The question nobody had answered
Every year around the NFL draft, the same argument plays out across sports radio, Reddit threads, and family group chats. Team A just used a first-round pick on a wide receiver. Team B traded up for a defensive tackle. And somewhere in the middle of all the takes, someone always asks: but how many of these guys actually make the team?
Not just make the opening-day roster. Stay. Contribute. Justify the capital spent on them.
We went looking for a real answer and couldn’t find one. Published analysis on the subject either covered a handful of seasons, focused only on first-round picks, or used a binary in/out roster flag with no sense of whether a player actually played. That felt like the wrong question answered with the wrong data.
So we built it ourselves.
The data
The foundation is a database we assembled from public NFL records. Every draft pick going back to 1980. Weekly roster data going back to 2002. Per-game snap counts going back to 2012. About 1.3 million rows total, sitting in a SQLite database we can query any way we want.
The key methodological decision was the hit metric. A player counts as a hit if they produce 500 or more snaps in any single regular season of their career. That is roughly one season of meaningful rotational playing time, the threshold where a player has spent at least one year as a real contributor. Not just making the 53-man roster on opening day, and not just suiting up for a few games as a backup. Earning a real role.
We started with a more forgiving threshold (100 snaps in a season, which roughly translates to “appeared in two or three games”). The numbers at that line were noisy. Almost every drafter looked above average and the league hit rate ran around 85 percent. Tighten the line to 500 snaps and the league baseline drops to 57 percent, and the picture sharpens. The teams that actually produce contributors separate from the teams that cycle bodies through the active roster. We rewrote the published articles against the new threshold in early May and have committed to publishing against 500 snaps going forward.
What the data actually shows
Two things became clear once the threshold was tightened.
First: drafting well and keeping the players you draft are barely correlated. Of the 32 NFL franchises, only two teams (Kansas City and Philadelphia) rank top-five in both 2012-2022 draft hit rate and 2017-2021 cohort retention on the 2025 roster. Every other contender excels at one or struggles at both.
Second: the league’s reputations don’t always line up with the data. Philadelphia leads the league in drafting at 67.5 percent over an eleven-year window and ranks third in retention. Buffalo ranks just thirteenth in drafting at 59.8 percent but leads the league in retention at 54.2 percent, nine points clear of second place. Cleveland, which looks elite at the 100-snap threshold, drops to sixteenth at 500 snaps. New England finishes thirty-second on the eleven-year hit rate. The Rams, who traded away the cohort for Stafford, Ramsey, and Miller, retain one of their 2017-2021 cohort hits.
Those findings are what pushed us toward building a publication around the data, not just an open-source repository.
The 53 Report
The 53 Report launched at the end of April 2026. The name comes from the 53-man active roster limit, which is the number every NFL team has to work with and the number that every front office decision ultimately comes back to.
The product is GM Performance Grading: a structured, data-backed evaluation of how well each team’s general manager drafts, retains talent, allocates capital by position, and builds around a franchise quarterback. We are going through all 32 teams, one article at a time, with three article shapes depending on what the data supports.
A scorecard piece covers a tenured GM with three or more completed draft classes. It grades four columns: hit rate against the league baseline, retention of drafted players, position construction across the roster, and the franchise pick. It closes with a letter grade. Brandon Beane got the first one and earned an A-, the deduction reflecting his middle-of-pack draft volume against league-best retention.
A narrative piece covers a paradox or anomaly in the data that doesn’t fit a clean scorecard. The Eagles piece is this shape. The original framing was a paradox: Roseman drafts elite, keeps fewer than expected. The threshold update flipped the second half of that. At 500 snaps, the Eagles rank first in drafting and third in retention. The piece reads now as the league’s clearest example of doing both well, with a meta paragraph on what changed and why.
A methodology piece covers league-wide findings that don’t belong to any single team. The opening article in the series established the core finding: the top five drafters and the top five retainers are almost entirely different organizations, and that split matters more than either metric alone.
Three pieces published. Twenty-nine teams queued.
Why publish, not just open-source
The database and the MCP server sitting on top of it are already open-sourced on GitHub. Anyone can query the data. But raw query results are not analysis. The work of figuring out which findings are interesting, which are misleading, and which actually say something about how NFL organizations make decisions is a separate task.
That is what The 53 Report is for. Not just the numbers, but the argument behind them. The counterpoint to every finding. The prior published work that this data extends or challenges. That work lives at the53report.com, published on a weekly cadence through the end of the 2026 season.
The long-term target is a paper for the Sloan Sports Analytics Conference in 2027. The abstract window opens around October 2026. We think the dataset is defensible against the published prior art, and we want to make that case in front of the people who will push back hardest on it.
This is what we do
Brooks New Media is primarily a web services and hosting company. We build sites, manage servers, run audits, and keep small business clients online. That is the core of what we do and we are proud of it.
But the honest version of the story is that we build things because we find problems worth solving. We built an automated SEO audit pipeline because we wanted a faster way to tell clients what was wrong with their sites. We built a personal finance tracker because spreadsheets weren’t cutting it. We built a self-hosted infrastructure that handles all of it because depending on someone else’s platform for things we can control ourselves didn’t make sense.
The 53 Report is the same instinct applied to a question we couldn’t stop thinking about. Fourteen years of NFL draft picks, one question, and a database that finally answers it.