Over the past three off-seasons, Adam Peters has used a combination of draft picks, free agency and trades to rebuild a roster that was left depleted by his predecessor.
Throughout his time in Washington, Peters has consistently preached building through the draft. But he has deviated from that narrative on a few occasions, most notably when he traded draft picks for veteran starters Laremy Tunsil and Marshon Lattimore.
Peters’ spending in free agency has been moderate, for the most part, with his
biggest splurges being Terry McLaurin and Tunsil’s extensions and this off-season’s acquisition of edge rusher Odafe Oweh. Aside from those deals, he has mainly signed players on mid-level to bargain short-term contracts.
Peters’ approach to roster renewal has been similar to that of the four successful rebuilds I profiled near the start of his tenure. That might not be a big surprise, since he was the EVP of personnel for one of the teams. Success in the draft played a big part in the revival of the Buffalo Bills, San Francisco 49ers, Detroit Lions, and Houston Texans under new GMs, following periods of mismanagement.
I have always taken it as gospel that building through the draft is the key to successful roster management. It is the mantra repeated by GMs of the most successful franchises. It makes sense from the standpoint of roster management in a salary-capped league. And the few attempts I’ve made to look into what makes successful teams tick seemed to back that up.
Others in the fanbase have expressed frustration with the slow pace of the rebuild and wished that Peters would lean more heavily on free agency to acquire marquee talents.
This article was inspired by an exchange in the Hogs Haven comments section during the lead-up to the draft. A regular commenter suggested that the draft was over-rated, citing the Patriots’ success in adding free agent talent ahead of the 2025 season. I was about to fire back with all the reasons building through the draft is essential to long-term success. But then I asked myself, “do I really know that’s true?” So I wrote down a note, “Look into building through the draft vs. free agency”. And here it is.
Best vs Worst Team Roster Construction – 2020 through 2025
To get a handle on how championship teams’ rosters are built, I examined how each of the Super Bowl teams from 2020 through 2025 acquired their players. To have a basis for comparison, I also examined the rosters of teams which earned the first and second positions in the draft in the same seasons. The resulting sample of 12 championship teams and 12 worst-placed teams was sufficient for robust statistical analysis.
There are many ways to approach roster construction. Most attempts I have seen to address this question have focussed on starting players. I opted for a different approach for two reasons. First, the emphasis on starters overlooks the significant contributions of rotational players, backups and special teams player to getting a team through a season and into the big game. Second, the “starter/backup” terminology is an obsolete vestige from the era before unlimited substitutions. Many positions on modern NFL rosters, such as defensive line, linebacker, running back and receivers are rotational. It is not uncommon for players listed as starters to get less playing time in a game than other players at their positions.
Snapshot Sampling: To get a more comprehensive view of how rosters were constructed, I used a snapshot approach. For the championship teams, I analyzed the game day rosters which played in the Super Bowl. This approach takes in the slated starters, as well as any backups who filled in for injured players to get their team to a Super Bowl.
In the case of the worst teams, I analyzed the game day rosters in the last win of the season, provided it was not in the last few weeks of regulation. This was intended to capture the teams’ best rosters, prior to any late-season substitutions for player evaluation or tanking. Not to imply that NFL teams ever tank on purpose.
Major Contributors: For the purposes of this analysis, I counted “major contributors”, arbitrarily defined as players who took more than 20% of snaps on offense or defense, specialists including punt and kick returners, and the two non-specialists on special teams. The latter category resulted in a lot of ties, which were broken by snap counts on offense and defense. When that didn’t work, I used the random number generator in Excel for a digital coin flip. The aim of this approach was to capture all the players who made significant contributions, and avoid biasing the results with players who only contribute on special teams.
Method of Acquisition: The method of player acquisition was classified using the following categories: Drafted, Undrafted Free Agent (UDFA), veteran Free Agent (FA), Trade, Practice Squad/Waiver Claim (PS/W Claim). Importantly, players were classified based on how they made their way onto the team’s current roster, rather than how they made it into the league. The UDFA category was restricted to players signed by the team in their draft years, without having played in a game for another team.
