Air Yards have been around for a long time. Elias Sports Bureau has been tracking complete and incomplete Air Yards on passes since 2006. Analysts like Chase Stuart and Brian Burke have both done informative analysis on Air Yards from the quarterback’s perspective. For wide receivers, most fantasy footballers are familiar with Mike Clay’s aDOT (average depth of target), and it was introduced over five years ago. Is there really anything else to learn? The answer may surprise you.
First, it’s probably best to define our terms. Receiver Air Yards are the total number of yards thrown toward a receiver on a play in which he is targeted, both complete and incomplete. If you add them up over a game or a season, you get a receiver’s total Air Yards.
Air Yards are charted orthogonally – or at right angles – from the line of scrimmage. So Air Yards aren’t a measure of how much distance the ball actually travelled. Instead, they measure potential yards a receiver would have produced if he caught the ball and then was immediately tackled. In other words, they are a measure of intent.
Volume is Important, and Air Yards Help Us Measure It
When it comes to forecasting football players, the thing we are best at projecting is volume -- carries, pass attempts, targets. Fortunately, volume correlates very well with fantasy points.
For wide receivers, we track volume either through snaps, routes run, or targets. Of these three stats, targets are the most common measure since they are most widely available. What Air Yards data does is allow us to see just how many yards are behind each of those targets. They are another way for us to quantify volume. Since volume is highly predictive of fantasy points, this is pretty exciting.
This graph shows how PPR fantasy points rise as target depth goes up, even when you adjust for catch rate.
The deeper the target, the more expected PPR fantasy points the target is worth. For instance, throws that are 40 yards deep are caught only about 30 percent of the time, but are still worth more expected fantasy points than a 10-yard throw. The picture is largely the same with non-PPR scoring. Points still go up and to the right in a linear way with the depth of the throw.
Using Air Yards to Predict Fantasy Points
Last season over at RotoViz, I ran Air Yards data through a complicated machine learning model to spot potential buy-low candidates at wide receiver and tight end. Each week the model looked at targets and Air Yards and spit out a list of guys that were under-performing their volume.
The thinking behind the model is simple: If a player is getting a lot of targets and there are lot of Air Yards behind those targets, but he has put up terrible games in terms of fantasy points, we should expect a rebound in the near future.
How can we say that with any confidence? It turns out that it takes just three games for a player’s target share and share of team Air Yards to stabilize to the point where we can make good predictions about them the rest of season.
Volume is sticky. In plain English, that means we can be pretty confident that if we see a receiver getting heavy targets and Air Yards early in the season, the volume will continue as the season progresses. This means if we can identify a guy getting good volume but not putting up points, we’ve found a solid buy-low opportunity.
Give Me Some Examples, Please
Last season the model identified Terrelle Pryor and Tyrell Williams as “Buy Lows” in Week 1. Pryor reappeared on the buy low list each week early on, despite almost constant change at the QB position in Cleveland. Owners that acted early to scoop Pryor off waivers got a great deal on a player who was a safe bet to return value.
Similarly, Tyrell Williams eventually ascended after Keenan Allen exited in Week 1 with an ACL injury. And he did so despite Travis Benjamin putting up crooked fantasy numbers the first couple weeks.
The model worked because volume is king in fantasy. Air Yards in particular help us predict fantasy points because they are an almost pure measure of QB and coaching intent. Each week a team has a game plan that includes key playmakers who they want to get the ball to. Air Yards help us better read the quarterback and coach’s minds to predict who those players will be.
Opportunity Rating and RACR
Creating and testing machine learning models is a lot of work. In the end, all the models are really doing is figuring out the proper weight for each piece of data and then making a prediction based on those weights. We can get most of the way there with simpler methods that use back-of-the napkin metrics. So here are a couple I created to get 90 percent of the results with 10 percent of the effort.
WOPR
WOPR stands for Weighted Opportunity Rating. It takes a player’s target share and share of team Air Yards and combines them in a way that best predicts both PPR and standard fantasy points. The formula for WOPR is:
1.5 * Target Share + 0.7 * Share of Team Air Yards
This opportunity rating combined with a player’s PPR points can help us spot players who under-produced their volume. In other words, WOPR helps us spot buy-lows both in-season and between seasons.
RACR
RACR stands for Receiver Air Conversion Ratio. RACR is an efficiency stat that answers the question: “How well does a player convert a yard thrown at him into receiving yards?” The formula for RACR is:
(Completed Air Yards + Yards After the Catch) / Total Air Yards
or, simplified as:
Receiving Yards / Total Air Yards
Essentially RACR combines catch percentage and YAC into one metric. I prefer RACR over Yards Per Catch because it is over twice as predictive, while giving you similar information. Like all efficiency metrics though, it regresses heavily to the mean each year.
2017 Buy Low Targets
That was a lot of talking about metrics. How can you actually use them to find buy-low candidates?
Below is a list of the top-36 receivers in the NFL last season ranked by WOPR, or Opportunity Rating. Most of the names are familiar. This makes sense since volume correlates very well with fantasy points.
RACR is included to get a feel for the player’s efficiency along with Air Yards Per Target, which is conceptually identical to Mike Clay’s aDOT. Since catch rate and YAC are dependent on the depth of the targets a receiver gets, you will often see players with low aYPT have a high RACR.
One shining example is Jarvis Landry. Landry had an aYPT of just 6.5 yards (the lowest on the list) and a RACR of 1.33 (the highest on the list).
~
Buy Low Targets
Players to buy low at this early stage of the offseason might include names like DeAndre Hopkins and Allen Robinson. They both saw a ton of volume and underperformed it. Chances are good they will see similar volume next season, and simply through the randomness inherent in efficiency they could end up posting better seasons. Similarly, Marvin Jones and Davante Adams are good bets to see enough opportunity to provide value at what may end up being WR3 ADPs.
Even the top name on the list might be undervalued, as crazy as it sounds. Mike Evans saw a tremendous amount of volume in 2016 and he produced, but he still underperformed his volume. You can see from his RACR of 0.53 that his catch rate and yards after catch were far lower than his peers that saw similar volume. How do we know?
Yards after catch and catch rate regress strongly to the mean year-over-year, and the mean RACR for someone getting Evans’ volume is around 0.69, almost nine percent over what Evans put up in 2016. Many people will mistakenly judge this as a negative for Evans. Don’t fall into that trap! While Evans is probably not as efficient as most of the other high-volume receivers on this list, it’s likely he’s more efficient than he showed in 2016. With that volume and some positive variance in 2017, Evans is a strong buy even at a first-round price.
Volume is king in fantasy football. It probably always will be. Thanks to Air Yards we have a new way to quantify it.
