2017 Air Yards Projection Countdown | Kelvin Benjamin – WR 34
Now that the draft is over I’m excited to report that I’ve finished my air yards model for NFL receivers.
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.
Air Yards Model Projection
Projected PPR Points per Game: 14.7
Projected WR Rank: 34
Actual ADP2: 35
Model Difference: + 1
Team Personnel Changes
With the eighth pick of the 2017 draft the Panthers chose running back Christian McCaffrey. Carolina then went on to select receiver Curtis Samuel with the 40th pick. Both appear poised to compete for immediate opportunity in the Panthers offense.
Ted Ginn accounted for 94 targets and 1170 air yards in 2016 but signed with the Saints in the offseason, freeing up roughly 20 percent of Carolina’s passing volume. The rest of the receiving corps remained intact, and the Panthers picked up Benjamin’s 2018 option in early May. Here is how passing opportunity was split in 2016.
After missing all of 2015 to injury, Benjamin’s air yards fell precipitously from 2023 in 2014 down to 1380 in 2016. His aDOT fell from 14 to 11.8 indicating that it wasn’t simply a matter of less targets, but where Benjamin was being targeted on the field. A plot of his 2016 targets by depth shows that Benjamin was targeted much less often on passes deeper than 20 yards compared to 2014.
Much of the decline in downfield opportunity was explained away as recovery from the season-ending left ACL tear he suffered in 2015, as well as a lack of conditioning late in the season. Due to this, it’s perhaps not all that encouraging that beat reporters have quoted coach Ron Rivera as saying that Benjamin is “a little heavy” and report that he’s been getting extra conditioning work in the offseason.
Pre-season narratives aside, Kelvin’s top competition for targets is an aging 32 year old Greg Olsen. Benjamin isn’t exactly a spring chicken himself at 26, but it’s certainly reasonable to project a slight increase in his target share in 2017. I’ll bump him to 23 percent of team targets on the assumption that he splits the lost Ginn looks with Olsen (who will get the lion’s share) and that new additions McCaffrey and Samuel will largely see targets at expense of Corey Brown and Devin Funchess.
Carolina passing attempts per game have been on an upward trajectory since 2011, the year Cam Newton was drafted.
The ARIMA projection3 for Carolina passing attempts sees this leveling off slightly in 2017. Last year the team dropped back to pass 559 times, but the model thinks a repeat of that type of volume is unlikely.
Floor: 488 attempts
Ceiling: 577 attempts
Expected: 532 attempts
With the addition of two intriguing new weapons in the draft, I think there is a good chance that the Panthers drop back to pass more than the model projects, so I’m taking the over. Bovada Sports Book puts the Panther’s win total at just 8.5, indicating that bookmakers have no specific read on their 2017 fortunes. Likewise their seasonal game script was close to neutral, indicating that a drastic move from their 55/45 pass/run ratio isn’t incredibly likely.
Instead, what will make or break the 2017 Panthers is their ability to extend drives and just be generally more efficient with their offensive play calling. The Panthers dropped from 78 to just 63 plays per game from 2015 to 2016.
Because plays per game correlates pretty well with winning, and running the football is something you do with a positive game script, it’s not surprising that in 2015 the Panthers’ pass/run ratio was 47/53 — almost the opposite of 2016. The bottom line of all this for Benjamin is that while his owners will definitely be cheering for a bigger pie of plays to slice up, he’s likely not to be the biggest beneficiary.
We’ll settle on 550 Carolina pass attempts, which gives Benjamin an expected target total of 127, a floor of 112 and a ceiling of 132. This is a reasonable range that lands squarely between his rookie year target total of 145 and his rather anemic 117 last season.
Quick note on methodology: In the past two articles I’ve used three year weighted efficiency to do these back of the envelope projections. I’ve since done some further testing and have found that for RACR and aDOT especially, career averages are the most predictive. So that’s what I’ll be using from here on out.
Benjamin is not what anyone would describe as incredibly efficient. When we roll up his catch rate and YAC by depth of target, we get a metric I call RACR. Here is what his performance compared to league average looks like for his two full seasons in the NFL.
The picture isn’t too terrible to be honest. Benjamin is roughly league average at most depths, and while you might want more than average from your WR1, he isn’t actively harming the offense. Again though, Benjamin’s problem area appears to be the deep ball. If he’s going to return excess value this season, he’ll probably need to make more than his share of catches stretching the field.
Let’s calculate receiving yards.
Floor: 112 targets * 13.0 aDOT => 1456 air yards * 0.58 RACR => 844 receiving yards
Ceiling: 132 targets * 13.0 aDOT => 1716 air yards * 0.58 RACR => 995 receiving yards
Expected: 127 targets * 13.0 aDOT => 1651 air yards * 0.58 RACR => 957 receiving yards
Benjamin’s career catch rate isn’t great at 52%. Let’s calculate receptions.
Floor: 112 targets * 0.52 Catch Rate => 58 receptions
Ceiling: 132 targets * 0.52 Catch Rate => 68 receptions
Expected: 127 targets * 0.52 Catch Rate => 66 receptions
Since Benjamin isn’t the type of guy to break off long runs consisting mainly of YAC, his TD value will come on end zone targets. Sample sizes for this type of target are small, but Benjamin has shown he is capable of using his tall frame and ample girth to grab his fair share of these high value looks over his career.
He’s been especially effective in the red area 10 yards and in. Let’s calculate his expected TDs.
Floor: 112 targets * 0.061 TD Rate => 6.8 TDs
Ceiling: 132 targets * 0.061 TD Rate => 8 TDs
Expected: 127 targets * 0.061 TD Rate => 7.7 TDs
If we combine all the above we get the following range of outcomes on a PPR points per game basis.
Let’s compare those with the WR Sim app:
As usual, the sim score app delivers a wider range of outcomes than our back of the napkin approach. This is incredibly useful, as it reminds us that we’re holding a lot of variables constant here that are in fact highly variable. Repeating another pattern, the ML model is more optimistic than the back of the napkin approach overall, suggesting that the things I’m not considering in this write up explicitly (like QB efficiency) are important and helping Benjamin’s overall outlook.
The model and ADP are in alignment on Benjamin’s expected value this season. While he isn’t a sexy pick and sports some early red flags this offseason in terms of conditioning, he is the clear WR1 on the Panthers heading into the year. Benjamin’s performance on end zone targets and deep passes – both of which are extremely difficult to predict – will likely determine if he delivers on his ADP this season for fantasy owners.
Subscribe for a constant stream of league-beating articles available only with a Premium Pass.
- 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)
- as of early May 2017 (back)
- these are posted for every NFL team on airyards.com (back)