The Air Yards Model countdown finale! Today we wrap up the series with the final list of the top 24 ordered into a subjective ranking with tiers.
Well-validated models are the best and most unbiased forecast of NFL receiver production there is.1 However there are things that they cannot capture that are important to consider.
Changes to team composition, injuries, coaching changes, and more can affect the accuracy of a pre-season model that is based purely on numbers. Finally, many models, including mine, do not give more than an expected value forecast for a player. Point estimates are great, but it’s also crucial to understand the range of outcomes.
In this series, I’m counting down the model’s top-36 wide receivers and publishing forecasts, along with a discussion of the other important factors the model doesn’t know about. Full projections will be released next month.
The Top 24
This series has stretched out for longer than I intended, but now it’s time to unveil the top-24 WRs according to my air yards model. I’ve added some notes on guys the model is higher on than consensus ADP would suggest is reasonable, and I provided a tier list of the top 24 based on the type of analysis I conducted throughout this series. As always you can head on over to the free Air Yards site and conduct your own research on these players.
|Rank||Name||Team||Tier||Air Yards Model PPR per Game||ADP|
Sort the table by air yards model PPR per game to get the model’s original rankings. I created tiers and then took information the model doesn’t know about2 and adjusted the rankings within each. Only Brandin Cooks feels entirely out of tier due to off-season changes.
The top eight of the model looks just about how you would expect any ordering of receivers to look in 2017. This was one of the best signs to me that the model was well calibrated and is measuring what matters. It helps me to put stock into what the model is saying even when ADP is wildly different from what the model recommends.
Jarvis Landry and Demaryius Thomas are well loved by the model and are the top Tier 2 WRs for redraft this year in my estimation. While Landry has seen upheaval at quarterback, there is nothing in Jay Cutler’s resume that points toward Landry not being a large part of the passing offense. Both Cutler and Ryan Tannehill have been well below average on deep passes during their careers, and Miami head coach Adam Gase will be looking to manage Cutler’s gunslinging.
Where Cutler has been steady in his career is in the short to intermediate areas. That’s where Gase will probably want him to focus, and it’s also where Landry shines.
Landry should continue to see a 24-25 percent target share and has as strong a chance as anyone in the top 20 to return value on ADP.
Thomas’ situation hasn’t changed. Trevor Siemian looks to be the starter at QB again this year for Denver, and the Broncos have the smell of an 8-8 team or worse, which means extra passing attempts. The model likes both Thomas and compatriot Emmanuel Sanders, so it clearly isn’t worried about Siemian. Thomas also looks to be a Tier 2 value, and I like him over Michael Thomas if only because his resume is so much longer.
Tyrell Williams and Kenny Britt are both screaming bargains at their ADP. We have spilled countless gallons of air time and digital ink on Britt this offseason, so I’ll just point you to those articles. Williams’ stock has been rising since the injury to rookie Mike Williams solidified his role in the Chargers offense. But what is that role?
Here are two charts that might help us understand.
When the Chargers finally decided to deploy Travis Benjamin as their downfield threat, Williams began to really shine. I think Benjamin will continue in the role of field stretcher this season, and that Williams will continue to see intermediate targets alongside Keenan Allen.
Speaking of Allen, you’ll notice he is not included in this countdown. This is because his projection is too noisy and uncertain for either the model or my brain to calculate. Players like Allen will always be a blind spot for quantitative, evidence-based analysis. If you want a well-reasoned gut-level approach on Allen, Ben Gretch makes the best argument I’ve heard this offseason.
The model’s top WR this year is Odell Beckham Jr. and it’s not particularly close. I agree. Beckham is in his prime and is a steady and explosive performer in fantasy. The one chink in his armor is his falling aDOT throughout his career, driven perhaps by his lack of production at around 25 yards deep. Everywhere else on the field Beckham is a star who will continue to see 28 to 29 percent target share in a revamped and potentially explosive Giants offense.
- The air yards model combined four separate machine learning models into one ensemble model. The models used included a Bayesian neural net, a nearest neighbors algorithm, a random forest algorithm and a tuned generalized linear model. Preprocessing steps included data normalization (centered and scaled), the elimination of near zero variance predictors, and principal component analysis transformation. The model data used were 838 observations of 35 predictors, including age, weight, previous season player volume and efficiency data, 3 year weighted player efficiency data, and previous season team passing volume and efficiency data. The dependent variable was fantasy points per game. The data were split 80/20 into training and test sets. 10 fold cross validation was performed. The measured R^2 on the held out data was 0.66. (back)
- Brandin Cooks to NE, Andrew Luck hurt, etc. (back)