Advice

2017 Air Yards Projection Countdown | Cameron Meredith – WR 32

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.

Cameron Meredith

Air Yards Model Projection

Projected PPR Points per Game: 14.7

Projected WR Rank: 32

Actual ADP2: 43

Model Difference: + 11

Team Personnel Changes

Few teams saw as much upheaval on the offensive side of the ball as Chicago this offseason. Jay Cutler was released and has since retired, wide receiver Alshon Jeffery left for Philadelphia, WR Marquess Wilson is no longer with the team, and WR Eddie Royal was released.

To make up for the personnel departures the Bears signed WR Marcus Wheaton, WR Kendall Wright, and WR Rueben Randle. As a trio they accounted for a meager 52 targets and 549 air yards in 2017, leaving a decent sized chunk of opportunity unaccounted for in the passing game.

At quarterback, the Bears signed Mike Glennon to a three-year 45 million dollar contract with 16 million guaranteed for 2017. In the draft, the Bears traded up to the second overall pick to snag QB Mitchell Trubisky. In the second round they added TE Adam Shaheen and in the fourth, they took RB Tarik Cohen. None of the picks project to have an immediate impact on the offense, though early season struggles could change that, especially at QB.

Here is how the opportunity was distributed throughout the receiving corps last season:

Screen Shot 2017-05-19 at 9.03.47 AM

While he only played in four games, when Kevin White was on the field it was clear he was the focal point of the passing offense. Despite his inefficiency and the general negative buzz surrounding him, a healthy White could negatively impact Meredith’s outlook in 2017. Of the roughly 100 targets and 1200 air yards from 2016 left unaccounted for by the various departures, it’s easy to see White commanding most if not all of it simply by being on the field.

Consider that White and Meredith’s target distributions are similar, indicating that they’ve been used in a similar manner.3

meredith-dot

white-dot

Due to his size and speed, White projects as the guy to take over for Jeffery in the stretch role. If he’s successful on the deeper routes, there is little doubt Chicago wants him to be the primary option in the Bears passing attack.

But even if you think White will be unable to rise to the occasion due to lack of polish or separation skills, I still don’t think it is necessary a good sign for Meredith. A limited White might still steal looks from Meredith in the areas of the field where he normally operates.

It’s slightly morbid, but the best case for Meredith is probably if oft-injured White again misses time to various ailments.

Projected Opportunity

The ARIMA projection4 for the Bears in 2017 sees a fairly large decline in passing. What the model doesn’t know is that there has also been a significant downgrade at the QB position.

Floor: 475 attempts

Ceiling:  548 attempts

Expected: 512 attempts

Bovada Sports Book has the win total for the Bears at 5.5, which is a strong signal of their overall badness. Chicago’s seasonal game script was negative last season, but their pass/run ratio has only just kept pace with the league average. This has been something of a trend for Chicago; being bad does not really cause them to pass that much more.

gamescript

pass ratio

I’m going to accept the low expected pass attempts from the ARIMA model on the assumption that Glennon, Meredith, and White will be less effective than Cutler, Jeffery, and Meredith in the passing game. This will lead to more stalled drives and a reliance on the run even when behind. My overall thesis is basically that the goal for Chicago this season is not really to win, but to lose slowly.

As for Meredith, I’m going to stick with an 18 percent target share. Nothing about how this team has handled him thus far has suggested they see him as their offensive focal point, and I expect the vacated targets to go to White if he’s healthy. That gives us the following range.

Floor: 85 targets

Ceiling: 98 targets

Expected: 92 targets

Career Efficiency

Meredith has been an efficient receiver over the course of his short career. For instance, his catch rate has been very good when you adjust for the depth of his targets.

meredith catch rate

As discussed previously he doesn’t do well deep, but that’s not where he is targeted or deployed. He’s a 20-yard-and-in type of guy.

Where Meredith is well below league average is YAC. He really doesn’t do much after he snags the ball.

meredith yac

When you combine his catch rate with YAC you get a metric I call RACR.  Because he catches most of what comes his way, Meredith is still an above league average receiver on short and intermediate routes. He’s reliable and profiles as a solid WR2.

RACR

Meredith’s career aDOT is 9.9 and his career RACR is 0.89. Let’s calculate his receiving yards.

Floor: 85 targets  * 9.9 aDOT => 841 air yards * 0.89 RACR => 748 receiving yards

Ceiling: 98 targets * 9.9 aDOT => 970 air yards * 0.89 RACR => 863 receiving yards

Expected: 92 targets * 9.9 aDOT => 910 air yards * 0.89 RACR => 810 receiving yards

Catch Rate

Meredith’s career catch rate is a strong 68 percent. Let’s calculate receptions.

Floor: 85 targets  * 0.68 Catch Rate =>  58 receptions

Ceiling: 98 targets * 0.68 Catch Rate =>  67 receptions

Expected: 92 targets * 0.68 Catch Rate =>  63 receptions

TD Rate

Finally, let’s calculate TDs. Meredith’s career TD Rate is 3.5 percent.

Floor: 85 targets  * 0.035 TD Rate =>  3 TDs

Ceiling: 98 targets * 0.035 TD Rate =>  3.4 TDs

Expected: 92 targets * 0.035 TD Rate => 3.2  TDs

Final Projection

Meredith’s final back of the napkin projection is:

Floor: 9.4

Ceiling: 10.8

Expected: 10.2

Let’s compare it to the Sim App:

Screen Shot 2017-05-19 at 11.14.11 AM

The back of the napkin projection is on the very low end of the sim apps range of outcomes, and well below the air yards model. The range of outcomes from the sim is large, but not overly so. Neither it nor the air yards model knows anything about the myriad changes to the Bears personnel this season, however.

Conclusion

Cam Meredith’s ADP is currently 11 spots lower than the air yards model, and I think that best ball drafters probably have it right. Nothing that happened in the offseason has substantially upgraded the probability of Meredith seeing more looks in 2017. Meanwhile the likelihood of a Glennon implosion and the prospect of a rookie QB learning on the job all point to significant downside risk. Kevin White hurts Meredith whenever he is on the field, and the Bears don’t appear to be a team that will sling the ball when they are behind – a position they will likely be in often in 2017.

Meredith is a guy I think the model has wrong, and who may even be a fade at his current ADP. The cheaper White is more appealing to me if you do decide to hold your nose and buy a piece of the Bears passing offense.

Subscribe for a constant stream of league-beating articles available only with a Premium Pass.

  1. 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)
  2. as of mid May 2017  (back)
  3. This far in their careers at least.  (back)
  4. these are posted for every NFL team on airyards.com  (back)
By Josh Hermsmeyer | @friscojosh | Archive

Comments   Add comment