From Houston's 7-for-44 Meltdown to Boston's 2026 Title: How the NBA Rewrote the Three-Point Equation

On May 28, 2026, at Toyota Center, a team that had won 65 regular-season games missed 27 consecutive three-point attempts. The Houston Rockets finished Game 7 shooting 7-of-44 from beyond the arc — 15.9% — in a 92-101 loss to the Golden State Warriors. No team in NBA playoff history had ever taken so many threes in a decisive game.

I stayed in a small bar on 8th Street in Manhattan until nearly two in the morning, rewinding the tape over and over. Not to find someone to blame. I wanted to know one thing: were those 44 attempts the product of a broken system, or the result of a loaded die?

By the next morning, the country had its own answer. Charles Barkley called it kids' basketball. The big sports shows repeated one line in unison: you cannot win a championship shooting threes. That night became the weapon of every analytics sceptic for the next half-decade.

Six years later, the Boston Celtics won the 2026-24 NBA title while averaging 42.5 three-point attempts per regular-season game — the highest figure ever recorded for a champion. They won 16 of 19 playoff games. They beat the Dallas Mavericks 4-1 in the Finals.

So the old story is dead? Not quite. It simply changed shape, and its new shape is far more uncomfortable than either side wants to believe.

Fifteen years of drift

In 2026-12, the average NBA team attempted 18.4 threes per game. By 2026-24, that number was 35.1. In twelve seasons, the league's three-point volume nearly doubled — and not because players suddenly developed stronger wrists.

A two-point shot converted at 45% yields 0.90 points per attempt. A three-point shot converted at 33% yields 0.99 points. In other words, a bad three-point shooter is still more efficient than an average two-point shooter. Daryl Morey saw that in Houston back in 2026 and turned it into a system: essentially erase the mid-range, shoot only threes and shots at the rim. People called it Moreyball, and eventually the entire league copied it.

The second consequence gets far less attention. When every team optimises the same equation, the marginal advantage disappears. Three-point shooting stops being a strategy and becomes the minimum condition for existing on the floor. By 2026, a team that could not shoot threes was not losing because it was outdated — it was losing because it had been removed from the geometry of the game.

And right there, another variable appears: variance.

The equation Houston solved correctly

In 2026-18, the Rockets attempted 42.3 threes per game and won 65 regular-season games. They built their roster around James Harden and Chris Paul, two players capable of generating an open shot from anywhere on the floor. Their process in Game 7 did not break. They still generated 44 three-point attempts, more than their own season average.

If each attempt is treated as an independent trial with a 35% success rate — their own season average — the expected number of makes is 15.4, with a standard deviation of roughly 3.2. Making only 7 sits about 2.7 standard deviations below expectation. The probability of such a night is under 1%. A run of 27 consecutive misses, considered on its own, carries a probability of roughly 1 in 112,000.

This is where I have to be blunt as a man who counts for a living: no system survives a night like that. But that is exactly why the real story lies elsewhere.

Chris Paul sat out Games 6 and 7 with a hamstring injury. The Rockets lost their second ball handler, lost their ability to create disruption in the second half, and forced Harden to carry a creative load beyond any reasonable threshold. The dreadful three-point percentage was a symptom. The cause was that Houston ran out of options once its second engine was on the bench.

Numbers do not score points, but numbers are quietly rewriting history: Game 7 in 2026 did not prove that three-point shooting fails. It proved that a system with a single engine dies when that engine dies.

What Boston did differently

The 2026-24 Celtics shot more threes than the 2026-18 Rockets. The difference lay in the quality distribution of those attempts — and on the other end of the floor, where analytics sceptics rarely bother to look.

That season Boston led the league in defensive rating. They had Derrick White, Jrue Holiday, Jaylen Brown and Jayson Tatum — four players capable of defending at a high level at any position. Al Horford and Kristaps Porziņģis protected the rim and allowed the perimeter defenders to press, to switch, to accept risk. In other words, they did not build a team to shoot threes. They built a defence good enough to keep generating open threes at the other end.

The point matters enough to isolate as a principle. Boston's three-point volume was not the cause of the championship; it was the consequence of a team controlling the tempo of the game at both ends. Across 19 playoff games, the Celtics won 16 and lost 3 — an 84.2% win rate, higher than any champion since the full 16-team playoff format was established. In the Finals they beat Dallas 4-1.

