The Limits of Data in Basketball: Revisiting the Golden State Warriors' 2026-2026 Season

The Limits of Data in Basketball: Revisiting the Golden State Warriors' 2026-2026 Season

Opening

Oracle Arena, the evening of June 19, 2026. Game 7 of the NBA Finals. The clock shows 53 seconds left, the score tied at 89-89. Kyrie Irving faces Stephen Curry on the right wing and rises for the decisive three. The ball drops. On the other side of the floor, LeBron James — who had just produced the block of a lifetime — screams through tears. The Golden State Warriors, the greatest regular-season team in NBA history at 73 wins and 9 losses, fall on their own floor. No stat sheet predicted that ending. The team with the best net rating, the best three-point efficiency, and the strongest case on paper, lost in seven games. Basketball always lives between two worlds — the world of data and the world of what cannot be measured. A dynasty does not fall with thunder; it falls with a missed shot in the final second.

In basketball, the most beautiful moment rarely appears on the scoreboard. It lives in the silence between the whistle and the scream, in the eyes of a player who knows his season has just ended. I began to care about those silences long ago. The more basketball I watch, the more I realize we spend too much time counting and too little time understanding. The Warriors' 2026-2026 season is the most beautiful reminder of that. A great season does not end with applause; it ends with a silence in the stands.

Context: The Data Revolution

The 2026-2026 season was the peak of a long process. From the early 2010s, NBA teams began to trust mathematics more than instinct. Three-point attempts per game soared: from roughly 13 per team per game in 2026 to more than 24 in the 2026-2026 season, then past 35 by the early 2020s. The principle was simple: a three is worth 3 points, a two is worth 2, and if the efficiency gap is not too large, the team that shoots more threes will score more. Daryl Morey, former general manager of the Houston Rockets, turned that principle into a philosophy the media nicknamed "Moreyball."

The Golden State Warriors under head coach Steve Kerr were the most complete expression of that philosophy. They had Stephen Curry — the best three-point shooter in history up to that point — and Klay Thompson, inside a system of constant ball movement. In 2026-2026 they averaged 114.9 points per game, leading the league. Curry averaged 30.1 points, made 402 threes across the season — a record — and became the first player in NBA history to win MVP with all 131 first-place votes. The data said this was a perfect team.

But data does not play the game. People do. To reach that Final, the Warriors had beaten the Houston Rockets and the Portland Trail Blazers, then come back from 1-3 down against the Oklahoma City Thunder in the Western Conference Finals. All season they lost only 9 games. But by June, when everything was staked on seven games, the story was different.

Core: Reading the Game with Both Eyes

Based on my own experience watching these games, there is a common mistake in reading basketball through data: people forget that every metric is an average, and every game is a small sample. Curry's 45% three-point rate is accurate — but accurate across 82 games, not within one possession. In Game 7, Curry shot 6-of-19, including 4-of-14 from three. That does not mean Curry played badly; that is simply basketball.

The tactical lesson here is the interesting part. The Cleveland Cavaliers under head coach Tyronn Lue did something many teams did not dare to do: they accepted a trade-off. Instead of chasing the Warriors' pace, they dragged the game toward their own — slower, more physical, more contact. They had LeBron James and Kyrie Irving alternate attacks into mismatches, forcing the Warriors into fouls and draining their energy. Over the final four games, the Cavaliers won three. That was not luck; it was tactical adjustment built on understanding that data points to an opponent's strengths, but victory lies in attacking the weakness data has not yet named.

Looking at advanced stats, the story becomes clearer. Across the first six games, the Warriors had a dominant plus-minus when Curry was on the floor. But by Game 7, when both teams had drained their physical and mental batteries, the deciding factor shifted from system to individual. Playoff basketball is a different sport, and that is no cliché. In the playoffs, pace slows, space narrows, and every mistake is punished harder. A system built on creating space through speed struggles when the opponent decides that speed will not exist.

This is where I find basketball closer to esports than people think. In League of Legends, a team can win 90% of its group-stage games through a stable meta and map control, then lose the final because the opponent has read its tempo. A stable meta is like a perfect 4-2-3-1 formation: it wins when everything goes to plan, and collapses when someone breaks the plan. The 2026 Warriors were that kind of team — a team of perfect plans, until they met a team playing no plan at all.

The 2026 series also taught a lesson about reading a series. The Warriors led 3-1, and statistically, no team had ever lost the Finals after leading 3-1. That number made many believe the title was settled. But sport does not run on probability; it runs on people. The Cavaliers won Game 5 on the road, Game 6 at home, then Game 7 at Oracle Arena. Three games, three entirely different psychological states. A good data model can compute the probability of each scenario, but no model can prepare a team to win three straight against the best team in the league.

There is a small detail I always remember about that series. After Draymond Green was suspended for Game 5 due to accumulated flagrant foul points, the Warriors lost their most important defensive organizer. At the same time, center Andrew Bogut was injured and could not continue. What stands out is this: before that, no one — not even the most sophisticated data models — treated Green's emotional control as a variable capable of deciding a championship. Data measures steals, passes, points; it does not measure impulsiveness, ego, and the limits of a person's endurance across seven battles. Those invisible variables, added together, made the difference. What is not counted is what decides what is remembered.

Contrarian: The Trap of Absolute Faith in Metrics

But if I said data is useless, I would fall into another trap — the trap of excessive romanticizing. Data made the Warriors great, and data helped the Cavaliers find a way to beat them. The problem is that people turn metrics into religion. When a coach eliminates the mid-range shot entirely — the weapon of Michael Jordan and Kobe Bryant — merely because it is inefficient, that is not analysis, that is dogma. The mid-range jumper still has value at specific moments, when the defense has locked both wings and no better option remains.

The same applies to how we evaluate players. A guard averaging 18 points on modest efficiency can matter more than a player averaging 25 points thanks to a system. The stat sheet sees the product, not the context. And across an 82-game season, context is everything. People talk about metrics that speak, but forget that a metric only speaks when we know how to ask the right question. A metric detached from context is like a quotation cut from its context: literally correct, meaningfully wrong.

I am not against data. I am against the habit of using data to end a conversation instead of starting one. In basketball, every metric is the starting point of a question, not the ending point of a conclusion. In Vietnam, basketball is growing, and with it a data culture. Domestic leagues are beginning to use statistics to evaluate players, select imports, and build tactics. That is a necessary step. But if they learn only the numbers and skip the human part, Vietnamese basketball will repeat the Warriors' 2026 lesson in its own way.

Takeaway

Basketball is a sport of metrics, but it is decided by what metrics leave out: fatigue, fear, a glance, a single distracted second. If you follow a season and trust only the stat sheet, you will miss the most beautiful part of the game. The question worth asking is not which team has better numbers, but which team can endure the moment when every number becomes meaningless. Very likely, that is the question every analyst — whether using Excel or instinct — must answer for themselves. And perhaps that is also the question Vietnamese basketball fans, who are approaching the sport with growing seriousness, will have to answer as domestic leagues begin building a data culture of their own.