Advice

The FF Industry Is Drawing All the Wrong Conclusions About the 2016 Season

Throughout the offseason, I wrote a series of articles focused on league-level trends. My interpretation of the impact they would have on the 2016 season was flat wrong. Naturally, I’ve been curious to understand why.

My curiosity has piqued with the season over and, with it, a rush of opinions on Twitter and elsewhere about the viability of Zero RB as a fantasy football draft strategy. If you follow FF Twitter at all, you’ve almost certainly seen a lot of explanations since the conclusion of the season discussing what 2016 meant.

Given that, and given my research throughout the offseason, I decided to dig into the data and see exactly what happened. My favorite part about data is there are no secrets – you can test a hypothesis and determine whether it was valid or not. In this case, my hypothesis — based on the fantasy scoring results we saw and the types of rosters I saw succeed in my own leagues — was that I was overconfident in the trends I analyzed last offseason.

Testing that hypothesis turned up some extremely interesting results.

Working Backward

There’s a lot to cover here, but I want to start with one major outcome of the 2016 season. There was a huge dropoff in high-end WR performances across every statistical category. Mike Evans led in targets, but his 171 ranked 15th over the past five seasons.

targets-last-five-seasons

Every other season in the past five years has featured multiple WRs with at least 10 more targets than Evans saw in 2016. Not surprisingly, the impact of reduced volume was lower production. Only three receivers had 100-catch seasons in 2016, and none hit 110 for the first time since 2006. Larry Fitzgerald’s league-leading 108 ranks 16th over the past five seasons.

receptions-last-five-seasons

The same was true with yardage. T.Y. Hilton’s league-leading 1,448 receiving yards ranks 15th over the past five seasons.

rec-yardage-last-five-seasons

And with respect to scoring, while Jordy Nelson tied for the second-highest total over the past five seasons with an impressive 14 receiving TDs, only five WRs reached double digits. From 2012-2015, a total of 40 WRs (10 per season) hit that mark.

League-Wide Trends

So what caused the drop in elite WR production? Last January I examined the extent of the league-wide shift toward more passing. Pass attempts have risen significantly in recent years while rush attempts have fallen. Coupled with a rise in league-wide completion percentage due to shorter passing across the league, completions have skyrocketed in the last half-decade.

If elite WR production dropped off so significantly at the expense of RB production, it would almost be necessary that those trends reversed course. Counterintiuitively, that wasn’t the case.

Total plays dipped, but rush attempts fell for the fifth straight season, setting a new all-time low in the process. Pass attempts remained almost entirely stagnant, and an ever-so-slight rise in completion percentage meant there was ultimately one single completion fewer in the NFL in 2016 than the record set in 2015 (11,527 to 11,526). Here is an updated table of that per-team data.1

YearPass CompPass AttRush AttTotal (includes sacks)comp%
2016360.2571.7416.31022.90.630
2015360.2571.9421.51030.50.630
2014350558.7427.81024.30.626
2013346.9566.8433.51040.70.612
2012338.5555.9435.21027.60.609
2011327544.1436.61017.80.601
2010327.8539.743510100.607
2009324.1532.3440.31006.90.609
2008315516.4441.29900.610
2007325.8532.7437.11004.20.612
2006306.1512.2451.510000.598
2005305.9514.5449.21000.70.595
2004305.4511.1450.9999.30.598
2003303515.4453.41002.90.588
2002322.3540.4440.71017.80.596
2001307.8522440.81001.40.590
2000306.4526.5441.21007.50.582
1999308.6540.643710180.571
1998292.2516.3452.31008.80.566
1997294.8524.3454.61020.70.562
1996306.6532.2453.11022.20.576
1995323.9556.64401032.50.582
1994312.1537.7448.21019.40.580
1993298.3514.84531005.50.579
1992275.2478.9439958.50.575
1991285.8498.2438.5972.20.574
1990270.4482.7444.4965.20.560
1989285.7512.1466.71018.30.558
1988273.9504.74851026.70.543
1986286.2516.8482.510420.554
1985282.5515.14871048.80.548
1984288.4511.6493.31051.80.564
1983285.5501.7507.51052.60.569
1981276.6506.4518.110610.546
1980275489.5513.710400.562
1979250.8463.5542.31043.50.541
1978224.2422.5574.110330.531

RB Production

So league-wide trends can’t explain the drop in elite WR production. Weird. Let’s look at RB production.

In an article supporting the Zero RB strategy last July, I shared a visual that I believed got to the heart of the decreasing value of the RB position to fantasy football. Since 2000, as backfields have shifted from hierarchies to committees, the advantage top RBs produce has shrunk from both ends. Top-scoring backs have scored less over time, but RB3 and RB4 types have also scored more.

