According to the Data: Using Analytics to Generate Value in Fantasy Football

NFL teams could learn something from us fantasy football owners. While organizations are finally trending toward analytics to make data-driven decisions, fantasy owners have been crunching numbers for years. Many argue football is too complex a game to be fully encapsulated by stats, but that's missing the point. We don't need numbers to explain every little nuance of football - whether it be the real or fake variety - for them to work.

NFL GMs and fantasy owners alike are in the business of making predictions. We predict the future performance of players and determine how their cost - whether it be a $10 million salary or a second-round draft pick - matches with their expected production. When that expectation exceeds the cost, we get value, and that value ultimately leads to success.

Thus, whether running an NFL team or competing against your Uncle Bruce in a fantasy football league, the only requirement for a particular metric to be useful is that it's predictive. If it can predict the future, it can be leveraged into value. The more accurately a stat can predict the future, and the more strongly it's correlated with NFL success, the more heavily we can weight it in our projections.

Over the last few months at RotoWire, I've been searching for the most predictive stats for each position. Do mobile quarterbacks really perform better? How much does speed matter for running backs? Such questions are important, and they can strongly influence our preseason fantasy rankings. In answering them, I've been attempting to reveal the "prototypical" player at each position. All other things equal, which traits do we want a quarterback or running back to possess?

Fantasy football production comes in all shapes and sizes, but our job is to maximize the probability of hitting on a particular player. If you understand which characteristics typically lead to success - those that make up the ideal prototype - you'll have a head start on the competition.

In the sections that follow, I'll label the three traits I've found to be most foretelling of success for each position, providing you with two players whose numbers suggest they could offer value in fantasy drafts and one whose stats hint to a potential down year. Note I'm not arguing the ideal players are risk-free selections or that the players to avoid are incapable of elite production; it's all about value, and I'm simply building an archetype for each position and using it to uncover that value.

Quarterback

A Wide Window of Opportunity
You're going to read about age for every position because I think it's absolutely vital in projecting players. While every player is different, it's essential to know when players at each position typically break out and break down. That's particularly true in keeper and dynasty leagues. And because most owners don't factor age into their rankings too heavily, it can be leveraged into a competitive advantage.

Quarterbacks have the largest window of opportunity of any position. Historically, quarterback play has peaked between ages 25 and 30. After a small drop in productivity in the early-30s, most quarterbacks can sustain a high level of play until around age 37.

On the Move
Although pocket passers have outperformed mobile quarterbacks in the past, the new breed of truly mobile passers has been a totally different story. There's no substitute for a quarterback who can throw the football - it's an essential trait - but nowadays, we have a lot of passers who can both pass and run. They've proven safer fantasy options than pocket passers because they have more ways to beat defenses, meaning they're relevant throughout games, regardless of the situation.

A Heavy Workload
If you want a productive fantasy quarterback, you absolutely have to emphasize a heavy workload. There's such a small deviation in quarterback efficiency; the league's top quarterbacks average just a yard or so more per attempt than mediocre ones. Meanwhile, some quarterbacks throw 300 more passes than others, and there's no amount of efficiency that can overcome that difference.

Ideal Values: Cam Newton & Matthew Stafford
Heading into his third season, Newton is set to explode in 2013. He'll be 24 when the season begins and has obvious mobility. And while Newton has averaged about 500 passes in his first two seasons in the league, don't forget his workload has been bolstered by an extra 125-plus attempts on the ground.

Stafford's 727 passing attempts in 2012 were the most in NFL history, but he failed to dominate the fantasy realm due to a 59.8 percent completion rate and only 6.8 YPA. Stafford can't be counted on for 700 attempts, but he's still good for 650-plus in Detroit's pass-happy offense. Entering the prime of his career, Stafford's efficiency metrics should improve enough to make up for any dip in workload.

Avoid: Robert Griffin III
Yes, RGIII is obviously mobile, but we really need to see how the Redskins plan to use him coming off of a knee tear. All signs point to a more conservative approach. Plus, Griffin was so efficient in 2012 that, barring a massive jump in attempts, he's primed for some regression.

Running Back

A Young Man's Game
While quarterbacks get better with experience, running backs are actually at their best almost immediately after they enter the NFL. Running back efficiency in terms of YPC peaks at age 22. Yes, the most likely season for peak running back efficiency is his rookie year. From there, it's a steady decline. By age 29, most running backs are rushing for just three-quarters of their career-high efficiency and overall production.

Speed Kills
NFL teams seem to value speed more for wide receivers than for running backs, but it's actually more predictive for the latter position.

In terms of approximate value - a metric that combines various stats relevant to fantasy football - running backs who have recorded a 4.50-plus 40-yard dash at the scouting combine have been largely ineffective in the NFL. You have your occasional Frank Gore-esque outlier, but for the most part, the best running backs are the fastest ones.