Draft Percentage: To answer the question posed in the title, I calculated the percentage of players on each roster who were acquired through the draft or as UDFAs, using the strict definition above. The reasoning for including UDFAs in their draft year is that these signings are an extension of the draft process. Teams generally sign the top players from their draft boards who are still available immediately after the draft and continue to add players they had rated highly throughout camp and preseason, and sometimes into the regular season.
Results
The breakdown of players by method of acquisition from the past 12 Super Bowl rosters is shown in the first table. Teams were listed in descending order of draft percentage.
Over the past six seasons, approximately 60% of players on the average Super Bowl roster were acquired through the draft or as UDFAs in their draft year. The draft was the most common method of player acquisition on all but one of the 12 championship teams, with veteran free agency coming in second 11 times. The lone exception was the 2025 Patriots, who acquired one more player through veteran free agency than the draft. In all 12 teams, players acquired through the draft or signed as UDFAs in their draft year accounted for 53.8% or more of the game day roster.
There was a 13.8% spread in draft percentage between the 2025 Seahawks, who were most draft-intensive, and the 2025 Patriots, who relied the most on veteran free agency.
How does that compare to the worst teams in the same seasons? The next table shows the breakdown of player acquisition method among teams that earned the first or second pick in the draft following the 2020 through 2025 seasons:
The average draft percentage on bottom-two teams was 49.5%, a full 10.8% lower than the Super Bowl teams. But the averages don’t even tell the whole story.
The bottom-two teams have much greater spread of draft percentage than the Super Bowl teams (33.5% vs 13.8%). The average draft percentage among the worst teams is pulled up by two draft-intensive rosters (2020 Jaguars, 2023 Commanders). But only four of the 12 worst teams had higher draft percentages than the least draft-intensive Super Bowl team.
More than half of the bottom-two teams had rosters with less than 50% drafted and UDFA-acquired players. That is not seen on any of the Super Bowl teams.
Statistical Significance
The difference in draft percentage between the best and worst teams is so large and clear cut that you really shouldn’t need a statistical test to be convinced that it’s real. But for the benefit of the hardcore stat-heads amongst us, a one-way, unpaired t-test returned a p value of 0.0018 for the difference between groups. For the non-statheads, that means that the probability of a difference this large occurring purely by chance is around two times in 1,000. That is a highly conclusive result to even the most skeptical analyst.
What Does It Mean?
The results lead me to conclude that there are many ways to build a bottom-of-the-league roster; but there appears to be only one way to build a Super Bowl team. That involves a heavy reliance on building through the draft.
The 2025 Patriots, who inspired this analysis, were the least draft-reliant Super Bowl team in the last six seasons. Even so, they had a higher percentage of players acquired through the draft process than seven of the 12 bottom-two teams over the same period.
The results do not mean that leaning on the draft will guarantee success. The most draft-intensive team in this sample was one of the worst. Ron Rivera’s 2023 Commanders were also a heavily-drafted team, and I don’t need to remind readers how bad they were. Obviously, drafting well is important too, as are coaching, scheme, cap management, team culture and success in acquiring players through other means.
The higher percentages of drafted players on Super Bowl rosters could be seen to raise a bit of a conundrum. Are these teams so much better than the basement dwellers because they emphasize building through the draft over free agency? Or are they just better at drafting players, resulting in higher retention of draftees and less need to seek players through other means? Ultimately, it doesn’t matter. Whichever answer is closest to the truth reinforces the importance of teams building through the draft.
Lastly, some readers may have noticed that two franchises account for half of the Super Bowl teams used in this analysis. It is possible the results were biased by the particular team building habits of the Kansas City Chiefs and the Philadelphia Eagles. I considered that a design feature, rather than a bug. If the aim of this analysis is to look for tips on best practice in team building, then it makes sense to over-represent the teams which have made repeat Super Bowl appearances.
