Trang chủBasketballFour Rewinds of the Tape: When the Official Box Score Gets It Wrong
Basketball

Four Rewinds of the Tape: When the Official Box Score Gets It Wrong

**Câu trả lời lõi**: Dữ liệu bóng rổ chính thức có thể sai vì người ghi nhận phải quyết định trong chưa đầy một giây và không có thời gian tua lại. Cách xử lý: đối chiếu tối thiểu hai nguồn độc lập, mở băng khi hai nguồn lệch nhau, và loại bỏ mọi con số không xác minh được. **Dữ kiện chính**: - Tháng 2/2019, bảng thống kê Duke gặp Virginia Tech ghi Zion Williamson 10 rebound; băng hình cho thấy pha thứ tám thuộc cầu thủ khác. - Ivan Perišić chạy 12,3 km mỗi trận ở World Cup 2018, nhưng chỉ 31% quãng đường hướng về khung thành đối phương. - 612 trận NBA từ tháng 3 đến tháng 10/2020: ném phạt của cầu thủ dưới 25 tuổi giảm 2,8% khi sân vắng; EuroLeague không thay đổi đáng kể. - Tháng 2/2023, Han Xu bị khai thác trung bình 14 lần mỗi trận ở tình huống pick-and-roll; đối phương ghi 1,17 điểm mỗi lần. **Nguồn**: Ghi chép cá nhân và bản ghi podcast của Matthew Chen; bảng thống kê nhà tổ chức NCAA tháng 2/2019; dữ liệu Second Spectrum tháng 2/2023. **Hỏi đáp liên quan**: - Vì sao bảng thống kê chính thức vẫn sai? Vì người ghi nhận quyết định trong thời gian thực, không có cơ hội tua lại, và hai người có thể ghi hai kết quả khác nhau cho cùng một pha bóng. - Chỉ số nào phát hiện lỗi hỗ trợ phòng ngự tốt nhất? Số lần bị khai thác mỗi trận trong pick-and-roll và điểm trung bình đối phương ghi được mỗi lần khai thác. - Sân vắng ảnh hưởng đến ném phạt như thế nào? Nhóm cầu thủ dưới 25 tuổi giảm trung bình 2,8% ở NBA, trong khi EuroLeague gần như không đổi.

In February 2026, at Cassell Coliseum, I sat in the freelance section with a phone in my left hand and a notebook in my right, logging every possession. Duke played Virginia Tech, and the crowd remembered the game for Zion Williamson. I remembered it for a much smaller number. The official box score credited Zion with ten rebounds. When I watched the tape back, the eighth one did not belong to him.

That night I rewound the clip four times. Half speed, then frame by frame. The ball came off the rim, Zion reached up, but a teammate touched it first. Under standard recording rules, that rebound was not his. I counted the tape four times over, and the error belonged to the source, not to me. A rebound the organizer logs incorrectly still counts — if you are willing to rewind.

I published a correction on my personal blog. Two hundred and forty reads. An editor at The Ringer shared it, and the following season I was hired as a statistics research assistant. My career began with a wrong number nobody had bothered to check.

The real story is not that data is wrong. The real story is that data fails through a predictable mechanism, and that mechanism lives in people, not in machines.

Every NBA game puts one or two people courtside, eyes locked on a monitor, fingers on a tracking application. They must decide in under a second whether a play is an assist, a block or an altered shot, a steal or a lost ball. There is no time to rewind. There is no timeout. There is crowd noise, a referee crossing the sightline, a coach standing up. And there is a detail rarely mentioned: the same possession, logged by two different people, can produce two different results, and both enter the official database.

Advanced tracking systems like Second Spectrum solve part of that problem with cameras that follow the ball and the players' joints. But they only answer who touched the ball, not who was responsible. A pick-and-roll that gets beaten appears in no box score at all. It only appears when someone sits down, rewinds, and counts.

Since the Zion piece, I have kept one rule: every number I use must come from two independent sources. If the sources disagree, I open the tape. If the tape is unclear, the number is cut from the article, however flattering it looks.

In 2026, while interning at a local radio station in New York, I was handed an assignment outside my field: analyze the defensive system of the Croatia national team at the World Cup in Russia. I rewatched all seven of their matches and charted every run by Ivan Perišić. The winger covered an average of 12.3 kilometers per game. But when I sorted those kilometers by direction, only 31 percent of them pointed toward the opponent's goal. The rest were lateral, backward, positional.

I wrote a nineteen-page internal memo. My editor passed on it, calling it too dry. Three weeks later Croatia reached the final, and he called me back. I wrote nineteen pages to extract one sentence worth saying. Croatia were not the team that ran the most — they were the team that ran in the right direction. That 31 percent figure is the number I wanted to talk about. In a match everyone measures by total distance, sorting direction is what creates the difference.

That lesson followed me back to basketball. The team that runs the most is not the best defensive team. The player with the highest distance covered is not the smartest mover. Counting is not enough. You have to classify.

In 2026, when leagues shut down, I defended a master's thesis on how empty arenas affect free-throw performance. I collected data from 612 NBA games between March and October and cross-checked it against previous seasons. Free-throw accuracy among players under 25 dropped by an average of 2.8 percent with no crowd present. In the EuroLeague, where crowds play a smaller role in free-throw routine, there was no significant change.

The review committee argued the sample was too small and the pandemic distorted the data. They were partly right. But they could not argue one thing away: the youngest players lost the most, and that drop vanished when they played in an environment already accustomed to silence. When the crowd disappears, young free-throw shooting disappears with it — unless you play in the EuroLeague. A thesis can be challenged; numbers do not argue back.

I used that thesis as my first solo podcast episode. And I left the limitations section in, read aloud, uncut.

In February 2026, the New York Liberty of the WNBA lost nine straight games. I built an investigative podcast series on the breakdowns in their switching defense. Data from Second Spectrum showed rookie center Han Xu being targeted an average of fourteen times per game in pick-and-roll situations, with opponents scoring an average of 1.17 points per successful attack. That number appears in no box score a fan can look up.

I reconstructed each possession, not to identify who was weak, but to find the mechanism: Han Xu was pulled beyond the three-point line, and once she was pulled out, nobody remained in the paint. Head coach Sandy Brondello declined an interview request. Three weeks later, the team changed its scheme: Han Xu was kept closer to the rim, and the guards took on more responsibility up top. That series drew eighty thousand listens, five times a normal episode.

I still credit the analytics assistants, because they supplied the underlying data. My sourcing network grew not because I was harsh, but because I never claimed credit that was not mine.

The counterintuitive part: most basketball arguments are not fought between right and wrong, but between two different ways of counting. When a reporter cites an official data feed and a fan pushes back with what he saw, both are usually correct within their own frame — and neither has opened the tape. Which means most social-media debate about basketball is an argument about method, not about the game.

I do not oppose the eye test. I oppose using it as a conclusion in place of data. And I oppose the reverse: treating a single number as final evidence. A player who scores zero points in a game can be the best player on the floor, if you count the possessions opponents avoided him. No box score records the times an opponent chose a different direction.

People see a mistake and laugh; I see a mistake and look for the source. That is the job. Unglamorous, but repeatable, and only what is repeatable can be verified.

What I keep after six years in this work is a habit: before publishing any number, I ask where it came from, who produced it, under what conditions, and how much my conclusion would change if it were wrong. Most of the numbers I have thrown away were the ones I liked best.

Next season, when someone tells me the data shows something, my first question will not be whether it is true. It will be whether they have watched the tape back. If they have not, we have nothing to discuss.

Cầu thủ liên quan