When Does Average Depth of Target Stabilize for Wide Receivers?

So far I’ve found that average depth of target (aDOT) stabilizes in 10 games (or 286 aimed throws) for QBs and in 14 games (or 30 targets) for RBs. Today, I’ll present and discuss the aDOT stabilization point for WRs.

Methods

It’s been a while since my last post, so below is the standard procedure I use to determine a stat’s stabilization point; this time applied to WR aDOT:

  1. I collected aDOT data for all WRs that had at least 8 games played for the same team from 2006 to 2016.
  2. Starting with WRs that played 8+ games for the same team, I randomly selected two sets of 4 games for each WR and calculated their aDOT in both sets.
  3. I calculated the split-half correlation (r) between the two randomly-selected sets of games.
  4. I performed 25 iterations of Steps 2 and 3 so that r converged.
  5. I repeated Steps 2-4 in 8-game intervals, from 16+ games all the way to 72+ games.
  6. For each “games played” interval, I calculated
    1. the number of games at which the aDOT variance explained, R2, would mathematically equal 0.5.1
    2. the “true” aDOT for a hypothetical WR that’s had an observed performance of 12.00 aDOT through that number of games.2
  7. I calculated a weighted average of the results from Step 6.3

Results and Discussion

Below is the familiar stability table, this time for WR aDOT:

GamesnrR2 = 0.50Avg aDOTObs 12.50 aDOT
Wtd Average412.0012.25
46250.47412.0412.27
84290.61512.0212.26
123080.71512.0112.25
162200.79412.0012.23
201640.85411.9412.22
241140.86411.8712.18
28790.87412.0112.26
32570.91311.9012.20
36490.91312.0212.26

As always, the bottom row labeled “Wtd Average” reveals my main finding: WR aDOT takes only 4 games to stabilize. Given an average of 5.3 targets and 26.1 routes run per game in my sample, this translates to only about 24 targets and only about 117 routes. Notably, this is faster than any WR stat I’ve analyzed thus far:

That first bullet point isn’t a typo: aDOT does, in fact, stabilize in nearly 1/10th the number of games/targets/routes it takes YPT. What’s more, this is now the second time I’ve found YPT to be far less reliable than a mathematically related WR stat (i.e., YPRR previously). Given the overwhelming popularity of YPT in both fantasy and non-fantasy football circles, these results are troubling, to say the least. Using it for describing or explaining the past is fine. For predicting WR production in the future, however, it’s becoming increasingly clear that relying on YPT is ill-advised. And I’m not the only one who’s come to this conclusion.

To drive this point home, it’s worth showing the reliability timelines of aDOT, YPRR, and YPT in graphical form:

Along the x-axis is number of games, while the y-axis represents how much each stat reflects “true” skill as opposed to random luck. The dots represent each stat’s stabilization point, defined as the point at which it represents 50 percent skill and 50 percent luck. There are two major revelations when displaying my results this way. First, although I used “1/10th” earlier to describe the ratio of YPT stabilization to aDOT stabilization, reliability trajectories are non-linear. It may take 1/10th as long, but at no time is YPT’s skill percentage 10 times (or more) lower than aDOT’s. That said, the differences are still considerable:

  • After 4 games, a WR’s actual aDOT represents 50.0% skill, but his actual YPT represents only 9.3%.
  • After 14 games, a WR’s actual aDOT represents 77.8% skill, but his actual YPT represents only 26.4%.
  • After 39 games, a WR’s actual aDOT represents 90.7% skill, but his actual YPT represents only 50.0%.

Now, one might argue, “OK. I get that YPT’s bad. But aDOT isn’t an analogous alternative because it’s more indicative of a WR’s offensive role than of his individual ability to gain yardage.” While I’m sympathetic to this logic, I do have two counterarguments. First, there’s the simple arithmetic that YPT = aDOT + YAC + a bunch of zeroes for yardage on incomplete targets, as well as the empirical truth that aDOT constitutes about 80 percent of YPT for the average WR. Indeed, among 130 WRs that played at least 25 percent of offensive snaps in 2016, only Cordarelle Patterson had a higher YAC per target (5.1) than aDOT (4.8). In short, YPT and aDOT are inextricably linked due to math.