Updating to include 2016, we see Le’Veon Bell and David Johnson were certainly fantastic, but they weren’t exactly throwbacks to prime Priest Holmes or LaDainian Tomlinson. Here’s that chart with 2016 included. (Note this only looks at per-game averages for backs who played at least eight games.)

rb-ppr-ppg-scoring-by-year-2016

Yes, there was a huge uptick for the RB1 (it was Bell in PPG) this year, but there were no significant changes at most benchmarks from RB10 on down. Ezekiel Elliott was the RB3 at 21.8 PPG, narrowly edging Devonta Freeman’s 21.7 as the RB1 in 2015.

By way of comparison, WR production was down at every benchmark.

wr-ppr-ppg-scoring-by-year-2016

As noted, rush attempts were down across the league. Rushing yardage doesn’t explain the increased production, either (more on this below, but the average team ran for almost exactly the same yardage in 2016 as 2015). We know Johnson and Bell were exceptional receivers, but they were in a class of their own on that front.

So did we shift back toward more workhorses, concentrating rushing production among fewer backs?

Another trend I examined last year was the decrease in rushing attempts for top fantasy backs. The following table showed the raw number of rush attempts for each of the top 30 backs over the prior decade. Note that this analysis did not look at points per game, but rather overall finish.

RB Carries

I noted two major takeaways: the decrease in 300-carry backs — there were as many as five in a season as recently as 2012, but no more than two in 2013, 2014, or 2015 — and the increased viability of RBs with under 100 carries.

Over the six seasons prior to 2012, only three backs finished among the top 30 PPR RBs with fewer than 100 carries. There are 12 examples over the four seasons since.

Updating for 2016 shows there was again just one back over 300 carries (Elliott), though three more crested 290. On the flip side, there were four more backs in the top 30 with fewer than 100 carries.

top-30-ppr-rb-carries-2

There was no Danny Woodhead 2015 performance from a receiving back, but keep in mind that Woodhead, Gio Bernard, and Theo Riddick — three of the most heralded of that archetype coming into the year according to preseason ADP — all got injured. Riddick still finished RB25, near Darren Sproles, James White, and Chris Thompson as sub-100-carry backs. Bilal Powell was RB16 on 131 carries, while Tevin Coleman finished RB19 with just 118.

The point: RB usage wasn’t significantly different in 2016 compared to recent seasons, and certainly doesn’t resemble the ’00s. The major difference in 2016? Touchdowns.

There were more rushing TDs across the NFL this season than in any year since 2008, and the third most since 2000.2 Here are league averages by season since 2000.

 Rush AttRush YdsYPCRush TD
2016416.31742.64.1913.8
2015421.51741.44.1311.4
2014427.81781.34.1611.9
2013433.51806.14.1712.8
2012435.21854.74.2612.5
2011436.61874.34.2912.5
2010435.01831.54.2112.5
2009440.31866.84.2413.4
2008441.21855.34.2114.9
2007437.11774.74.0612.1
2006451.51876.94.1613.3
2005449.21799.54.0113.5
2004450.91865.94.1413.0
2003453.41885.74.1613.3
2002440.71858.14.2214.4
2001440.81788.44.0611.8
2000441.21800.94.0813.3

Overall scoring was actually down from 2015, so that wasn’t the culprit. A whole bunch of TDs just went from passes to runs.

We can examine how that manifested itself by looking at all the 12-plus rush TD seasons over the past five years. Five of the 14 occurrences came in 2016, including the top two.

12-rush-tds-last-5-years

But TDs Are Random, Right?

So passing trends stayed fairly stagnant, but elite WR production was down anyway. David Johnson and Le’Veon Bell were fantastic, but RB production beyond those two wasn’t that far out of line with what we’d seen in recent seasons. And touchdowns shifted heavily toward RBs.

So what happened in 2016?

Variance.

That’s it. I know that sounds like an excuse from someone who was living it up on the Zero RB train, but the data speaks for itself. There’s no secret sauce and we shouldn’t have seen this coming. There are three major ways variance impacted what happened.

  1. Injuries. Per RotoDoc and Josh Hermsmeyer’s preseason article reviewing expected injury rates, RBs drafted in the first five rounds suffered major injuries at an astronomically low rate in 2016.

2. Touchdowns.

3. Whatever happened with the distribution of WR scoring. Maybe elite cornerback play is on the rise, or maybe teams simply spread the ball around more. I’ll note one more thing I looked at last offseason, which was positional reception rates. I’ll spare you more updated charts, but know that the top 100 receiving RBs combined for fewer receptions in 2016 than 2015, while the top 100 WRs combined for more receptions than any year dating back to 2000.3 So it wasn’t a matter of decreased significance of the WR position as a whole, just the concentration of production.

