Not so long ago I made some analytics-based guesses on the 2019 NFL Draft. How’d they hold up?
One of the more-celebrated traditions leading up the NFL Draft are predictions from so-called “football experts.”—which are often off the mark. Well I’m no football expert but I made some guesses anyway in true “me” fashion: data! Under the premise that even billion dollar organizations can’t afford to waste all their time leading up to such an important event, I took guesses based on the 2019 Buffalo Bills “time spent” (here) and if the front office tended to draft players they had in for private visits (here). Let’s see how they held up.
Volume of visits by position
One of the first things examined was how much attention was spent based on position groupings. The idea here is that coveted positions would lead to a larger volume of players with whom the team met.
The highest amount of meetings went to wide receivers (14 meetings). The Buffalo Bills selected zero receivers in the draft. So far the analytics model is not paying off. I actually used this as a major component of my specific prediction that the Bills would draft D.K. Metcalf. I wasn’t even close.
Next up was a tie for offensive tackles and defensive ends. The Bills drafted one of each. They did wait awhile for the defensive end (Darryl Johnson Jr. in the seventh round). However, Cody Ford was not only a higher pick, but a much-desired player if the Bills’ “war room” video is any indication.
Next up was defensive tackles and corners with eight meetings each. If you hadn’t heard, Buffalo selected Ed Oliver with their first pick. The Bills didn’t select a corner, but there may have been some real smoke here (see below).
Volume wise, with seven meetings the running back and guard positions followed. The Bills selected Devin Singletary in the third round. The team did not take a guard.
Tight ends (five meetings), outside linebackers (four), and safeties (four) were next by volume. Dawson Knox and Tommy Sweeney were taken for the tight-end group. The Bills also picked up linebacker Vosean Joseph and safety Jaquan Johnson.
Visits by school
Another premise explored was the notion that teams may focus on certain schools to gain insight into a specific player. Rather than focus on the player alone, teammates might give a better perspective into a personality.
In the pre-draft predictions, for the sake of brevity I focused on Ole Miss and Penn State. It was noted that Penn State counts as “local” for the Bills and may have had inflated counts as visits with those players don’t count against the allocation of 30 visits per team. The Bills didn’t grab anyone from Penn State, but did select Dawson Knox from Ole Miss.
Arbitrarily looking at schools that had three or more players meet with the Bills, we add these schools to the list:
- Florida: Buffalo selected Vosean Joseph
- Georgia: No players selected from this school
- Oklahoma: The Bill drafted Cody Ford
For the other draftees, Buffalo only met with one player each from Houston, Florida Atlantic, North Carolina A&T, and Miami. The Bills respectively selected Ed Oliver, Devin Singletary, Darryl Johnson Jr., and Jaquan Johnson from those schools. Tommy Sweeney hails from Boston College. The Bills met with two players from that school.
Did the Bills draft the players with whom they met?
In a separate article we looked at how often the Bills selected players they took the time to meet with. Under Sean McDermott and Brandon Beane, the only time the answer was “yes” was at the quarterback position. Well what about this year?
The Bills met with half of their draft picks prior to the draft. This includes Ed Oliver, Devin Singletary, Dawson Knox and Vosean Joseph. That’s four of their top five selections. The only exception was second-round pick Cody Ford.
Interestingly, Ford did not come up on any pre-draft visit list but indicated on stage at the draft that he had met with the Bills at his pro-day. For analytics purposes I’d never count a meeting that flew under the radar. For practical purposes though, Ford was so impressed with that meeting that he stated he was hoping to be selected by Buffalo.
Summary
The analytics approach wasn’t a good enough model to accurately guess the first player selected. With the number of unknown variables it’s unlikely it ever reaches that level either. As an example, if meetings are thought of like job interviews, a high number of candidates could indicate a high number of players who weren’t “right” for the job. An explanation such as this could explain why no receivers were drafted, but would be completely unknown prior to the draft.
Disclaimers aside, the 2019 NFL Draft results for the Buffalo Bills had enough go right where some trends could help point toward interest in the future. With the exception of wide receiver, the Bills did follow the positional volume trend fairly well. Similarly, for schools where the Bills met with three or more players, they selected a player from that school three out of five times. If we factor out Penn state being artificially inflated due to being “local” it becomes three out of four schools. The Bills also heavily pulled from their pre-draft visits with their top five picks all having confirmed meetings at some point with the team.
If we’re looking to summarize in a convenient analogy, the analytical predictions probably will have a low rate of spotting the fire. However, in the case of the Buffalo Bills’ current front office there’s reason to believe it’s pretty good at finding the smoke.
from Buffalo Rumblings - All Posts http://bit.ly/2JL2GzX
No comments:
Post a Comment