Second, a WR’s offensive role is inextricably linked to his individual skill set in a chicken-and-egg sort of way. For instance, was DeSean Jackson’s consistent Top-20 aDOT the past three seasons the result of his role as Washington’s deep threat or was his role as Washington’s deep threat the result of an individual skill set best suited for high air-yard targets? What about Torrey Smith in San Francisco or Ted Ginn in Carolina? On the other end of the spectrum, what about the Cowboys’ Cole Beasley or the Rams’ Tavon Austin? What I’m getting at with these rhetorical questions is that aDOT is an indicator of individual yardage-gaining skill because WRs fill certain roles based on how their skill sets help maximize the team’s goal of gaining yardage (and ultimately scoring points). This is an unavoidable aspect of how the game of football works.

Finally, below is a table showing actual and “true” aDOT for every WR in 2016 that played at least 25 percent of snaps per PFF:

PlayerTmGTargaDOTRkTrue aDOTRk
Sammie CoatesPIT114921.6118.51
J.J. NelsonARI146718.1316.52
Robbie AndersonNYJ147517.5416.23
DeSean JacksonWAS158916.9515.94
Mike EvansTB1616816.0915.55
Marquise GoodwinBUF146316.6715.36
Roger LewisNYG91919.3215.27
Josh BellamyCHI123816.9515.08
Will FullerHOU148715.81115.09
Terrelle PryorCLE1613215.51615.010
Corey ColemanCLE106216.0914.911
Dez BryantDAL128915.61314.812
Michael FloydARI126715.81114.813
Kenny StillsMIA167515.61314.714
Alshon JefferyCHI129215.31714.615
Julio JonesATL1412514.52414.116
John BrownARI146614.81814.117
Brice ButlerDAL143015.61314.018
Breshad PerrimanBAL156414.72114.019
Rishard MatthewsTEN1610414.32713.920
Marvin JonesDET1510114.32713.921
Sammy WatkinsBUF85014.62313.822
Russell ShepardTB123914.81813.723
Allen RobinsonJAX1614413.93313.624
Torrey SmithSF124514.42513.625
Chris HoganNE155414.22913.526
Cody LatimerDEN71416.1813.527
Phillip DorsettIND155614.13113.528
Devin FunchessCAR145414.13113.529
Deonte ThompsonCHI93614.42513.430
Ted GinnCAR168913.83413.431
Jaron BrownARI62414.81813.432
Andre RobertsDET132514.72113.433
Brandon MarshallNYJ1512113.63613.334
T.Y. HiltonIND1614313.53713.335
Cobi HamiltonPIT112814.22913.236
Tajae SharpeTEN167713.53713.137
Jermaine KearseSEA167613.53713.138
Brandin CooksNO1511313.34113.139
Travis BenjaminSD147113.34113.040
Vincent JacksonTB53013.73512.941
Emmanuel SandersDEN1513113.14412.942
DeVante ParkerMIA158213.24312.943
A.J. GreenCIN109513.14412.944
Paul RichardsonSEA113213.44012.845
Jordy NelsonGB1614612.84712.746
Charles D. JohnsonMIN133713.14412.747
Tyrell WilliamsSD1611112.84712.748
Mike WallaceBAL1610912.84712.749
Brian QuickLA167512.75012.550
Kelvin BenjaminCAR1611212.65212.551
Kenny BrittLA1510912.55312.452
Chester RogersIND93112.75012.453
DeAndre HopkinsHOU1613812.45512.354
Davante AdamsGB1611812.45512.355
Jordan TaylorDEN112412.55312.356
Aldrick RobinsonATL133112.45512.257
Terrance WilliamsDAL156112.25812.158
Kamar AikenBAL164612.25812.159
Marqise LeeJAX1610112.16012.160
Brandon ColemanNO133812.06112.061
Victor CruzNYG146812.06112.061
Dontrelle InmanSD168912.06112.061
Brittan GoldenARI71411.96412.064
Jeremy MaclinKC127211.86511.865
Bennie FowlerDEN102111.56611.866
