Basketball is dying from data addiction: Lessons from the analytics revolution sinking deep

I have been sitting in the broadcast booth of an NBA game in Chicago for the past 44 years, and here is something I have never said publicly: the analysts sitting in their computer rooms are killing basketball in a way no one realizes.

Don't get me wrong. I'm not against data. I'm against the way data is being used to replace eyes instead of supplementing them. And believe me, that difference is more important than all the numbers they throw out every day.

Last week, a young colleague at the Chicago Bulls sent me a 47-page analysis report. He excitedly wrote: "Check this out, everything has data." I opened it, scrolled through cascading graphs, probability matrices, and PER or EPM indices calculated to the third decimal place. And I asked only one question: "Do you know what that kid is thinking — the one sitting at the corner of the court, watching the opponent but not looking at the ball?"

He couldn't answer.

There, ladies and gentlemen, is why basketball is losing itself.

Forty-four years of watching games from an insider's perspective gives me a privilege few have: I can see the silences between numbers. I saw the moment an 18-year-old named Jayson Tatum walked onto the Boston Garden court in 2026 with an expression of fearlessness — something no analysis could quantify. I saw how a shy kid at the academy, whom every mock draft had going in the third round, was the only one who understood how to read the spaces between two defensive lines.

And I also saw how an entire industry is losing that ability.

Context: When charts replace eyes

Let's talk about how sports analytics infiltrated professional basketball. Everything started in the 2000s when teams began realizing there were things the naked eye missed but data could catch. A triple-double could be a trap. A 40% three-point shot could be a phantom number. And how a team moves when the ball isn't nearby — what insiders call "ghost football" — is what truly decides games.

But consensus became fanaticism. Teams started hiring MIT Sloan Sports Analytics Conference graduates instead of experienced scouts. They built data centers right next to locker rooms, placed massive screens displaying real-time stats next to the coach's bench. And they began making decisions based on algorithms rather than intuition.

Daryl Morey's Houston Rockets was the perfect symbol of this trend. They abandoned mid-range shots — what the Moreyball model called "inefficient behavior" — to cram three-pointers into every possession. The result? They won more regular-season games than ever, but collapsed at the exact most important moments — every single time.

Why?

Because when playoffs arrive, when opponents can prepare specific strategies for a series, when pressure turns athletes into real people with real fears — algorithms are no longer enough. That's when things that can't be quantified become the only difference between winning and losing.

Core: Three cracks in the analytics empire

Based on my game-watching experience, I've identified three serious flaws in how the basketball industry is using data.

First, the disconnect from real-world rhythm. Data analysts work with aggregate numbers — season averages, career trends, matchup histories — but they aren't on the court to feel the opponent center's breathing getting heavier with each passing minute. They don't see the moment a defender starts leaning his shoulder an inch to the left before the offensive player fakes. They only see the final result, not the process.

Take 2026, for example: when Kawhi Leonard hit the buzzer-beater to beat the Philadelphia 76ers in the playoffs, all post-game analyses focused on how many shots he made, his success rate, and how much Toronto's win probability increased. But nobody — not a single person — mentioned what I saw from the stands: how Leonard stepped onto the court before that shot. His eyes never wavered. He didn't look at teammates, didn't look at the coach, just stared straight at the basket. That wasn't confidence. That was the stillness of someone who had accepted the outcome beforehand. And that, ladies and gentlemen, doesn't appear in any statistics table.

Second, the dependence on large sample sizes. Analytics requires sample size — the more data, the more reliable. But basketball, especially at the highest level, is a sport of small moments. One pick-and-roll in minute 88 of Game 7 can decide an entire season, but it's just one game out of thousands. The data volume isn't sufficient for statistical analysis, but its significance far exceeds any number.

