Ask a simple question. How many people do American police kill in a year? For a long time, nobody in government could tell you.
In August 2014, a police officer shot and killed Michael Brown in Ferguson, Missouri. In the months of protest and reporting that followed, journalists went looking for a basic number: how many people do police in this country shoot and kill each year? Nobody had a trustworthy answer. There was an official FBI figure, but when the Washington Post checked it against the real cases, the FBI was missing more than half of them.[src]
So the Post did the counting itself, one shooting at a time, from local news and public records. It called the project Fatal Force. The Guardian ran its own count, called The Counted. A group of organizers built Mapping Police Violence. And one man, D. Brian Burghart, had already started a fourth, Fatal Encounters, going all the way back to the year 2000. Four serious efforts, and not one of them run by the government.
That is the number of people police shot and killed from 2015 through 2024, counted case by case in the Post's records. It works out to more than a thousand a year, and 2024 was the worst year they ever logged. On the last day of 2024, the Post stopped counting.[src]
Sit with that for a second. The best record of how often American police kill people was built by a newspaper, a nonprofit, and a guy with a spreadsheet. Years after Ferguson, the FBI still could not put a solid national number on the table. By the Post's own account, the FBI in 2021 caught only about a third of the fatal shootings the Post found.[src]
Here is why that matters, and it is simpler than it sounds. A government cannot manage what it cannot see, and it cannot see what it never counts. No count, no budget line, no policy, no accountability. The counting comes first, before any of that. On this one, the government left the counting to strangers.
This is a series about a broken measurement, told in plain terms. It never singles out an officer, a department, or a case, and it takes no political side. Here is the honest surprise the rest of the series builds to: these databases actually agree with each other about who died. They just disagree about the totals, because each one decided to count a slightly different thing. Follow that thread and you learn something about how the whole country keeps its records.