Defensive Backs and Linebackers

The latest newsletter from KCSN Sports Data Scientist, Joseph Hefner.

This week we wrap up the series on draft curves and values with defensive backs and linebackers. I’m just going to jump right in and start with a very unfortunate confession regarding player positions in my data.

Regrettably, no dataset that I have access to has reliable information about whether a DB played as a cornerback or a safety. More players are listed as simply DB than are listed as CB or S. That means I am forced to combine them into just defensive backs. That’s going to severely limit the usefulness and accuracy of the graphic for DB’s.

Safeties and linebackers are two of the more difficult positions to assign a specific value to, because they have such varied roles, and they’re often more of the “clean up” players when someone else has made a mistake.

Offenses can and do dictate mismatches against safeties and linebackers, and that can make them look worse than they are. Total tackles is a really weird stat because you want them to be making tackles, but if they’re making too many, that probably means they’ve been out of position, and now have to make a tackle because of that mistake. So you want that sweet spot of not too few, but not too many tackles.

Coverage stats are problematic, because slants and crossers are incredibly difficult to defend, and end up leading to high tackle numbers. Linebackers often wear the green dot as well, which means that a significant portion of their value lies in getting everyone lined up correctly, and that’s not going to show up in any stat sheet.

They’re just weird positions, and hard to grade correctly. Even PFF employees have talked about how their grades for safeties and linebackers have the least predictive power and probably shouldn’t be strongly valued when discussing a players’ contributions.