How to Draft a Running Back

   Earlier in the year, I wrote about how to draft a quarterback and I decided that I should write on a topic more relevant to the Patriots. What do all the successful running backs have in common, and what connects all of the draft busts? The Patriots are in the market for an additional running back and you might be surprised at how on the money they are with their contacts (or you might not be surprised at all). I used a combination of CBS NFL Draft and Pro Football Reference for my statistics and if you ever need draft or stats information, they're two of the best sites available.

   For this project, I assembled data on all college running backs with combine or pro day data, and some backs without, since 2005. My hope was to find a common thread amongst the top performers in the NFL and find out ways to weed out the underperformers. I used Pro Football Reference to make a list of every single running back who have made their debut since 2005 and their average yearly production to determine their contributions at the NFL level. The top five running backs who have debuted since 2005 are:

1. Chris Johnson, Tennessee Titans
2. Adrian Peterson, Minnesota Vikings
3. Maurice Jones-Drew, Jacksonville Jaguars
4. Frank Gore, San Francisco 49ers
5. Matt Forte, Chicago Bears

   After agreeing with the rankings of 280 running backs, I set forth to analyze the data in order to see what connected these players.

   In order to find statistics to define these players, I looked at the following combine numbers:

Height, Weight, 40 Yard Dash, 20 Yard Split, 10 Yard Split, Bench Press, Vertical Jump, Broad Jump, 3 Cone Drill, Shuttle Drill

...as well as looked over the following college numbers:

Career Yards, Career Yards/Carry, Final Year Yards, Final Year Yards/Carry

   After compiling all the data, I was able to see a running back's history. Unfortunately, just looking at the numbers weren't very helpful. How can the 197 lbs Chris Johnson (impact 10) have the same impact as the 217 lbs Adrian Peterson (impact 10)? How come some players with the same 40 yard dash times have completely different careers? Other than factors outside of their control (injuries, varying levels of quality offensive linemen), there are some qualities that make running backs more likely to have success than others.

   I ran regressions (which is statistics for "can Quality Y [NFL success] be determined by Factor X?") in order to see the relationship between NFL success and these different factors. You have to keep in mind that correlation does not imply causation, but you also have to keep in mind that I'm not looking at correlation. I'm looking at the impact Factor X has on NFL success.

   After analyzing all of the data, I found that the following combine qualities have an impact on NFL success:

40 Yard Dash, Broad Jump, 20 Yard Split, 10 Yard Split, 3 Cone Drill

   However, the impact of the 3 Cone Drill depends on the type of running back. If you want a shifty (think: Ray Rice, Jamal Charles, Matt Forte) running back, you want a player with a sub-7 second 3 Cone Drill. If you want a pounding (think: Brandon Jacobs, Adrian Peterson, Marshawn Lynch) running back, the 3 Cone Drill doesn't matter as much. If you want a do-everything running back, a sub 7.1s 3 Cone Drill will suffice.

   10 Yard Splits and 20 Yard Splits also have minimal impacts. 10 Yard Splits will show how a player can get into the open field and the rest of the dash shows what the player will do in the open space- however, the 10 yard split relies on the fact that the player will be untouched. As a result, the impact of these splits on NFL success are small. Honestly, the numbers show that a player just has to break 1.6s in the 10 Yard Split and 2.7s in the 20 Yard Split. While these numbers have a connection with NFL success, the marker is almost negligible.

   As a result, the two combine factors that have the largest impact on NFL success are the 40 Yard Dash and the Broad Jump.The bench mark for the Broad Jump is 9 feet, 8 inches. Anything lower and the player appears to lack the explosion to burst through the holes in the offensive line. That leaves players like Shane Vereen, Alex Green, Mark Ingram and Dion Lewis with large red flags.

   The elite running backs were sub-4.5s in the 40 yard dash and the above average running backs were sub-4.55s. This means that any running back who cannot run sub-4.55s should be examined with extreme caution. That includes Mikel Leshoure, Ryan Williams, and Daniel Thomas.

   Unfortunately, NFL success is not as simple as drafting a running back with a sub 4.5s 40 Yard Dash and a 9 foot 8 inch Broad Jump. There are different weights of importance between the speed and explosion of a player and, hooray for statistics, I'm able to come up with a formula that can help determine if a prospect has the physicals of an NFL football player. Keep in mind that "NFL Physicals" does not mean "NFL Success;" we have to incorporate their college production to see if they can actually play in the NFL.

