DFS NFL Research: What Drives Fantasy Points In The Passing Game?
A few days ago I started off my NFL research series by looking at the correlations between fantasy points rushing and other statistics. The conclusions of that study are pretty much what I expected: higher margin of victory means more rushing attempts and production, and therefore more expected fantasy points rushing. This supports the theory that teams that are big favorites are great for targeting RBs. Today, I’m going to look into fantasy points receiving.
There were a few details I had to work out before diving into the stats. First off, I decided to look at fantasy points receiving alone, rather than count both QB and receiving fantasy points. I did this mostly because I felt like fantasy points receiving would have more interesting relationships with other statistics because of points gained from receptions. I also used DraftKings scoring, sans bonuses, for that same reason, as they give a full point per reception.
The first thing I did is look at the correlation between fantasy points receiving and team score. The correlation was just as strong as with fantasy points rushing.
> cor(gameData$Score, gameData$FPReceiving)
Not so shockingly, the correlation between margin of victory and FP receiving was much less than we saw with FP rushing. MOV stands for margin of victory.
> cor(gameData$MOV, gameData$FPReceiving)
Since margin of victory is ostensibly correlated to score (the more you win by, the more your team is likely to have scored), it’s not surprising to see a positive correlation between the two stats. But the big drop off between the correlations of score and margin of victory with FP receiving suggests that losing can often result in more fantasy points receiving than winning.
To look at how margin of victory was really impacting receiving fantasy points, I decided to compare margin of victory to fantasy points receiving… per point scored (for short, FPRPPS). Calculating FPRPPS is simple, I just take the fantasy points receiving and divide it by team score. What this stat shows me is how many fantasy points receiving a team is “milking” out of its offensive production. High FPRPPS implies a team got a lot of fantasy production out of not a lot of scoring. The relationship between margin of victory and FPRPPS ended up being a bit shocking.
> cor(gameData$MOV, gameData$FPPerPReceiving)
The relationship is strongly inversely correlated. The more a team loses by, the more fantasy points receiving we can expect per point they score. This has some interesting implications, mainly that a team projected to score less points than another team could have more expected fantasy points passing if one team is a big favorite and the other is a big underdog.
If we plot the data, we can see the inverse relationship quite clearly, but it gets more intense as teams lose by more than a touchdown.
The relationship between margin of victory and FPRPPS looks linearly inverse until we get to teams losing by 7 or more, then the graph gets out of control. Teams that lose by 7 or more are routinely getting twice as many fantasy points receiving per point than teams that win.
So are teams that lose by 7 or more actually gaining more fantasy points receiving than teams that win?
> mean(gameData$FPReceiving, MOV < -7)
> mean(gameData$FPReceiving, MOV > -7)
The answer is no, but not resoundingly. Score still dominates here. Teams that win by larger margins are more likely to score more points, and therefore get more fantasy points receiving. But what happens when we limit the team score to a certain range, say 20-30 points?
mean(gameData$FPReceiving, MOV < -7)
> mean(gameData$FPReceiving, MOV > -7)
A bit crazy. Teams that lose by 7 or more points actually score over 10 more fantasy points receiving than teams that lose by 7 or less points or win the game when they score between 20 and 30 points.
Let’s say we have a game with team A and team B. The over/under of this game is 50, and team A is favored by 10. This means team A is expected to score 30 points, and team B is expected to score 20 points. Who is expected to score more fantasy points receiving, assuming we know nothing more about each team? Using this research, the answer appears to be team B.
If we do a simple linear regression between team score and fantasy points receiving, we get a slope of .846. This means for every point scored, we expect a bit less than one fantasy point receiving.
Because teams that lose by 7 or more points on average score approximately 10 more fantasy points receiving than teams that win or lose by 7 or less, team A does not surpass team B in fantasy points receiving based on 10 more points in score.
While score is still a large driver in expected fantasy points receiving, the amount of fantasy points a team expects to score receiving is partially driven by the production of the opposing offense. The more an opposing team scores, the more a target team has to throw the ball and the more fantasy points receiving we can expect. When looking for teams to target for receiving fantasy points, look for good passing teams that are big underdogs.
View all posts by Daniel Steinberg