Is the NFL Draft Efficient?

The latest newsletter from KCSN Sports Data Scientist, Joseph Hefner

Every year an incredible amount of paper and ink, both digital and material, is spilled by teams, scouts, analysts, and amateur prognosticators about the NFL draft. People spend incredible amounts of time and energy watching tape, studying plays, looking through stats, building models, creating color-coded spreadsheets, and more.

If you’re reading this, you are probably the exact sort of person for whom these people are doing all this work. Frankly, we can’t stop writing about the draft, and we can’t stop consuming draft content. Is there any time of year that is better than the month leading up to the draft? Mock drafts are flying, pseudo depth charts are being created, and emoji boards are being requested.

In the midst of all this frenetic activity, it can be easy to wonder what actually matters. What draft analysts should be trusted? Who studies film the best? Which data points are relevant? If that’s you, you’re not alone. Even NFL teams can struggle with this. Every year, there are major draft busts, and others who may not quite be busts, but certainly do not live up to the pre-draft hype. People often cite NFL teams and front offices as the best at this (“if you’re so good at this, why aren’t you working for a team?!”), but that just misses the point that this is hard.

I wanted to know how good the NFL is at drafting. This is a difficult question to answer, because it requires that you create a standardized value for every player across all positions. PFF WAR is one such standardized value, which uses their internal grades for players. Approximate Value is another.

I decided to use second contract value for this analysis. Basically, how much is a player paid after his rookie contract is over. That’s a specific value that actual NFL teams put on players. It is possibly the most objective public measure we have. Of course, the cap goes up every year, so just using the raw monetary values will not work. Instead, I used the percentage of the cap at the time of signing. So a $10M contract with a cap of $100M is the same value as a $20M contract with a cap of $200M (10% of cap).