For seventy years, basketball coaches believed in the midrange jumper. Fifteen feet from the basket, pull up off the dribble, knock it down. It was the signature of greatness — Jordan, Kobe, Wade. Coaching trees passed it down like scripture. Scouts valued it at the combine. Summer workouts were built around it. The midrange shot was what separated good scorers from great ones.
When Daryl Morey became General Manager of the Houston Rockets in 2006, he ran the numbers. The midrange jumper, it turned out, was the worst shot in basketball.
Not the lowest percentage — layups are lower percentage than corner threes. But the worst value. A midrange shot converts at roughly 40%, yielding 0.8 points per attempt. A corner three converts at 37% but yields 1.11 points per attempt. A layup or free throw is better still. Every possession that ended in a midrange jumper was leaving points on the floor relative to every other option available. The math was not subtle. It was not a close call.
The question wasn't whether the data was right. The question was whether the data ever reached someone with the authority to do something about it — and whether the decision environment made acting on it possible.
The information was already there
Analytics departments across the league had run similar calculations for years before Morey arrived. Box scores had always contained the raw material. Shot charts existed. Points-per-possession data was available. The information was not hidden. What was hidden was a path from that information to actual roster construction and play-calling decisions.
In most NBA organizations, analytics staff produced reports that went to front office executives who discussed them in meetings that were separate from the conversations coaches were having about schemes and lineups. Coaches made decisions based on film, feel, and decades of basketball intuition. Analysts made reports based on data. The two worlds rarely met at the level of a specific decision about a specific shot by a specific player in a specific moment of a game.
The information was there. The decision environment wasn't built around it.
What Morey actually changed
Morey didn't walk into the Rockets' practice facility and announce that midrange shots were banned. He changed the infrastructure of decisions. Roster construction criteria were explicitly tied to shot quality data — the question wasn't "can this player score?" but "does this player's shot selection create value?" Players who lived in the midrange were evaluated differently than players who attacked the basket and the arc. Contract offers reflected this. Drafts reflected this.
He also changed what coaches saw and when they saw it. Shot quality dashboards replaced generalized stat summaries. Defensive assignments were informed by data on where specific opponents shot poorly, not where conventional wisdom said to shade them. The feedback loop between information and decision tightened.
James Harden, who arrived in Houston in 2012, was a perfect test case for this environment. His natural game — draw fouls, attack the rim, pull up only from three — mapped precisely onto what the data said was valuable. In a different organization, a coach might have pushed him toward more midrange work to "create a more complete offensive game." In Houston, the environment supported what the data showed: let him do what the numbers say works.
The result
By 2018, the Rockets led the league in three-point attempts and had the best record in the NBA. Harden won MVP. The midrange attempts that other teams were still taking — out of habit, out of intuition, out of a coaching culture that had never been forced to confront the math — were showing up now as a measurable inefficiency that opponents could exploit.
Other teams eventually caught up. But they caught up only after they changed their decision environments too — after analytics staff were in the room when lineups were being set, after coaches were seeing shot quality data during timeouts instead of three days later in a film session.
The shot was always available. The players capable of eliminating it from their game always existed. What took twelve years to change was the path from the information to the person making the call — and the organizational structure that determined whether acting on the information was even possible.
The midrange jumper didn't die because the data got better. It died because the data finally reached the right people at the right time.