Walt PowellBUF72311.46811.767
Ryan GrantWAS111811.37011.768
Taylor GabrielATL124611.56611.769
Allen HurnsJAX116711.46811.670
Tyler LockettSEA146211.37011.571
Robert WoodsBUF137111.37011.572
James WrightCIN101910.88211.573
Jeff JanisGB101710.78311.574
Ricardo LouisCLE73211.07911.475
Michael CrabtreeOAK1614211.37011.476
Odell Beckham Jr.NYG1616111.37011.477
Adam ThielenMIN168811.27511.478
Kendall WrightTEN114111.07911.479
Nelson AgholorPHI156211.17611.380
Donte MoncriefIND95411.07911.381
Demaryius ThomasDEN1614111.17611.282
Antonio BrownPIT1515111.17611.283
Andre HolmesOAK142310.39011.284
Malcolm MitchellNE114510.68411.185
Corey BrownCAR154510.58511.086
Danny AmendolaNE112810.09410.987
Chris ConleyKC166510.58510.988
Dorial Green-BeckhamPHI146710.48810.889
Justin HardyATL11309.89610.890
Pierre GarconWAS1611010.58510.891
Jordan MatthewsPHI1410910.48810.792
Brandon LaFellCIN1610110.39010.693
Braxton MillerHOU9279.310210.694
Charone PeakeNYJ11339.59910.595
Cameron MeredithCHI149510.19210.596
Sterling ShepardNYG1610310.19210.597
Amari CooperOAK1612710.09410.398
Jaelen StrongHOU6238.511010.399
Eli RogersPIT13659.69710.2100
Quinton PattonSF13599.410010.1101
Quincy EnunwaNYJ16999.69710.1102
Jordan NorwoodDEN14328.610910.1103
Seth RobertsOAK16759.410010.0104
Rod StreaterSF13278.211410.0105
Tyler BoydCIN16769.21039.9106
Cody CoreCIN6267.51219.7107
Steve L. SmithBAL141029.01059.6108
Jeremy KerleySF161089.01059.5109
Julian EdelmanNE161469.11049.5110
Doug BaldwinSEA161178.81079.3111
Stefon DiggsMIN131098.71089.3112
Andrew HawkinsCLE16508.01189.3113
Tyreek HillKC16818.41129.2114
Mohamed SanuATL15788.31139.2115
Golden TateDET161318.51109.0116
Jamison CrowderWAS16928.21149.0117
Albert WilsonKC14507.51219.0118
Michael A. ThomasNO151198.21148.8119
Eddie RoyalCHI9446.91248.7120
Bryan WaltersJAX8316.11288.7121
Larry FitzgeraldARI161478.11178.6122
Willie SneadNO15967.81198.6123
Tavon AustinLA151017.71208.5124
Randall CobbGB12807.31238.4125
Anquan BoldinDET16886.61267.7126
Jarvis LandryMIA161226.81257.6127
Cole BeasleyDAL16926.51277.6128
Adam HumphriesTB15776.11287.5129
Cordarrelle PattersonMIN16644.81306.8130

DT : IR :: TL : DR

Deploying my normal procedure for reliability analysis, I found that a WR’s average depth of target (aDOT) stabilizes in only 4 games (or 24 targets or 117 routes run). This is the fastest of any WR stat I’ve examined thus far and about 10 times faster than yards per target (YPT). Comparing the reliability timelines of aDOT and YPT reveals that, by the time a WR’s actual YPT represents half skill and half luck, his actual aDOT represents about 90 percent skill and only 10 percent luck. Taken together, my results suggests that we should be using aDOT — or at the very least, yards per route run (YPRR) — rather than YPT for the purposes of predicting WR yardage production.

Email to someoneShare on Facebook0Tweet about this on TwitterShare on Reddit0

  1. The formula is (Games/2)*[(1-r)/r]

  2. The formula is [(Observed Performance * Games) + (League-Average Performance * Stabilization Point)] / (Observations + Stabilization Point)  

  3. Weighted by group size. 

Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *