Rewind, Count, Repeat: The Manual Work Behind Basketball Analysis

Ask a video coordinator what they did last night and the answer is almost always the same: they watched the game again. Not once. Possession by possession, with a laptop open, pausing, rewinding, typing. Who set the screen. Which coverage the defense played. Where the shot came from. Who was on the floor.

That work is invisible in the final product. A coach sees a clean scouting report or a twelve-minute player edit. What they do not see is the four hours of pausing and typing that produced it. This post is about that job: where it came from, why teams came to depend on it, and why it is still done largely by hand.

Before the film: counting in the margins

Basketball has been counted since it was invented. Newspapers printed field goals and free throws for college games in the 1890s, and when the Basketball Association of America, the forerunner of the NBA, played its first season in 1946–47, the official scorer tracked points, field goals, free throws and assists. Rebounds were not recorded until 1950–51. Steals and blocks had to wait until 1973–74, and individual turnovers until 1977–78. For its first three decades, professional basketball's official record was a handful of columns, written by one person at a table.

The first coaches to look past the box score did it with a pencil. Dean Smith at North Carolina charted his team's games by possession as early as the 1960s, tracking points scored per possession rather than per game, because he understood that a slow team and a fast team could not be compared by raw totals. He was, in effect, running an analytics department of one, and its entire output was a sheet of paper.

The film era: when the job became a job

The tool that turned counting into a profession was film. College programs were exchanging 16mm game reels by the 1950s, and a coach who wanted to prepare for an opponent would sit in a dark room with a projector and a notepad. The work was slow enough that only the head coach or a trusted assistant did it.

The VHS tape changed that in the 1980s. Suddenly anyone could rewind, and the volume of film a program could review went up sharply. Teams created a new entry-level role to handle it: the video coordinator. The job description was simple and brutal. Record every game and practice, cut clips for the coaches, log what happened on every possession, and have the opponent's tendencies ready before the next shootaround.

That role became one of the most reliable paths into coaching. Erik Spoelstra joined the Miami Heat as a video coordinator in 1995. Mike Budenholzer started in the San Antonio Spurs video room in 1994. Frank Vogel got his start as a video intern with the Boston Celtics. Many of the sharpest basketball minds of the last two decades learned the game by watching it frame by frame and writing down what they saw.

The tagging era: the possession becomes data

The next shift came in the mid-2000s, and it was less about cameras than about people. In 2004, Synergy Sports Technology began building a database in which every possession of every game was watched by a human logger and tagged with a play type: pick and roll, isolation, post-up, spot-up, transition, and so on. For the first time a coach could ask, "How efficient is this player in isolation against a switch?" and get an answer built from thousands of possessions rather than a hunch.

Two years later, in 2006, Hudl was founded in Lincoln, Nebraska, and brought the same idea down to high school and college programs: upload the film, tag the plays, share the clips. What had been an NBA and major-college privilege became something a small-school assistant could do on a laptop at the kitchen table.

Around the same time, the front office caught up with the film room. Dean Oliver's Basketball on Paper was published in 2004, giving the sport its "four factors" of winning, and the Houston Rockets hired Daryl Morey as general manager in 2007, the same year the MIT Sloan Sports Analytics Conference was first held. Basketball had decided, publicly, that data mattered.

But notice what powered every one of these advances. Synergy's database was built by people watching film and pressing keys. Hudl's clips were tagged by coaches and student managers. The data was only as complete as the hours someone was willing to put in.

The tracking era: cameras arrive, and the tagging stays

In 2010, four NBA teams installed SportVU, a six-camera system originally derived from missile-tracking technology, that recorded the position of every player and the ball 25 times per second. By the 2013–14 season the league had installed it in all 29 arenas. Second Spectrum took over as the official tracking provider in 2017–18, and Hawk-Eye followed for 2023–24.

You might expect that to be the end of manual work. It was not. Tracking data tells you where every player was. It does not tell you what they were trying to do. A camera can see that two players came close together at the top of the key. Only a person, so far, can reliably say whether that was a ball screen, a slip, a dribble hand-off, or two players simply crossing paths, and whether the defense played it in drop, at the level, or switched it.

So even at the top of the sport, staffs of full-time video analysts still tag film by hand, on top of the tracking feed. The volume of data went up. The hours did not go down.

What the manual work actually looks like

Below the NBA, where most basketball is played, there is no tracking feed at all. There is a camera in the corner of a gym, a coach, and a laptop. Here is what the work looks like for an assistant at a college or a club program after a Friday night game:

• Logging the possessions. Play through the game, pausing on each possession to record the outcome, the shooter, the shot location, the assist, the turnover, and who was on the floor. This alone is often three to five hours for a single game.
• Tagging the actions. On a second pass, note the offensive set, the primary action, and the defensive coverage. This is the layer that makes the film searchable later.
• Cutting the edits. Build clip reels: an offense edit and a defense edit for the team meeting, individual edits for players, and a set of the opponent's plays for the scout.
• Writing the scouting report. Personnel tendencies, top sets, out-of-bounds plays, and what the opponent does after a timeout, all pulled from several games of their film, each of which has to be watched.
• Keeping the season totals. Updating the running stats, per player and per game, so that season-to-date numbers exist when a coach asks for them.

Multiply that by a schedule of two or three games a week, and the video assistant is spending more of the week watching basketball than the players spend playing it.

Why it matters enough to do anyway

If the work is this expensive, why has every serious program done it for seventy years? Because film is the sport's shared language, and the counting is what makes the film usable.

• Players believe clips more than words. A coach can say "you are late on the weak-side rotation" ten times. One clip, with the number of times it happened this month, ends the argument.
• Opponents are predictable if you have counted them. A team that runs the same set out of every timeout will keep running it, and the staff that logged it will be ready.
• Development needs a record. A player's shot chart in November and in February tells a story that neither the player nor the coach could tell from memory.
• Accountability needs a number. Minutes, plus-minus, turnovers, shot locations: the numbers a staff trusts are the ones someone actually counted, from the film, every game.

The insight itself is the valuable part. The counting is just the tax you pay to get there.

The part that should not be manual

Look again at the list of tasks above. Some of it is real basketball thinking: choosing which sets to prepare for, deciding what a player needs to see, spotting the coverage that broke down. That is the work a video coordinator signed up for, and it is the work that made Spoelstra and Budenholzer head coaches.

The rest is counting. Who scored, from where, with whom on the floor, for how many minutes. That part does not need a coach's judgment. It needs to be right, and it needs to be back before the next practice.

That is the piece Playtics takes off the analyst's desk. Upload the game film, from any camera, and the next day you get back a player-by-player box score, the opponent's box score, shot locations by zone, minutes and plus-minus for every player, and highlights for every player, every team and every game. Nothing to install in the gym, nothing to tag. If something looks wrong, edit it yourself or tell us and we fix it. The output belongs to the person who used to spend the evening building it.

The film room is not going anywhere. The people in it are the ones who decide what the numbers mean. We just think they should get to start on that part at seven o'clock instead of midnight.

Tell us about your team and how you film. Talk to us and we will get you set up for your next game.