The National Football League is littered with unlikely success stories.

Quarterback Kurt Warner took until age 29 before he was given a real chance to start in the league, and then he engineered two career comebacks that included multiple Super Bowl appearances with both the St. Louis Rams and Arizona Cardinals. Seattle Seahawks running back Marshawn Lynch went from first-round pick to waiver-wire fodder and back to offensive dynamo and Super Bowl hero in an era obsessed with the pass.

While these unheralded heroes are impressive -- especially due to the rarity of such big misses -- in no part of a football team are there more surprise success stories than the defense.

While the top tiers of defensive linemen are littered with first-round picks, there are a fair amount of mid-round and later draftees who come to make impacts on the NFL. For every high-value J.J. Watt and Chandler Jones who goes off right away, there is at least one mid-round Jurrell Casey who develops into something that scouts whiffed on.

With so much seeming randomness in who breaks out as a defensive lineman in the NFL, how can we project these players into the pros?

Theory

First, let me begin with a caveat. This was a massive data project intended to look at tendencies and trends. While people may point to the results of any study and demand to know why we ignored the exceptions to the rules, the reason is exactly that: they are exceptions, not rules. Numbers may be massaged, altered, and shifted to show what we want, but over a large enough sample, the biases of the analyst tend to fade away.

I hope that’s abundantly clear when I tell you that the results of this study are not what I expected at all.

To begin, I wanted to make sure I had an accurate and representative sample size. For this study, I selected every player who was drafted as a defensive end or defensive tackle -- scheme notwithstanding -- between 2009 and 2013. This way they, would certainly be the position I was looking for -- rather than a defensive end who transitioned from outside linebacker -- and have a three-year body of work to stabilize their production in the NFL.

I took their NFL Combine or pro day data points from NFL Draft Scout, adjusted them for a player’s weight and height thanks to Football Perspective, and calculated the correlation between athletic measurements and their games played, sacks, tackles, and round drafted in the NFL Draft, courtesy of Pro Football Reference.

A correlation means that the relationship between two things is stronger when it's closer to 1 or -1 and the relationship is weak or non-existent when it's closer to 0.

Everyone has his or her theories about what makes for a successful defender in the NFL. But can we even project them from their physical attributes?

Practice

All in all, this led me to 152 correlation calculations (19 athletic factors versus 8 measures of success), which I am not going to fully list below, for both the sake of your eyes and my fingers. If you’re interested in the full depth of my defensive depravity, you can find the total spreadsheet here.

From this enormous study, a few revelations emerged, and I want to discuss those in some depth.

1. The 40-yard dash matters. A lot.

In fact, the adjusted 40-yard dash was the only metric that was broadly predictive for defensive linemen as a whole. This measurement has one moderate correlation to sacks (-0.28, meaning that the lower the 40 time, the more sacks) for a defensive lineman. When we look just at defensive ends, though, there are strong correlations between the adjusted 40 and draft round, career fantasy points, and sacks. For edge rushers, surprisingly, the drill actually matters a fair amount. The vertical jump is much less correlative but still seems to hold some weight when looking specifically at defensive ends (0.26 correlation to sacks).

2. The bench press doesn’t matter. At all.

Across the board, there were no meaningful correlations for the bench press, showing that strength on the barbell doesn’t translate to game strength.

3. We should not ignore what teams are telling us; draft results are important.

Many people in the NFL Draft community have tried to get ahead of the curve on some players by saying that we should stick solely to our valuations and ignore the way players fall in the draft. The data does not bear that out. The table below shows the correlations between NFL Draft round and various measurements of success.

The negative on the correlations means that the closer the number of the draft round is to 1, the more production they racked up in the various categories. Whether or not it’s right, NFL teams give more playing time to the players they draft highly and sometimes ignore the athletic marvels (hey, Grady Jarrett). If your favorite edge rusher or defensive tackle goes in the first round, grab them in your fantasy rookie drafts; chances are, they’ll be given a long leash to try to succeed.

4. Sacks are the only measure of success even moderately correlated to athletic measurables.

Across the board, sacks correlate moderately well to a number of athletic measurables. This is the only success metric that does so, and it means that if we are hunting for sack production from these defensive linemen, the more athletic ones will be the ones to turn to. The table below shows these correlations.

So, we can see that nine of the 15 categories above have moderate (0.25 or more) correlations to sacks for defensive linemen. The strongest ones are the raw 40-yard dash, the raw three-cone, and raw broad jump. The 40 and broad jump fall into the category of explosiveness drills, so while we perhaps want to see lightning-quick blitzers dance around tackles, the real value comes from straight line power generated. The three-cone is an agility drill but is the only such drill that shows any real association with sack generation.

So, while it may seem that the NFL Draft is impervious to our inquisitions, we can find a few ways to look at the defensive linemen and separate them from the rest of the pack.

In addition, this should serve as a reminder: for all that we try to learn, we can never know everything the NFL does before it’s time to make their selection.

We should trust NFL general managers -- if for nothing else -- that they’re going to use their first-round picks early and often.

That knowledge can help our fantasy teams greatly.