How important are efficiency stats for fantasy?
Over the past three seasons, Martavis Bryant ranks behind only Julio Jones in PPR fantasy points per snap. On a fantasy points per target basis, only DeSean Jackson, Doug Baldwin, and Jordy Nelson rank higher than Bryant. The implication here is that Bryant is good, or at least efficient, from a fantasy standpoint. I agree with this implication and I think the metrics supporting it are sound. Still, it seems, every time I list some sort of efficiency metric, I’m met with a chorus of boos.
The twitter machine replies, “Don’t you know? Efficiency metrics hardly correlate as well as volume-related metrics year-over-year. Efficiency stats are really only valuable as a means of predicting future regression.”
I posted a statistic several weeks ago, looking at wide receivers in terms of depth-adjusted yards per target. Adam Thielen posted one of the best marks of the past decade. Allen Robinson posted one of the worst. The immediate response was that Thielen was due to regress negatively, and Robinson was due for a major positive regression and possible return to WR1/WR2 status for fantasy. My response was, “Well, yes. Maybe. I’m not sure.”
Is Thielen likely to again make his way onto the top-10 list of the past decade in terms of depth-adjusted yards per target? No, almost definitely not. Is Thielen likely to see an increase in targets on the back of an uber-efficient season? Yes, I believe so. What about Robinson, is he likely to improve by depth-adjusted yards per target in 2017? Yes, I believe so. Is he likely to see an improvement in targets? I don’t think so, no. Is he not risky heading into 2017? No, I believe his potential for another disappointing season is now much higher than most wide receivers posting only middling efficiency stats in that target range.
Looking at fantasy points, yards, or really anything on a per-target basis is inherently flawed. A target, in and of itself, is a positive indicator of talent. The quarterback throwing him the ball — due to trust, due to the fact that he’s open, or for whatever reason — is likely a sign that he is good. So dividing one “good” thing (targets) by another (fantasy points, yards, etc.), usually obscures the results. (Shout out to Adam Harstad for continually beating this drum.) One could compare it to something like dividing a kicker’s total field goals made by extra points made. Does that really tell us anything about how good a kicker is? No. While per-target efficiency stats are far more useful than this made up kicker statistic, it’s important to understand the context in which this is true.
In general, efficiency stats (per target, per carry, etc.) aren’t as important as volume statistics (targets, carries, air yards, etc.) when building out fantasy projections, but they are very important at the extremes of each spectrum. It’s why when looking at a player like Rob Gronkowski, we need to adjust for efficiency on top of volume in our projections. It’s why no one is predicting an increase in targets for Tavon Austin, in spite of minimal target competition, on top of another grossly inefficient season. As ESPN’s Mike Clay put it when I asked for his opinion on this topic, “[I found the] polar ends are predictive, but the middle bulk is not.”
So, let’s just dig into the year-over-year correlations for some different efficiency statistics for wide receivers. (Notes: For this exercise we are using a sample size of all wide receivers with at least 50 targets from 2007 to 2016. This provides us with a sample of 862 player-seasons. Also, when referencing fantasy points, we’re only concerned with receiving fantasy points — e.g. I’m not counting Tyreek Hill’s rushing or special teams fantasy production.)