Loading scores...
Columns - Magazine

Seasonal Stats Evolution

by Jonathan Bales
Updated On: October 4, 2018, 4:09 pm ET

This article started out as a simple question for me—how do fantasy stats change over the course of the NFL season?—that I figured would have a simple answer.

 

To determine how play-calling and/or offensive efficiency shifts as the season progresses, I used Pro Football Reference’s Game Play Finder to sort stats into various buckets based on when games occurred. When I perform any study like this, I just do a very simple test to see if there’s anything of interest before I dig deeper. Sometimes I have really cool ideas—or what I think are really cool ideas—that end up amounting to nothing because there’s nothing really interesting or actionable in the data.

 

When I ran the initial numbers for this particular study, I just sorted some simple stats from all games since 2010 into two buckets based on the week number: Weeks 1 through 9 versus Weeks 10 through 17. Here are the results.

 

FS1

 

Okay, so there’s pretty clearly something here. In the second half of the NFL season, offenses throw the ball less frequently (and with worse efficiency) than early in the season, and they run the ball more often (and with greater efficiency). Of course, opportunity multiplied by efficiency equals fantasy production, so we of course see fewer total passing yards and more total rushing yards late in the year.

 

I already think this information is interesting and suggests something significant is happening with play-calling and success rates, but I thought it would be even more telling to break things down more. This is the same data—all games since 2010—broken down into four buckets. Here’s a look at overall play-calling.

 

FS2

 

In each bucket, passing attempts decrease and rushing attempts increase. Meanwhile, the effect is similar (but not as strong) for efficiency.

 

FS3

 

Offenses certainly run the ball worse in the beginning of the season and pass it worse later in the season, but the effect doesn’t appear to be as linear as it is with play-calling. That is, the first quarter and final quarter of the season are much different from one another in terms of play efficiency, but the middle part of the year—Weeks 5 through 13—appears to treat offenses the same throughout.

 

With the play-calling and efficiency trends, here’s how the overall numbers shake out.

 

FS4

 

NFL offenses average around 255 passing yards through the first four weeks of the season. That number drops about 10 yards per game in Weeks 5 through 13, then dips another 10 yards in the final quarter of the season. On average, a team that totals 255 passing yards per game in September can be expected to throw for just 235 yards per game in December. That’s a meaningful decline.

 

The reverse effect is apparent with rushing yards.

 

FS5

 

Rushing yards per game are basically a mirror image (in the opposite direction) of passing yards—an early jump, sustained production, and then another late jump. On average, a team that averages 109 rushing yards per game in September will run for about 10 extra yards per game in December.

 

Conclusions

 

There are probably a few reasons that passing production declines and rushing production improves throughout the course of a season, but I think what we’re seeing is the effect of inclement weather. I’ve done a lot of research on this in the past—namely, how much precipitation, cold temperatures and wind affects fantasy stats.

 

The numbers from this analysis support a hunch I’ve had that anecdotal evidence has supported: most “inclement” weather doesn’t actually affect football stats all that much, but harsh winds can wreak havoc on a game.

 

The reason I think this new data supports that idea is how it matches up with increases in wind speed across the country. I researched the average wind speed, temperature and precipitation by month for a bunch of cities, and we see the same general trend in many of them: a steady decline in temperature in fall and winter, a spike in precipitation in November, and a jump in wind speeds in December.

 

Here’s a look at a normal temperate progression by month (this graph is for Boston).

 

FS6

 

If temperature had a major impact on fantasy stats, we’d probably see a steadier decline in efficiency than we do. Meanwhile, here’s a look at Boston’s average wind speed.

 

FS7

 

It’s the highest in December. Note that while a difference in wind speed of a couple miles per hour probably isn’t massive, the higher average is representative of the fact that certain times of the year come with significantly worse weather. It would be one thing if the wind speed were exactly 7 mph at all times, but it’s probably lower than that on many occasions and then significantly higher at times. It’s those times with really harsh winds that we see fantasy production plummet, and it’s likely the outlier poor performances that skew the overall fantasy numbers. In short, strong winds significantly decrease the floor production of quarterbacks and pass-catchers.

 

None of this is to say that you should be sitting Aaron Rodgers or Tom Brady late in December, but I think there’s enough anecdotal evidence at this point to conclude that passing efficiency declines late in the season, and wind speed is probably the main culprit. I’m not going to use wind speed as a standalone tool for fantasy football decisions, but it’s certainly a consideration for me, especially in tiebreaker situations.