I was sitting in a dimly lit student research space in the top floor of EBE at the University of Missouri. It was almost one in the morning during the start of the fall semester in 2013. Looking over my shoulder I heard some quiet rustling ... and the only place it seemed to be coming from was the community fridge located on the side wall of the research space.
Curiously I walked over to the fridge and opened it. Immediately I was confronted with a horrendous smell. The inside of the fridge was crawling with insects who were enjoying the contents of a fridge which had been shut off for far too long. I closed the fridge and reported it to EBE’s facilitator.
It’s strange the things that stick with us. Oddly enough this is one of my first memories when I think about the time of my life I spent working on my masters thesis. A thesis which had the goal of finding and selecting players in the NFL draft with greater success than current NFL front offices could consistently achieve.
The thesis used machine learning to achieve it’s goal. I don’t want to get into too much of the nuts and bolts information, but in simple terms machine learning is essentially statistics on crack while using a computer.
So where does Albert fit in all of this?
The study primarily used college statistics and player combine data. I was looking for two types of players when I was running the algorithm over the data:
- Players who would be elite
- Players who would play in 75% or more of their team’s games after coming into the NFL
I used a lot of player statistics to try and quantify who was elite or not. The rationale for searching for these types of players is pretty obvious. Drafting elite players is pretty helpful to a football team’s success isn’t it?
The rationale of searching for players who appeared in 75 percent of their games while in the NFL is not as obvious. Essentially I was searching for a player who could contribute and be healthy. I also assumed that if a team kept playing a certain player, it meant they were successfully keeping their job and playing well enough to garner time on the field. Essentially I was looking for under the radar contributors who might typically get missed on draft day.
I ran the algorithms countless times and one player kept showing up over and over again as being a successful ‘75 percenter’. That player was Albert Wilson. At the time I knew nothing about him, I just knew he had decent combine stats and a successful career at Georgia State; a university which obtained FBS eligibility during Albert Wilson’s senior year.
Albert Wilson stats at Georgia State
Wilson also had a decent combine with an official 40 time of 4.43 (the unofficial time was 4.35).
After learning that Wilson had a good college career, solid combine numbers, and was a projected as an undrafted free agent I began to hope the Chiefs would pick him up after the draft.
Much to my happiness, the Chiefs signed Albert Wilson as an undrafted free agent and he made the team. I felt legitimized in my work and Albert Wilson instantly became my favorite Chiefs player.
In a way, I would link my work’s success to Albert Wilson. Albert Wilson could be living proof that my efforts and research were working.
There Were Other Players
The algorithm selected other players too. I haven’t gone back and revisited my work so it’s interesting to see who was selected. Below is a list of the selections who were projected as starting over 75 percent of their games during their respective careers:
(Keep in mind these are players from the 2014 draft, sorted by overall draft pick)
Sammy Watkins, Mike Evans, Brandin Cooks, Kelvin Benjamin, Marqise Lee, Jordan Matthews, Paul Richardson, Davante Adams, Cody Latimer, Allen Robinson, Jarvis Landry, Josh Huff, Donte Moncrief, Jalen Saunders, Bruce Ellington, Devin Street, Jared Abbrederis, Robert Herron, TJ Jones, Michael Campanaro, Tevin Reese, Jeremy Gallon, Brandon Coleman, Austin Franklin, Marcus Lucas, Albert Wilson, Willie Snead, Josh Stewart, Isaiah Burse.
There are a lot of good players in this list, and some players who have fallen by the way side. The goal was never about perfection, it was about drafting better than an average NFL team could. It was about finding guys like Albert Wilson and Michael Campanaro; undrafted players no one saw coming who help their teams win on special teams and/or the regular offensive unit.
Albert Wilson is a Winning Player
I won’t be able to remember every time Albert Wilson helped the Chiefs win a game, but here are a few.
This is a basic pitch and catch, but ball placement is the difference between 1st down and potential touchdown. pic.twitter.com/6Z5oSNdHxI— Seth Keysor (@RealMNchiefsfan) December 18, 2015
This is a game in 2015 where Wilson caught a touchdown pass on a quick slant. The Chiefs beat the Chargers in this game by a score of 10-3.
There was the fake punt touchdown in 2016 where the Chiefs beat the Falcons by a score of 28-29.
Also to note is a touchdown Wilson scored where the Chiefs beat the Jaguars by a score of 19-14.
When will defenses learn you shouldn't cover Albert Wilson with a LB? pic.twitter.com/sHzmz0VFUe— Gary McKenzie (@Super_G_Chiefs) November 8, 2016
There are two times this year where Albert Wilson pretty much saved the day and made winning plays which directly contributed to the Chiefs winning the game. Both of them occurred during the Redskins game. Wilson stopped an interception from happening and made a brilliant fourth quarter catch with seconds remaining to put the Chiefs in field goal position to win the game.
Of course you can’t forget this block either...
Wilson trashing Talib, LOVE IT! pic.twitter.com/mp1dMLXwFm— Gary McKenzie (@Super_G_Chiefs) November 28, 2016
I Get It, I’m Biased
Yes, I understand I’m incredibly biased when it comes to Albert Wilson. Like I mentioned above, I feel that my success as a sports data scientist has been hitched to a wagon named Albert Wilson. That being said I do believe Wilson is a winning player and has helped the Chiefs win football games. What more could you ask for?
Lastly, I’m just going to throw this out there.... If any NFL team is reading this and is looking for help in their front office analytics department feel free to contact me. I’ve contacted a number of teams but it’s very difficult to get your foot through the door.