -- should be paying attention to the Rockets' strategy. But just how crazy does a team have to play to beat Steph Curry & Co.?
Houston launched a veritable hailstorm this year. James Harden, Eric Gordon, Trevor Ariza and Ryan Anderson all ranked among the top 11 in the NBA in 3-point attempts, the first time one team has had four players take more than 500 apiece. Amazingly, Harden also drew more fouls on 3-point shots and assisted on more trifectas than anyone in the league. That's all part of playing smarter: By taking 81.8 percent of their shots from the restricted area near the hoop or beyond the 3-point arc, the highest proportion in the NBA by nearly 10 percentage points, the Rockets scored 114.7 points per 100 possessions, second best in the NBA.
But Morey was on to something deeper than efficiency, with lessons for any team that finds itself facing the Warriors. If your club scores 105 points a night while your opponent scores 115, there's not much you can do to close the talent gap. But you can increase your scoring variance, so that you put up, say, 90 to 120 points per game instead of 100 to 110. You're more likely to get blown out -- but you have a better chance of grabbing a win against a superior foe. Relying on long-distance shooting is a crucial way to go high-risk/high-reward. It's quite possible for a team to go 6-for-37 on 3-pointers in one game, then 22-for-50 in the next; in fact, the Rockets did just that in Game 5 of their series against Oklahoma City and Game 1 vs. San Antonio.
The idea that inconsistency helps underdogs was first explored by statistician Dean Oliver in his 2004 book Basketball on Paper. Over the past decade, The Mag's Giant Killers project has found that NCAA tournament Cinderellas tend to deploy high-variance strategies. They work in the NBA playoffs too. To analyze factors that might lead lower seeds to overperform, I used regression analysis to study all playoff games over the past three years between teams separated by at least four points, according to Basketball-Reference.com's Simple Rating System. I found that for every 1 percentage point an underdog increases the proportion of its shots devoted to 3s, its winning percentage increases by an average of 2.2 percentage points -- even though its overall scoring stays about the same. Again, that's because it's more likely to eke out victories while suffering bigger losses.
Given the scoring stats, Houston would have had about a 30 percent chance of beating Golden State in any given game. So if the Rockets could have boosted their 3PA/FGA by 10 percentage points, to about 48 attempts per game, they'd have been near even money.
It's a nice theory, I have to say, but this ain't March Madness. An opponent would have to shoot 3s at a record pace for a week or two to compete with the Warriors, not just in a single game. And there's a tipping point at which defenses overload the perimeter. Even these Rockets, who posted far more of the top bombs-away games in NBA history than any other team, would have had less than a 1 percent chance of posting 48 3PA four times in a seven-game series; for the Cavs, who shoot fewer 3s to begin with, the odds are infinitesimal.
So the Spurs -- built on the opposite model of patiently moving the ball to create efficient shots from various distances -- might have saved the Rockets some trouble. And while the Cavaliers have more room to follow Houston's blueprint, by the metrics, they would be deeper underdogs to begin with. Against the Warriors, the numbers say you need more than an outside shot to have more than an outside shot.