Last week, we debuted an approach to understanding NFL prospects using text analytics on their scouting reports. Using the awesome resource that is Dane Brugler’s draft guides from 2015 to the present, we went through and derived a similarity score between every pair of prospects during that time frame.
Once these similarity scores were computed, we looked at early-career WAR for each comparable player and weighted them by the similarity in their scouting reports. If a player is a close comp to a player who produced a lot during his rookie deal, the score will go higher. Similarly, the same thing would happen if a player was a far away comp to a player who played poorly.
Here, we provide a ranking using this approach for the top cornerbacks in this week’s draft, along with some under-the-radar players the method likes. Enjoy!
A.J. Terrell, Clemson