I absolutely abused my opponent last week in part because he lost Aaron Rodgers and had drafted most of the rest of Green Bay along with him (Jordy Nelson, Randall Cobb, Ty Montgomery). It made me curious though about how much worse WRs did when the QB1 was out.
This one is for you, Nick. Let’s see what the rest of your season looks like.
As in my award-winning article cataloguing the effect of DST score on starter vs. backup QBs, I have some caveats:
- Everything I’m going to do here will be as reasonable as I can be, but there’s a chance you’ll disagree with my methods. If you want to tweet at me (@FFDataStream) I’d love to hear from you and we can discuss my analysis (and life) choices.
- I had to decide how to split games with and without starting QBs. I chose to only keep the unambiguous games, i.e., games in which ONLY the starter (QB1) played and games when ONLY a backup (QB2) played. There are games like last week in which Rodgers went out early and my filtering treats that exactly the same as if it was a blowout and they just brought in the backup for the last few minutes of the 4th quarter because the game was over. In my explorations I found very little difference between keeping or discaring these, but I wanted to remove as much ambiguity as I could and hoped to find larger effects. This leaves me with 908 games with QB1 only and 146 games with QB2 only, tossing 152 games in which both played at least a little (note: this is for the 2015, 2016, and 2017 seasons, up through week 6).
- I had to decide how to rank each wide receiver (WR1, WR2, etc.). I made this decision based on the mean fantasy points for each player on a team in a given year as opposed to the total fantasy points over the year. This leads to the occasional weirdness; Geronimo Allison has a mean of 7.3 fpts this year while Randall Cobb has 7.2. I would probably consider Cobb to be the WR3 on the team, so this method isn’t perfect. On the other hand, just total fantasy points also didn’t work great since there were some WRs who were out for enough games to push their ranking up or down. I’m sure there’s a nice, weighted system that tries to solve these, but I’ve made my choice and now you all have to live with it.
- There’s a TON to unpack in this data set; total fantasy points to each WR, points given up by opposing defenses, games when the top receiver was out, games when the #2 guy was out, and on, and on, and on. I’ve done some of this on my own already, but I’m not going to show you all of it. It would take too long and some of it just isn’t interesting. The short version of the most intersting finding so far is that the WR1 on a team tends to get about 40% of the fantasy points, WR2 gets 30%, WR3 gets 20%, and 10% for the rest of the receivers. This varies heavily based on the team, of course, with Seattle, Kansas City, and the Chargers spreading the ball around much more evenly and Cleveland, The Jets, and Miami giving closer to half of the points to the WR 1 (with honorable mention to Pittsburgh, shout out to my guy Antonio Brown!). Again: neat, but not the point of today’s analysis.
- I’m going to be looking for statistical differences, not narrative ones. If you know that Jordy Nelson and Brett Hundley have been playing catch in the park every Tuesday for the last 6 years, then there’s a chance that any horror that the statistics predict may not come to pass. On the other hand, my guy Antonio Brown has never caught a TD pass from a QB other than Ben Roethlisberger. Although I’m calling out Green Bay and Pittsburgh as specific examples here, I’m just going to be showing the average (actually median) change rather than break it out by team. Not all starters and backups are created equal.
First, how much of a downgrade is it overall with the backup QB in:
That looks pretty clearly like the QB2 is a downgrade. The tail is certainly longer when the QB1 is in, and the peak is slightly higher as well. Since those distributions aren’t Gaussian, let’s just look at the median. I found that the median number of points that all WRs get when the QB 1 is in is 21.7, and 18.4 when the QB2 is in. That’s a difference of 3.3, not too bad considering how terrified people usually are when they find out the backup is coming in. All backups aren’t created equal and I can think of a few who have done alright when they’re brought in.
So, definitely a downgrade, but for whom? Let’s see that same plot, but split out by WR rank:
It looks like the WR1 might take the biggest hit. The tail is dropped at the high end for the WR1, but looks mostly unchanged for WR2 and WR3. The peak also looks mostly unchanged for WR2 and WR3, though WR3 is missing some of the middle section when the QB2 comes in. Here’s a table of the median values:
|rank||QB1 only||QB2 only|
That’s neat. Remember that point difference from above? It looks like the WR1 takes the biggest hit with a drop of 2 points when the backup QB comes in. Interestingly, the WR2 stays almost the same. WR3 drops, but by very little, less than half a point. If we’re dipping into narratives, I would probably say that the WR1 is the top receiver because of the rapport that they’ve built up with the QB1. The starter QB knows that he can trust his top receiver and throws to him when in doubt. The backup doesn’t have this relationship and so spreads it around more.
With this in mind, adjust your hopes for Jordy accordingly, Nick.