D.J. Moore leads the way as Anthony Amico presents his early wide receiver projections for the 2018 draft.
Building the Model
I took 247 FBS WRs drafted since 2002, looking to test relevant variables and measure success, using roughly half the sample for training, and the other half for testing. I built the model to predict a WR’s best season within his first three years in the league. The resulting score can be thought of as the quality of the season on a scale from 0 to 100. After running multiple variables through regression testing, here were the three that tested as most significant and were used to build the model (listed in order of statistical significance).
The model tested with an out of sample r-squared value of 0.3203.
Here are the results of the model when interacting with the 2018 WR class. Included is every WR inside of NFL Draft Scout’s top 30,3 along with some other players I examined in my NFL Draft Preview series. Since I did not use any non-FBS WRs to build the model, you will not see any in the results.
The output is listed in two different ways. “Predict” is the raw total for fantasy points in the player’s best season (first three years). “Prediction Score” is the result of scaling that number from 0 to 100.
|Player||School||Age||Breakout||Avg Age||NFL DS Rank||Dominator||Predict||Prediction Score|
|Christian Kirk||Texas A&M||21.1||20.1||20.6||2||0.368||155.4||72.9|
|James Washington||Oklahoma State||21.8||20.8||21.3||4||0.329||131.7||64.2|
|Michael Gallup||Colorado State||21.8||20.8||21.3||12||0.307||112.6||57.2|
|Allen Lazard||Iowa State||22.1||20.1||21.1||17||0.318||109.8||56.1|
|Equanimeous St. Brown||Notre Dame||21.3||20.3||20.8||11||0.230||103.7||53.9|
|Cedrick Wilson||Boise State||22.1||21.1||21.6||20||0.340||102.0||53.3|
|Keke Coutee||Texas Tech||21.0||21.0||21.0||23||0.308||98.6||52.0|
|Korey Robertson||Southern Mississippi||22.5||22.5||22.5||30||0.427||91.5||49.4|
|Ricky Jeune||Georgia Tech||24.1||22.1||23.1||56||0.594||77.2||44.2|
|DaeSean Hamilton||Penn State||22.8||-||22.8||13||0.254||74.7||43.2|
|Jaleel Scott||New Mexico State||22.9||22.9||22.9||22||0.299||68.4||40.9|
|Byron Pringle||Kansas State||24.1||23.1||23.6||33||0.363||53.8||35.6|
|Cam Philips||Virginia Tech||22.0||22.0||22.0||47||0.345||50.4||34.3|
|Marcell Ateman||Oklahoma State||23.3||-||23.3||19||0.223||49.1||33.8|
And here are all of the prediction scores presented visually.
D.J. Moore is the clear top WR in this class. Even though he is only the seventh-rated prospect by NFL Draft Scout, Moore has the lowest average age and second-highest dominator rating. After very quietly being one of the most productive WRs in college football, Moore is sure to be a RotoViz favorite. His prediction score would rank eighth in the overall sample.
The second tier of WRs appears to be a one-man show as well. It belongs to Christian Kirk. Kirk was another young breakout, and he ranks highly in the NFL Draft Scout ratings. In addition to being a great WR, he was a special teams dynamo. Kirk had almost 2,000 return yards and seven return TDs to his name at Texas A&M. He may be even better than his score suggests.
If you’re looking for Calvin Ridley, he’s all the way down at 12 – this despite being the top-rated WR on NFL Draft Scout. Working against Ridley are his average age of 22 and final year dominator of just 0.268. With a profile that is lacking much to love,4 Ridley is sure to be one of the more heavily-debated prospects this draft season.
Deontay Burnett and Auden Tate both appear in the top eight of the model, despite being ranked outside of the top 15 by NFL Draft Scout.5 They are interesting names to watch going forward due to their youth and final season dominance.
Something to keep in mind while observing the results is that they will likely change leading up to the draft. In particular, I would expect a lot of the rankings to change across the industry once the NFL Combine and various pro-days take place. My plan is to release another iteration of these scores sometime after the combine and before the NFL draft occurs, then follow up the draft with a post-draft model utilizing draft position.
- This is essentially a proxy for draft position. (back)
- Breakout age is defined as the first season a player posts a dominator rating of at least 0.30. If a player did not break out or broke out in his final season, the input was just his final age. (back)
- Except those for whom I had no age. (back)
- Ridley has strong career market share numbers but was an older college entrant. (back)
- Tate is actually outside of the top 20. (back)