I still remember the 2026 game, Game 6 of the Western Conference Finals between the Golden State Warriors and Oklahoma City Thunder. My colleagues in the upper deck were calculating the Warriors' win probability after falling behind 3-2. The odds were so low that nobody believed in a comeback. But I didn't need the numbers. I just needed to see how Stephen Curry left the court after the first quarter, how he threw his towel on the ground harder than usual. That was the signal of someone getting angry — not angry from frustration, but angry from being insulted. And when a star gets insulted to that degree, the algorithm might as well shut down.

Third, and most importantly: the loss of connection between humans and systems. Analysts create models based on the past, but basketball is a sport of adaptation. Lineups change, tactics change, people change. A model trained on three years of data can become useless when a new coach changes how the team moves without the ball. And while algorithms are trying to catch up, on the court, humans are adapting in completely unpredictable ways.

Take Denver Nuggets and Nikola Jokic as an example. In the 2026 season, when Jokic won his third consecutive MVP, analysts tried explaining his success with numbers. How many triple-doubles he had, his true shooting percentage, where his EPM ranked in the league. But nobody — I emphasize, nobody — could explain with data how Jokic reads the game like a chess player. He doesn't just see his teammates' current positions; he sees where they'll be next. He doesn't just see the gap ahead; he sees the gap that will appear three seconds later. And that ability, ladies and gentlemen, cannot be learned from any dataset.

Contrarian angle: Why comebacks matter more than ever

Now, this is the part I know will cause controversy. And yes, I like that.

People call me old-fashioned because I still believe in things like "game feel" and "star instinct." They call me outdated because I still emphasize the importance of plays that don't show up in statistics. And I accept those accusations, because I know something data addicts never admit: sometimes, the crowd is wrong. And sometimes, correctness lies in moments nobody recognizes.

In 2026, when Mohamed Salah had just moved to Liverpool for £36 million and was being mocked across forums for inconsistent form, I was one of very few people who declared he would break the Premier League scoring record. I didn't have an algorithm. I didn't have a machine learning model. I only had my eyes and 44 years of experience telling me: when a player moves with that speed, when he can change direction without losing momentum, when he looks at the goal as if it's the only prey in existence — then no defender can keep up, no matter how sophisticated the defensive system.

The result? Salah scored 32 goals in his first season, breaking the 38-game record. And analysts immediately wrote articles explaining why this had been predicted — based on data. They took credit for a prediction they never made.

That's the nature of the problem. Data analysts are often very good at explaining the past. But the past doesn't win championships. Only the future can do that.

And to predict the future, you need to understand people. You need to understand how they react under pressure, how they adapt when everything falls apart, how they find themselves again after a devastating loss. Not their win probability. Not their PER. But the real person, with all their flaws and strengths.

Every giant that falls is a slap in the face for those who collect names instead of collecting people. And that's, ladies and gentlemen, why people like me are still necessary — no matter how much algorithms want to replace us.

Takeaway: Basketball needs both, not one

So what's the solution? I'm not demanding the abandonment of analytics. I'm just demanding balance.

Modern basketball needs both: eyes and data, intuition and algorithm, emotion and reason. The problem isn't bad data. The problem is that data is being prioritized while other things are forgotten.

Teams need analysts who can read the spaces between numbers. They need people who can ask "why" instead of just "how many." They need people who understand that a $60 million player won't necessarily make more of a difference than a shy kid from the academy who knows how to observe.

And most importantly, they need people who understand that basketball, ultimately, is a sport of human beings. Not computers. Not algorithms. But real people, with real fears, real dreams, real ability to do extraordinary things in moments nobody can predict.

I've seen that too many times in the past 44 years. I've seen teams built on data get beaten by teams built on instinct. I've seen players "not good enough by the numbers" become stars because they had something no number can measure. And I've seen perfect systems collapse before moments that nobody — not a single person — could explain.

That's, ladies and gentlemen, why I love this sport. And that's also why I'm worried about its future.

Don't let data addicts kill what makes basketball beautiful: the unpredictability of human beings. That's the most important lesson 44 years of watching has taught me. And that's what I will continue to say, no matter how many people want to silence me.