The most striking detail lies elsewhere: Jayson Tatum shot just 38.8% from the field in the Finals and 26.3% from three. A team won a championship while its number one star shot that badly. Jaylen Brown took Finals MVP. For any team dependent on one individual, that script ends in defeat. For Boston, it ended in a trophy.

That is the entire difference between 2026 and 2026. Houston had one engine and lost it. Boston had six and needed only two of them running properly in June.

The 2026 title and the return of defence

Had the story stopped at Boston, people would have concluded that volume three-point shooting is the road to a championship. The following season broke that conclusion.

On June 22, 2026, the Oklahoma City Thunder claimed the first title in the franchise's Oklahoma City era, beating the Indiana Pacers 103-91 in Game 7. The Thunder finished the regular season 68-14. Shai Gilgeous-Alexander led the league in scoring at 32.7 points per game and took both regular-season MVP and Finals MVP.

What matters is that Oklahoma City won with a completely different formula to Boston. They led the league in defensive rating, in steals and in forced turnovers. Their three-point volume was not among the league leaders. They won by suffocating their opponents' space, converting defence into points at the other end, and letting Shai generate rim pressure and free throws in the closing minutes.

Two consecutive champions, two different formulas. That is the single most important data point of the whole period, and it demolishes both versions of the simple story: shooting a lot of threes guarantees nothing, and shooting few guarantees nothing either.

Possession control is an illusion; points are the bare truth — but where those points come from is the question worth asking.

The four factors and the death of empty stats

Dean Oliver laid out the four-factors framework in the early 2000s, and NBA analytics departments still rank teams on exactly those four variables: effective field goal percentage, turnover rate, offensive rebounding rate and free-throw rate. Three-point attempts are not on that list. They are merely one input into eFG%. A team shooting 40 threes at 32% posts a lower eFG% than a team shooting 30 threes at 39%. Volume alone says nothing. This is why I am always wary of stat tables that count attempts while ignoring shot quality, shooter quality and defensive context.

The same problem appears at the individual level. A player averaging 25 points on a 20-win team can look superb on a box score and completely meaningless in the standings. Based on my own experience tracking games across many seasons, this group of players tends to shoot heavily in games already decided, shoot heavily in the fourth quarter when the team is down double digits, and shoot very little when the game is still balanced in the final two minutes. I do not need a complex model to spot that; I only need to read the box score quarter by quarter.

At the operational level, teams have recognised the gap between the media narrative and actual value. On February 1, 2026, Luka Dončić moved to the Los Angeles Lakers in a three-team deal, with Anthony Davis heading the other way to the Dallas Mavericks. A trade like that cannot be explained by scoring averages. It is decided by age, contract status, physical foundation and internal data no analytics department publishes. The distance between what fans see and what front offices see is precisely where analytical work still has room to dig.

Where I could be wrong

I have spent half a decade arguing that three-point shooting is misunderstood. But there are at least four points that force me to doubt myself.

First, Boston's 2026 championship path was far softer than it looks. Miami was without Jimmy Butler. Cleveland was without its strongest line-up for stretches. Tyrese Haliburton went down injured during the Eastern Conference Finals. A title won through four weakened opponents is still a title, but it is not strong enough experimental evidence to settle a philosophy.

Second, I have just used one game to refute an argument built on another game. Game 7 in 2026 proves nothing in the long run, and neither does Game 5 of the 2026 Finals. A sample size of one is never a sample size.

Third, three-point volume growth is showing signs of plateauing. The NBA tightened its rules on unnatural foul-drawing motions from the 2026-22 season, and switch-everything defensive schemes have made open threes more expensive. If over the next five seasons volume flatlines while accuracy falls, Moreyball will be revisited as an optimisation bubble rather than a revolution.

Fourth, and this is the point that bothers me most: I may be mistaking correlation for causation. Teams that shoot a lot of threes tend to be teams with a lot of good players. Teams with a lot of good players tend to win titles. Three-point shooting may simply be a marker of roster quality rather than its cause.

What to watch next

If there is one conclusion I would bet on, it is that the next competitive edge will not lie in three-point volume but in the ability to generate three different kinds of points in the final two minutes of an elimination game: an open three, a rim attack, and a trip to the free-throw line. The team that can do all three in the same quarter will win the title. The team that can do only one will be eliminated in the second round.

The three-point equation has been solved. What remains unsolved is the variance equation — how to build a system that does not collapse on the exact night when everything goes wrong. If no team wins a title with more than 45 three-point attempts per regular-season game over the next decade, we will have our answer. And if one does, I will have to rewrite this very article.

The standings do not lie. And they have never told the whole story either.