If you watched an NBA game in 2005 and then watched one today, you’d barely recognize the sport. The rules haven’t changed much. The court is the same size. The basket is still ten feet high. But the way teams play — where they shoot from, how they space the floor, which players they value — has been completely transformed by data analytics.
This isn’t just a story about numbers replacing instinct. It’s about how a mathematical insight reshaped a billion-dollar sport from the ground up, changing everything from how coaches draw up plays to how general managers build rosters worth hundreds of millions of dollars.
The Moreyball Revolution
The story starts with Daryl Morey, who became general manager of the Houston Rockets in 2007. Morey, a MIT Sloan graduate with a background in data science rather than basketball, looked at shot data and saw something that old-school basketball minds had been ignoring for decades.
The math was brutally simple. A three-point shot is worth 50% more than a two-point shot. So even if a player makes three-pointers at a lower percentage than two-pointers, the expected value per possession can be higher from beyond the arc. Specifically, a 35% three-point shooter generates 1.05 points per attempt, while a 50% mid-range shooter generates only 1.00 point per attempt. That five-hundredths of a point difference, compounded over thousands of possessions across a season, translates to wins.
Morey’s conclusion was radical at the time: the mid-range jumper — the bread and butter of basketball legends from Michael Jordan to Kobe Bryant — was the worst shot in basketball. Teams should either shoot threes or drive to the basket for layups and free throws. Everything in between was a mathematical mistake.
The basketball establishment scoffed. Analysts mocked the approach. Then the data started speaking for itself.
The Three-Point Explosion By The Numbers
In the 2004-05 NBA season, teams averaged 16.0 three-point attempts per game. By the 2024-25 season, that number had ballooned to over 37 attempts per game. Some teams regularly launch 45 or more threes in a single game — numbers that would have been unthinkable two decades ago.
The Houston Rockets under Morey led this charge, but the Golden State Warriors proved the concept could win championships. Stephen Curry, Klay Thompson, and the Warriors’ system combined elite three-point shooting with ball movement and spacing to create one of the most dominant dynasties in NBA history, winning four championships between 2015 and 2022.
The mid-range shot didn’t just decline — it nearly went extinct. In 2005, mid-range jumpers accounted for roughly 40% of all field goal attempts. By 2025, that number had dropped below 15%. An entire category of basketball skill — the pull-up eighteen-footer, the elbow jumper, the baseline fadeaway — became strategically obsolete almost overnight.
Players who built their careers around mid-range mastery found themselves analytically devalued. The DeMar DeRozan type — a gifted scorer who lived in the mid-range — went from All-Star to “inefficient” in the spreadsheets, even as they continued to fill stat sheets by traditional measures.
Beyond Shot Selection: How Data Reshapes Everything
The analytics revolution didn’t stop at three-pointers. Data has infiltrated every corner of basketball strategy.
Defensive schemes have been transformed by tracking data. Every NBA arena now has cameras that record the position of every player and the ball 25 times per second. This spatial data reveals defensive patterns invisible to the naked eye — how quickly a team rotates to help, which defenders contest shots most effectively, where gaps open up in zone coverage.
Teams use this data to prepare for opponents with a precision that would stun coaches from previous eras. They know that a specific player tends to drive left 73% of the time. They know which pick-and-roll actions generate the best looks. They can quantify which defensive matchups cause the most problems and adjust rotations accordingly.
Roster construction has shifted dramatically. The premium on “3-and-D” players — wings who can shoot threes and defend multiple positions — has skyrocketed. Traditional big men who operate primarily in the post and can’t shoot from outside have seen their market value collapse. Centers who can stretch the floor with three-point shooting, like Nikola Jokić, have become the most valuable players in the league.
Load management is another analytics-driven innovation. Teams now monitor player fatigue through biometric data, tracking everything from sleep quality to muscle recovery rates. The controversial practice of resting healthy players during regular-season games — pioneered by the San Antonio Spurs with Tim Duncan and later adopted league-wide — is grounded in data showing that fresh players perform significantly better in the playoffs, where games matter most.
The Backlash and the Pendulum
Not everyone loves what analytics has done to basketball. Critics argue that the three-point revolution has made the game repetitive and less aesthetically pleasing. When every team runs similar offensive schemes — space the floor, shoot threes, drive and kick — games can start to look the same.
There’s a legitimate aesthetic argument. The mid-range game produced some of basketball’s most beautiful moments. Jordan’s fadeaway over Byron Russell. Kobe’s footwork in the post. Dirk Nowitzki’s one-legged step-back. These shots were art, and analytics essentially argued that art was inefficient.
Some teams and players have pushed back. The mid-range has seen a modest revival in recent seasons, driven by players like Kevin Durant and Kawhi Leonard who are so efficient from mid-range that the math actually works in their favor. The analytics community has evolved too, recognizing that shot quality depends on context — an open mid-range jumper from an elite shooter can be a better look than a contested three from a mediocre one.
Coaches have also gotten smarter about defending the three-point line, which has slightly reduced three-point efficiency league-wide. As defenses adapt, offenses may need to re-diversify. The pendulum never stays at its extreme.
What Comes Next
The next frontier of basketball analytics isn’t about shot selection — that revolution is largely complete. Instead, teams are investing in predictive modeling and real-time decision-making tools.
Imagine a coaching staff that receives real-time recommendations during games based on live data analysis. “Run action X against this defensive set — it’s generated 1.2 points per possession in similar situations this season.” Some teams are already experimenting with this kind of in-game analytics support, and it will only become more sophisticated.
Player development is another growth area. Analytics can now identify specific skills that a young player should develop to maximize their career value. A rookie who shoots 30% from three but has good mechanics might receive a personalized development plan designed by data scientists to push that number to 37% within two seasons.
Draft evaluation has been revolutionized too. Teams combine traditional scouting with statistical models that predict how college and international players will translate to the NBA. These models aren’t perfect, but they’ve reduced the frequency of catastrophic draft busts by identifying red flags that the eye test alone might miss.
The Bigger Picture
Basketball’s analytics revolution mirrors what’s happened across professional sports — from baseball’s Moneyball era to soccer’s expected goals (xG) models to football’s fourth-down decision-making. The common thread is the same: data reveals truths that tradition and intuition miss, and the teams that embrace those truths gain a competitive edge.
But the best teams — the ones that sustain success — don’t choose between analytics and traditional basketball wisdom. They integrate both. The numbers tell you what to do. The coaches and players figure out how to do it. And occasionally, a transcendent talent like Curry or Jokić comes along and does things the models didn’t even know were possible.
That’s the beauty of sports in the analytics age. The data makes the game smarter, but it can’t replace the human element that makes it exciting. And as long as that tension exists, basketball will keep evolving in ways nobody — not even the algorithms — can fully predict.