Kevin Love, 38.5%, and the Gap the Box Score Forgot

Kevin Love finished Game 5 of the 2026 NBA Finals with an eFG% of just 38.5%. On the box score, it was a forgettable night for a star. I read it that way too, until I spent 72 hours taking apart the final 14 offensive possessions and saw an entirely different story. On six of them, Love stretched the Warriors' defense and opened space for LeBron James to score 10 points directly. Not a single column recorded those six plays. For the first time in my life, I understood that data is never neutral: it records only what someone chose to record, and forgets the rest.

To understand why the number 38.5% misled viewers, you have to place it in the right moment and the right system. The 2026 NBA Finals pitted the Cleveland Cavaliers — the reigning champions — against the Golden State Warriors, a team that had just added Kevin Durant to a roster that had won 73 regular-season games. The Warriors led 3-1 going into Game 5 at Oracle Arena on June 12, 2026, and they closed the series with a 129-120 win.

Tactically, the Warriors defended by switching almost everything. Draymond Green and Durant could trade assignments from guard to forward without collapsing the structure. That forced Cleveland to attack through two primary weapons: LeBron James and Kyrie Irving. As the defense collapsed toward those two, the role of the third man, Kevin Love, became strange. He was no longer a scorer, but a maker of the space that let others score.

This is where the traditional box score fails. It measures the shot, but not the space. It credits the finisher, but not the man who dragged a defender out of position so the shot could exist.

I started by redrawing the floor for the final 14 possessions. On each one, I marked Love's position as the ball left James' or Irving's hands. The result made me redo the work three times, because I could not believe my eyes.

On six possessions, Love stood in the corner or on the wing, pulling a Warriors defender out of the paint. On five of those, the man guarding Love was one of the opponent's best defenders. The space he created showed up not in the points column but in the fact that the defense could not send another body into the middle. LeBron James scored 10 points directly from the possessions where Love was the gravitational pull.

This concept is now called gravity — a player's pull on the defense, measured by the distance at which opponents are forced to guard him. But in 2026, player-tracking data was still young. SportVU had only been installed in every NBA arena a few years earlier, and teams were just beginning to learn how to use it. The general audience had almost no access. They had only the box score.

And the box score, as a system, has no cell for space. It was designed to count events, not to measure influence. That is a structural limit, not a personal failure. Every statistic you read is the product of a design choice: someone decided what was worth counting and what was not.

What stands out is that the Warriors understood that value better than anyone. They never left Love alone in the corner without a defender. But precisely because they had to guard him, they stripped themselves of the ability to send another man into the middle. Love did not score, but he forced the best defense in the league to choose between two bad options. That is a form of influence the box score has no name for.

I remember building my own spreadsheet from four NBA stat sites, comparing Love's eFG% against his number of spacing actions. The two figures moved in opposite directions. One said he was useless; the other said he was the link that kept Cleveland's offensive machine running. Same player, same game, two contradictory truths.

That was when I understood the thing that later became my working principle: every result is a deliberate lie, because it is created by someone to serve some story. The number 38.5% is not mathematically wrong. It simply does not answer the question fans actually wanted to ask: did Love help the team win?

I started a blog. The first post ran two thousand words and got exactly 47 reads. But it laid the foundation for how I have looked at basketball ever since: never narrate the game, but always start from a shocking number, then use tactical context to explain why that number is lying.

This is also why I am always skeptical of the idea of empty stats. When someone says a player only scores on a bad team, the next question has to be: in what context did he score, against what defense, and what did his teammates gain from his presence? The box score cannot answer that, but it does not follow that the answer does not exist.

Years later, when the pandemic shut every league down, I went back to old data and spent nine weeks measuring the average distance between two defenders in Olympiacos' pick-and-roll coverages in the EuroLeague. Out of that, the podcast Vùng phủ sóng was born. A podcast is not born in a studio; it is born in the silence of the world. And that silence, in the end, is also a kind of unread data.

But if I stopped at "don't trust the box score," I would have fooled myself once again. The deeper problem lies elsewhere: when data is missing, the analyst's reflex is to invent meaning to fill the gap.

I have seen this repeat hundreds of times over ten years of watching the industry. A player posts a bad stat line, and the media immediately builds a story about attitude. A team loses three straight, and pundits assign it a locker-room crisis. No one verifies, because inventing a story is always easier than admitting we do not have enough information.

In data analysis, there is a state called null input. When it occurs, the only honest answer is: not enough information to conclude. But the sports-content industry does not run on that principle. It needs articles, headlines, conclusions. And so people invent. They fill a null input with plausible-sounding speculation, then present it as fact.

The only way to resist that temptation is to build data discipline. Before I write a single conclusion, I force myself to answer: how many information points do I have to back this sentence up? If the answer is none, I stop. Not because I have no opinion, but because an opinion without data is just a dressed-up guess.

The biggest risk in this profession is not analyzing wrongly, but analyzing when there is nothing to analyze. That is when data turns into literature, and the reader loses any way to tell the two apart.

So what is the variable for the next game? It is probably not a player. It is us — the readers and the writers. Basketball never ends with a buzzer; it ends with a question. And the right question is sometimes not who played well, but what data am I missing to answer that.