Humans are pretty wonderful creatures when you think about it. We’ve figured out how to harness the energy of the world we live in: first through fire, then through electricity, and now through the very matter of the universe. We are the only species on the planet to have invented language, a complex system of idea representation with infinite grammatical and semantic possibilities. We freaking figured out how to fly, to swim, and to go to the moon, all without wings, flippers, or space-gills.

Evolutionary biology is working on that last one; we’ll get there.

Still, despite these amazing discoveries and achievements we’ve made, we still take a lot of things as given and refuse to question them. Nowhere is this willful ignorance clearer than in sports, where we defend old, unproven adages such as “defense wins championships”. Most notably, some of the smartest fantasy football minds insist that if an NFL team has a good defense, their offensive players will be better in fantasy football. Is this true?

Dark Ages

I don’t like to call people out, but I recently heard a highly-respected fantasy analyst on a podcast wondering quite seriously if R correlation, which helps to provide us with a relationship between two variables. This is measured on a scale of 0.00 to 1.00 (either positive or negative). The closer the value is to 1.00, the more direct of a relationship there is; the closer it is to 0.00, the more random the association is.

So let’s get into it! I compiled the fantasy points for each position, for each team, over the past three years (from 2012 to 2014), and then matched them up with their defense’s value production each year as well. The table below shows the correlations between each position and Adjusted Defensive NEP, which shows the expected values allowed by opposing offenses. The more negative the Adjusted Defensive NEP is, the better. What that means is we want to see a correlation of -0.40 or lower (low points allowed, high fantasy points); what do we find?

So, what makes a strong correlation? Anything from 0.40 to 1.00 is considered a strong or very strong correlation, indicating a likely relationship between the two. Our variables, however, sit in the 0.00 to 0.19 threshold: no or negligible relationship. There is absolutely nothing to indicate that defensive value matters to fantasy production. One might have a great defense that routinely gets the ball back to the offense, and if the offense can’t make anything happen with it, their fantasy production will flounder.

To show this clearly, the next table shows the top five teams in average fantasy points from 2012 to 2014 and their average Adjusted Defensive NEP over that time.

If this adage were true, we'd see a lot more teams like Denver in the top-five of this ranking. As it is, however, we have two top-10 teams, and three bottom-third. For what it's worth, the top-five by Adjusted Defensive NEP includes Arizona (30th in fantasy points since 2012) and San Francisco (20th). This seems to shut the door on the old way of thinking pretty soundly.

Reformation

There is one other part of the adage that people claim and it’s that the “quicker” a defense gets the ball back to the offense, the better. I decided to run a few other correlations to check in on this idea of getting the ball back sooner, and found a few interesting things.

The table below shows these variables, including Time of Possession (T.O.P.), Defensive Yards Allowed (D YA), and Defensive Points Per Game Allowed (D PPG).

What we see here is our first strong correlation, a negative relationship between Total Fantasy Points and Time of Possession. What this means is that the longer a team’s offense held the ball, the higher the fantasy points they scored. Well, this is great! That means that if a defense got the ball back more to the offense, they scored more fantasy points, right?

Hold your horses. When we look at Adjusted Defensive NEP and its relation to T.O.P., there’s only a weak correlation that says the fewer points the defense allows, the longer the offense has the ball. We miss a link in here that looks at how good the offense is at hanging onto the ball once they have it. This isn’t accounted for by the defense. Houston had an incredible defensive season in 2014 (-53.43 Adjusted Defensive NEP, 1st in the NFL), but ranked just 20th in fantasy points scored.

To that end, I even ran correlations between Yards Allowed and Points Per Game Allowed. They have negative relationships, which we’d expect, but negligible ones.

There is no statistical evidence that on a larger scale good defense allows the offense to produce better. This notion simply does not hold up when put under scrutiny. So bring your fantasy team into the 21st Century and do away with silly narratives like this.