Advice

How Much Does Week 1 Tell Us About the Rest of the Season?

Now that we’ve got a week’s worth of games under our belt, fantasy pundits are busy trying to extrapolate the rest of the season. How much weight should Week 1 performances be given?

Using the RotoViz Screener, I calculated the relationship between Week 1 performance and rest of season (Week 2 through Week 16) fantasy performance.1 The numbers I’m presenting are on a per-game basis and cover the years 2000 through 2016.2 A full table of relationships can be found at the bottom of this post, following a position-by-position breakdown.

Quarterback

When it comes to predicting rest of season fantasy points per game, the three best Week 1 variables for quarterbacks are fantasy points, passing yards, and expected points.3

Position Variable ROS R-Squared
QB PaYds 0.232
QB PaEP 0.183
QB PPR 0.222

Unfortunately, none of the relationships are very strong. Fortunately, quarterback is a “onesie” position and most fantasy teams are just fine with their current starter.

Running Back


rb

For running backs, the best Week 1 variables are market share of rushing attempts, total rushing attempts, and rushing expected points. Fantasy points (PPR) in Week 1 and total rushing yards are also strongly correlated with rest of season fantasy performance.

Position Variable ROS R-Squared
RB RuMS 0.468
RB RuAtts 0.431
RB RuEP 0.431
RB PPR 0.407
RB RuYds 0.386
RB ReEP 0.173
RB ReREC 0.163
RB ReYds 0.146
RB RuTD 0.127

By themselves, Week 1 receiving stats aren’t very strongly correlated with rest of season performance. Those relationships improve if we narrow our date range to the last five years, as you’d expect. We can also get a nice boost by combining market share of rushing attempts and receptions.

Capture

Combining those two numbers gets us to an r-squared of 0.492, meaning almost half of a running back’s performance in weeks 2 through 16 can be explained by their market share of attempts and total receptions in Week 1.

Wide Receiver

wr

Week 1 target market share, raw targets, and expected points are the strongest predictors of rest of season fantasy points per game at the wide receiver position.

Position Variable ROS R-Squared
WR ReTrgMS 0.342
WR ReTrg 0.312
WR ReEP 0.311
WR ReRec 0.293
WR PPR 0.269
WR ReYds 0.257
WR ReTD 0.06
WR ReCR 0.019

I included a number of weaker variables as well. Note that neither Week 1 TDs nor catch rate has any notable relationship to rest of season fantasy points per game. These are two stats that pack more emotional impact than actual relevance.

Tight Ends

te

Week 1 target market share is also the variable most strongly correlated with rest of season performance at the tight end position. Total targets, expected points, and total receptions also relate quite well.

Position Variable ROS R-Squared
TE ReTrgMS 0.352
TE ReTrg 0.349
TE ReEP 0.349
TE ReRec 0.301
TE ReYds 0.282
TE PPR 0.267
TE ReTD 0.066
TE ReCR 0.007

Just as at WR, Week 1 catch rate and TDs don’t appear to have much to do with rest of season performance.

Takeaways

Perhaps the biggest takeaway is the metrics that don’t show up. None of the efficiency stats, like yards per carry or reception, touchdown or interception rate, or even success rate or points over expectation, achieved more than a 0.20 r-squared. Efficiency stats are nice for describing what happened but don’t seem to offer much as a way to prognosticate the near future.

Market share numbers do seem to be very useful. This should come as no surprise to regular RotoViz readers, but it’s interesting to see it pop up as a valuable within-season measure in addition to its usual usage, which is projecting future seasons.

Actionable Insights

  • Lamar Miller led the league with 74 percent of his team’s rushing attempts.4 D’Onta Foreman, Alfred Blue, and Tyler Ervin don’t appear to be threats yet.
  • Jonathan Stewart tallied a lot (18) of carries (47 percent market share) and garnered a pair of receptions in Week 1.
  • Meanwhile, Christian McCaffrey earned a huge share (29 percent) of Carolina’s pass targets. It’s possible that both Carolina RBs are viable fantasy starters going forward.
  • Ty Montgomery (68 percent) tied for the second highest share of team rushing attempts and added four receptions. Both are very promising for his future performance.
  • Jordan Howard got upstaged by Tarik Cohen, but still earned 68 percent of Chicago’s rushing attempts. It’s obviously bad if he loses a lot of work in the passing game, but he should have a stable floor of rushing attempts.
  • Speaking of Cohen, he’s a priority waiver add based on his incredible 30 percent share of targets. That’s a benchmark only eight players (none of whom are RBs) exceeded in Week 1. Cohen’s eight receptions were fourth-best regardless of position in Week 1.
  • Carlos Hyde, (60 percent), LeGarrette Blount (58), and Ameer Abdullah (58) all earned RB1-level rushing workloads.
  • DeAndre Hopkins’ 16 targets equated to 44 percent of Houston’s total. The workload is certainly there.
  • Surprisingly, Jermaine Kearse earned 24 percent of the Jets’ targets. There are obvious offensive concerns, but he could have some value going forward.
  • Breshad Perriman earned 24 percent of Baltimore’s targets, although the total number (4) was very low. Still, it’s the same share and total that Jeremy Maclin received.
  • Kenny Golladay earned 18 percent of Detroit’s targets. Golden Tate earned 31 percent, but Marvin Jones had just five percent. Golladay (7) and Tate (12) both easily trumped Jones’ two targets. Golladay is an obvious add, but Tate might be undervalued.
  • At tight end, Charles Clay earned 33 percent of Buffalo’s targets. That easily led all TEs, and only Hopkins, A.J. Green, and Amari Cooper had a higher percentage. Eventually Zay Jones and Jordan Matthews might take some work, but for now Clay looks like a valuable commodity.

R-squared table

PositionVariableROS R-Squared
RBRuMS0.468
RBRuAtts0.431
RBRuEP0.431
RBPPR0.407
RBRuYds0.386
TEReTrgMS0.352
TEReTrg0.349
TEReEP0.349
WRReTrgMS0.342
WRReTrg0.312
WRReEP0.311
TEReRec0.301
WRReRec0.293
TEReYds0.282
WRPPR0.269
TEPPR0.267
WRReYds0.257
QBpayds0.232
QBppr0.222
QBep0.183
RBReEP0.173
RBReREC0.163
RBReYds0.146
RBRuTD0.127
TEReTD0.066
WRReTD0.06
WRReCR0.019
TEReCR0.007

Subscribe to the best.

  1. PPR points/game.  (back)
  2. I also calculated smaller and more recent time frames, and the trends were very similar.  (back)
  3. Excludes QBs who posted a zero in Week 1.  (back)
  4. Through Sunday.  (back)
By Charles Kleinheksel | @Spidr2ybanana | Archive

No Comment

Leave a reply