Discussion

This is a look at macro trends, but we know on the micro level that individual players have ranges of outcomes. One season’s outcomes wouldn’t dispel a strategy as strong and well-argued as Zero RB. But in a hilarious twist, those arguing that to be the big takeaway from 2016 don’t even have data on their side; the only real explanation is we saw some elite RBs hit the high end of their ranges and elite WRs fall short of their statistical ceilings.

The main thing 2016 told us is that variance happens. Injury rates can fluctuate, TDs can fluctuate, and the concentration of production can fluctuate. And when all of those things lean one direction, you get an outcome like 2016.

Taking a victory lap on that outcome is like hitting on 18 with the dealer showing a 5, getting your 3, and then gallivanting around the casino telling anyone who will listen that you just have to know when the 3 is coming.

The good news is that ADP is going to shift back toward RBs. Needless to say, I’m very excited to go Zero RB again in 2017.

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

  1. Note that 1982 and 1987 are omitted due to being strike-shortened seasons.  (back)
  2. A remarkable feat considering how much less run-heavy the entire league is relative to a decade ago.  (back)
  3. Which is how far back I went in my initial analysis.  (back)
By Ben Gretch | @YardsPerGretch | Archive

Comments   Add comment

  1. Excellent article; thanks for the great work this season. Data pointing to variance is the best news for Zero RB drafters in 2017.

    Do you know if there much change in the average bust rates between RBs/WRs this year?

  2. Dennis says:

    Really awesome stuff! And yeah, the variance exhibited is making me want to double down on ZeroRB. @FF_Contrarian suggested I allocate about 20-30 of my 200 auction bucks for HPPR on a couple guys like Melvin Gordon, J. Hill, T coleman, and spend the rest on WRs. that was great advice. Ben, your podcasts with @RotoDoc down the stretch were unbelievble for daily picks, you and he must've done pretty well. thanks and congrats.
    im curious...totally an eyeball hypothesis from watching lots of redzone...but could defensive penalty's inside the 5 have helped these RB tds and might that not necessarily be going away? maybe DB's are learning to take the flag and live to fight another play

  3. Eric says:

    Great article Ben. As somebody who had a bunch of WR heavy teams with little success in 2016 it's nice to know that things have a chance to look up next year. Maybe as owners switch to a more RB heavy approach I can take advantage of the swing draft strategy.

  4. Great article. A couple of points....

    So rushing TDs are up as compared to receiving TDs? I have to wonder if rushing TDs from the 1-yard line SPECIFICALLY are up at all. DPI isn't going away. Teams lob it up to covered receivers in the end zone all the time and get bogus PI penalties when the receiver stops to come back to the poorly thrown pass. I would assume that PI calls in the end zone are trending upward, and I wonder if that explains the rise in rushing TDs compared to receiving. This may well be a trend and not simple variance, IMO. I wish the NFL would fix the way DPI is called, personally, but that's a different conversation.

    Something weird happened this year with the decrease in targets to primary receivers and increase to #2's and #3's, and that something crippled Zero RB. I have no idea if this a trend or an aberration. In past seasons, our Zero RB teams were all built upon the premise of purchasing targets, and filling our lineups with as many targets as possible. The targets simply weren't there this year. The article lists this as reason #3 for the failure of Zero RB, behind variance and touchdowns, but my armchair common sense opinion is that it was reason #1. Not only did our lineups lack the distinct advantage at WR provided by target hogging receivers, but our opponents' lower-tier options were more viable starters. I didn't feel like I was losing FF games this season because of my opponents' RBs, but I was losing because I had no real advantage at WR despite the heavy draft investment. Is this because defenses adjusted to being repeatedly butchered by the AB's and Julios of the world? Or did offensive coordinators decide they needed more balance/spreading the ball around in their attacks? I have no idea, but this shift IMO is what went wrong with Zero RB this year. It was completely aberrant to the trend of the last few seasons. Whether this is a new trend or merely one weird year for some reason is what will determine the success of Zero RB in future seasons.

  5. Matt says:

    I agree that the fairly dramatic decrease in targets for the top WR was pretty concerning to me. It's not targets but if you look at the graph on WR PPR PPG Scoring by Year, you can see that the gap between WR 10 and WR 40 dramatically shrank this year. WR 40 has been on a steady rise the last several years, while WR 10 had been fairly steady until last year. Even looking at last year, it appeared that scoring was on a steady rise for pretty much all WR, making it look like drafting WR was the way to go. Even though this was my first year of doing 0 RB, I'm pretty much undeterred due to the decreased bust rate and decreased injury risk of WR, as well as the fact that people are likely to draft RB earlier because of what happened in 2016. Would love to see more analysis of this year and what we can learn from it.

Discuss this article on the RotoViz Forums

3 more replies