Should We Look for Quantity or Quality in Fantasy Football Quarterbacks?

Also: here are a few reasons why I'll be drafting a lot of Eli Manning this year.

My earliest exposure to sports journalism was when I had a subscription to Sports Illustrated for Kids.

As a child, I pored over every page so that I could absorb as much sports news as I could handle. I still remember one issue in particular that had a collaged image on the cover of a Frankenstein-esque hybrid of the “perfect athlete”: Tiger Woods’ eyes, Mark McGwire’s torso, Wayne Gretzky and Mia Hamm’s legs, Brett Favre and Sammy Sosa’s arms.

If only athletes were like Mr. Potato Head toys.

This is an interesting idea for fantasy football, though; if you could build the perfect fantasy quarterback, which components would you use? I’d probably take Cam Newton’s legs, Peyton Manning’s eyes, Ben Roethlisberger’s torso, and Aaron Rodgers’ arm -- I’ve given this some thought.

Since we can’t pick and choose pieces of players, however, we have to think through what statistics make for better fantasy options at the position. This, to me, boils back down to the age-old question: do you prefer quality or quantity from your fantasy quarterback?

It’s Alive!

This all comes back to a fundamental question of how you play fantasy football. In recent years, that’s meant the decision to abide by the credo of the “Late Round QB” theory or not; this is a whole other conversation entirely. I don’t even want to be considering draft value for quarterbacks at the moment. Instead, think only about what creates real production for your fantasy signal-callers: are you one to go for the high volume throwers who might chuck 20 interceptions in a season, or do you aim for the options who will barely throw 15 touchdowns but won’t turn the ball over? Which is better?

To answer this question, I concocted a simple study using a few key quarterback statistics and assessing which ones correlated best with fantasy points. Using R correlation -- which functions on a scale of 0.00 (no statistical correlation) to 1.00 (perfect statistical correlation) -- I assessed which statistics will be most predictive of a good fantasy season.

I compiled data from each quarterback season with 100 or more drop backs since 2010, looking at total drop backs and completions as quantity categories, and completion percentage, Passing Net Expected Points (NEP), per-play Passing NEP, and Total NEP (Passing NEP plus Rushing NEP). NEP helps us take the numbers we get from the box score and assign them contextual value so they relate even closer to the game on the field. By adding down-and-distance value, we can see just how much each play and each team as a whole influence the outcome of games. For more info on NEP, check out our glossary.

I Will Cause Fear

How does this all break down when we examine the machinery behind a fantasy passer? The table below shows our correlations between each statistic and annual total fantasy points from 2010 to 2014. What do we find matches up the best?

Drop Backs0.90Quantity
Completion %0.60Quality
Passing NEP0.77Quality
Passing NEP/P0.73Quality
Total NEP0.81Quality

To put this simply: holy crap. I have never seen a higher collection of correlations since I’ve done statistical research, but this really knocks it out of the park. Among quality, we have a broad range of predictive values for fantasy points.

Completion percentage gets the lowest mark, likely because -- as a rate stat -- it doesn’t take into account injury-shortened or short-term substitute seasons. Passing NEP and its per-play variation are the next lowest; per-play Passing NEP is again a rate statistic, and Passing NEP fails to account for rushing value, which is factored into fantasy points.

Finally, even Total NEP, which assesses both passing and rushing value, isn’t the highest metric on the board. It’s very good -- 0.81 is well above the “very strong relationship” threshold in statistical terms -- but it’s not the best in terms of identifying fantasy points, which can often be racked up in garbage time, which NEP weeds out.

By far the best possible measures you can use to predict fantasy value for a quarterback are your basic counting stats of quantity: drop backs and completions. It’s simple: the more chances a player has to make value, the more value they will make. That’s why when -- just for kicks -- I checked the correlation between touches (drop backs plus rushing attempts), the correlation was somehow even stronger. When you stretch it out to three decimal places, completions had a 0.915 correlation with fantasy points; touches’ correlation actually gave us a 0.917.

A Great and Sudden Change

Don’t shirk quality on the whole, however. Make sure you’re aiming for high-volume players first but then angle for the better passers. Interestingly, there is a certain point where the correlation between touches versus fantasy points and Total NEP versus fantasy points equalizes, and at 150 total touches it levels out at its highest equal point: about 0.89 for each.

To further illustrate that in the higher ranges we can actually be more predictive with quality, let’s aim for the 90th percentile of fantasy points over the past five years from quarterbacks: 300 points. Let’s also see how many 90th percentile seasons in the other two categories we can find that match up with this mark.

There are 11 seasons that scored more than 300 points with at least 670 touches -- its 90th percentile mark. There were 18 seasons that scored more than 300 points but had a minimum of 118 Total NEP.

Once we look at the elite, we not only want volume, we also want value. The only quarterbacks who have met the 80th percentile of each of these categories multiple times are Drew Brees, Matt Ryan, Peyton Manning, and Tom Brady. This isn’t to say that Aaron Rodgers isn’t still one of the best fantasy quarterbacks available, but he's only had to touch the ball 600 times once in his career to reach 300 fantasy points; he's the hyper-efficient outlier, not the norm.

What does this mean for fantasy owners on a practical level? Especially if you decide to draft with the "Late Round QB" theory, your first goal should be to target quarterbacks who throw for high volume or are in high-volume offenses. More opportunity to be on the field simply means more production for you over the full course of a season. This is why players like Matthew Stafford -- despite ugly Total NEP marks the past few years -- are still very fantasy-relevant.

A few others to target who will be cheaper but have had good volume over the past few years: Tony Romo, Ryan Tannehill, Ben Roethlisberger, and Eli Manning.

My lesson to you: don’t shirk value for volume entirely, but go for quantity over quality when building your perfect fantasy passer.