How a Modern Big Board Actually Gets Built
Every spring several million people look at a big board and disagree with it. Far fewer have looked at how one gets assembled, which is a shame, because the process explains most of the disagreements.
A board is not a ranked list of opinions. It is four layers of information with a human argument sitting on top, and almost every fight about a ranking is really a fight about which layer to believe when they contradict each other.
Layer one: production
What the player did on the field. Snap counts, targets, pressures, missed tackles forced, yards after contact, and the situational splits underneath all of it.
The trap here is the raw counting stat. A back with 1,600 yards behind a line that sent three players to the league has told you more about the men in front of him than about himself. Any evaluation worth reading adjusts for context before it compares two players, and the adjustment is where the actual work lives.
Layer two: athletic testing
The combine and the pro days. Height, weight, arm length, hand size, the forty and its ten yard split, vertical, broad, shuttle, three cones.
This is the most consumed and the most misread layer on the board. Testing numbers describe an athlete in shorts. They correlate with football ability at some positions and barely at all at others, and a fast forty has ended more evaluations than it has ever rescued.
The step that separates a serious board from a fan board is converting raw numbers into position-adjusted scores, so that the same time means something different for a 250 pound edge rusher than for a 190 pound slot receiver. Skip it and you end up ranking track athletes.
Layer three: context
Level of competition, scheme, usage and role. A corner who played press man for three years in the SEC and a corner who played off coverage in a smaller conference have not been asked the same questions, and their tape cannot be stacked as though they had.
This layer also holds the dull facts that move players further than anything else: age on draft day, number of career starts, position changes, and whether the production came against starters or against backups in the fourth quarter.
Layer four: medical and background
The layer outsiders never see and teams weight most heavily. A player with a clean board and two shoulder surgeries can slide two rounds on a Wednesday for reasons that will never be published anywhere.
The part nobody sees: assembling it
Here is the practical reality behind every board you have ever read. The four layers do not live in the same place. Conference sites, school athletics departments, combine results, pro day numbers filed by regional outlets, injury notes buried in local coverage, transfer portal records that change weekly.
None of it arrives as a file. A draft site tracking three hundred prospects across five months is pulling from hundreds of sources on a schedule, and the sources that matter most are frequently the smallest. The only person who filed the pro day numbers for a mid-major prospect is often one beat writer in one town.
Two problems come with that. Small regional sites are the least tolerant of automated access, because they run on modest hosting and treat anything unusual as an attack. And a fair amount of college athletics data is served differently depending on where the request comes from. Analysts working across conferences generally pull through a residential ip proxy so requests arrive looking like ordinary readers rather than one machine hammering a small site, which keeps the sources reachable and keeps the operator on decent terms with people whose work the whole exercise depends on.
Unglamorous, and it is the difference between a board built on three hundred prospects and a board built on the eighty everybody already knows.
Where the model stops
The last step cannot be automated. Somebody watches the tape and decides whether the numbers actually describe the player.
Boards that skip that step produce the same recognisable failures every year. They fall in love with the small school athlete with elite testing and no competition. They fade the productive player whose testing is ordinary and whose tape is not. They rank without positional value, so a very good guard finishes above a good tackle, which no team in the league would ever do.
How to read a board properly
Next time a ranking offends you, ask which layer it is leaning on. A board with a player forty spots above consensus is usually not wrong so much as weighted differently, and the honest ones will tell you which layer they trust and why.
That is the argument worth having every April, and it is a better one than saying "he is a first rounder" again with more conviction.