Workhorse Runners
The difference between an elite running back and an average one can be less than one yard per carry. That means the best fantasy backs are typically those who get the ball the most. Look at the 16 backs who rushed for 1,000 yards in 2012; with the exception of C.J. Spiller, each had at least 220 carries.

Ideal Values: Trent Richardson & Lamar Miller
I can't emphasize enough how important it is to seek youth in running backs. You want to look for backs whose past production doesn't measure up with what they'll do in the future. Both Richardson and Miller are young, talented backs who are expected to see heavy workloads in 2013.

Avoid: Steven Jackson
Jackson has been a fixture in the first few rounds of drafts for nearly a decade, but he'll be 30 when 2013 begins, and he's racked up 2,395 career carries. He could see better efficiency in Atlanta than he did in St. Louis, but he's not going to receive the same heavy workload in the Falcons' pass-first offense. Jackson might have a solid season, but he's not worth his current early-second round ADP.

Wide Receiver

A Long Wait
Since 2006, only six rookie receivers have finished in the top 24 at the position. That's pretty amazing, and it highlights that receivers typically take a long time to develop. Although it's popular to use a late-round pick on a rookie receiver, that selection would probably be better spent on a rookie runner. Most wide receivers don't even reach 90 percent of peak production until age 25 or later.

A Tall order
The average height and weight of the top-10 fantasy receivers in 2012 is 6-3, 220. The wide receiver position has been taken over by the big boys, primarily because they stay relevant in all areas of the field. While smaller slot receivers have become in vogue, they don't typically produce much inside the red zone. When the difference between 12 touchdowns and four touchdowns is the equivalent of 480 receiving yards, emphasizing height and weight is a smart move.

Moderate Speed
Contrary to popular belief, not all receivers need blazing speed. Actually, the fastest receivers have historically been just barely more productive than those with moderate speed. A bunch of big receivers with average speed have succeeded recently - Brandon Marshall (4.52), Kenny Britt (4.51), Jordy Nelson (4.51), Dwayne Bowe (4.51) and Hakeem Nicks (4.50) - but they're all very large. The cutoff seems to be in the 4.55 range.

Ideal Values: Dez Bryant & Torrey Smith
Both Bryant and Smith had big seasons last year, but their ceilings are so high they're still offering value in 2013. Both are big receivers with decent speed heading into the peak seasons of their careers.

Smith failed to improve much upon his impressive rookie numbers, but this is still a 24-year-old receiver with good size and speed heading into his third NFL season. Now the clear No. 1 option in Baltimore, Smith and his gaudy 15.2-percent career touchdown rate should provide value at his current mid-round ADP.

Avoid: Percy Harvin
If you play in a PPR league, Harvin can be gold. In standard scoring, however, he's not nearly as valuable. That's because he becomes relatively ineffective in the red zone; his career high in touchdown receptions is just six, and he still hasn't had a 1,000-yard season, though he was on pace for 1,200 last season before getting hurt. His rushing ability makes him a relatively safe pick, but his upside is limited by his size (5-11, 184).

Tight End

A Narrow Window
Believe it or not, tight ends have even fewer seasons than running backs at peak production.

The typical tight end produces 90 percent of his career peak in just four seasons, and 80 percent of his peak in six. Those prime seasons have usually come late - between ages 27 and 30 - meaning tight ends take longer to develop than any position. By age 26, the average tight end is producing only 85 percent of his eventual peak. That means when a tight end is successful early in his career, it's usually a sign of great things to come.

Going to Extreme Lengths
There's almost no correlation at all between tight end speed and production. Instead, height seems to be most predictive of success at the position. Whether a tight end has Vernon Davis-like speed or not, he needs to be tall, preferably at least 6-5.

Efficiency Rules
There's a pretty big gap in YPC between some tight ends simply because the position is used so differently across the league. Some tight ends are now basically wide receivers, meaning they can average 15.0 YPC with ease. Rob Gronkowski's career YPC is actually 51 percent greater than what Jason Witten posted in 2012. While efficiency isn't all that important for quarterbacks or running backs, it holds weight for tight ends. Look for playmakers who don't need 100 receptions to have 1,000-yard seasons.

Ideal Values: Rob Gronkowski & Jared Cook
Gronkowski is really the ideal tight end: big, an excellent red-zone threat and young. Most don't realize that Gronkowski is still just 24. If his career follows the typical tight end career curve, it means he's been producing at only around three-quarters of his peak. Assuming he's healthy in the early weeks of the season, look out. Meanwhile, Cook is an obviously talented player in a place where he can finally flourish.

Avoid: Jason Witten
If there's one player I won't own on any fantasy teams in 2013, it's Witten. Aside from the Cowboys' recent draft moves, and the increased focus on getting the ball to Bryant, Witten just isn't very efficient anymore. He broke the record for catches by a tight end in 2012, but that's only because the Cowboys forced him the ball. In reality, Witten's yards-per-target dropped for the fourth consecutive season. Heading into 2013 at age 31, this could be the year Witten's production plummets, especially in non-PPR leagues.