   However, just looking at the "Physicals" part, the following players have the physical qualities of a successful NFL player:

DeMarco Murray, Oklahoma
Jordan Todman, Connecticutt
Derrick Locke, Kentucky
Roy Helu Jr., Nebraska
Mario Fannin, Auburn
Kendall Hunter, Oklahoma State
Da'Rel Scott, Maryland
Brandon Saine, Ohio State

   Now some names stick out as players who shouldn't be on the list, like Saine and Fannin- well that's completely true. That's where the college production factor is important in eliminating poor prospects and that's what we'll be looking at tomorrow.

   I bolded the names of the players the Patriots have been looking at, which is a great sign.

   So now we have the following knowledge:

The 40 Yard Dash is actually a positive indicator for a running back's NFL success
The Broad Jump is another solid indicator
Three Cone is important if the running back is a shifty back (ie: Ray Rice)

   Knowing these three facts, we were able to narrow down our list of intriguing prospects to the above grouping. Keep in mind that these players meet the physical requirements of a successful NFL running back. They have the explosion and burst to get through the hole and past/through defenders [Broad Jump] and the speed to run away in the open field [40 Yard Dash]. However, players need more than just the physical qualities to be a successful running back. As a result, we must look to their college production.

   We understand that all players come from different teams with different quality offensive lines in front of them and different talent surrounding the offense. We also understand that talent is talent and talent will prevail. There is not a top running back in the league, who premiered since 2005, who did not have a 1000 yard from scrimmage season in college. That's a solid bench mark to think about- a player must have a 1000 yard from scrimmage season in college. Think about it- if they can't do it in college against weaker defenses, why would they do it in the NFL? Now we still expect players to improve, but if they can't reach that mark in college, they have a very low chance of doing it in the NFL.

   Now the question is: does it matter when the 1000 yard season takes place? After examining the top players in the NFL, we come to find that it's not imperative that the 1000 yard season take place in their final year. In fact, most of the late round steals are players who reached the bench mark the year prior to their final season and saw a drop in production in their final year. Of course, it is preferred that a player continue to play at a high level until they graduate, but if an injury or some other factor outside their control slows them for part of a season, there's no need to worry- in fact, the player might be hiding some extra value as they drop in the draft.

   There are a couple types of players:

Ryan Grant - This player has great success early in their career, but for some reason, their production drops for their last seasons at college. They showed the talent early on, but their lack of production drops their value. These players can be found in the later rounds and in free agency.
Compares to: Ryan Williams, Da'Rel Scott

Chris Johnson - This player plods along as the back-up early in their career, but when they're the starter they explode for one tremendous season before leaving for the draft. These players are usually found in the middle rounds of the draft.
Compares to: Mikel Leshoure

Adrian Peterson - This player is a star their entire college career. They put up big numbers each year and it seems that defenses can't stop them. These players are usually found in the beginning rounds of the draft.
Compares to: No one in this draft, but sort of Jordan Todman, Kendall Hunter

   And when I say "compares to", I don't mean anything more than their college production is similar. Is there a best kind of player to draft? All three are successful. It's just a matter of how much risk you're willing to take on with your prospect. The Peterson type players come with low risk of bust at a high cost, while the Grant players have high risk of bust, but at a low cost. It's a matter of preference- I believe that the Patriots are willing to try the low cost players because that's how they've had success in the past couple seasons.

   No matter the category, however, a few statistical facts hold true:

A player should have at least one 1000 yard season while in college.
A player should accomplish that 1000 yard season at a 5.0+ yards/carry pace.

   Now there are some exceptions to the rule, but if the prospect meets these two qualities, they have an astronomically higher chance of success in the NFL.

   Since we've looked at all of the factors, we can come to the following rules:

A player must have at least one 1000+ yard season while in college.
A player should accomplish the 1000+ yard season at a 5.0+ yards/carry pace.
A player should run a sub 4.55s 40 Yard Dash if they're an above average player.
A player should run a sub 4.5s 40 Yard Dash if they're an elite player.
A player should jump at least a 9-8 Broad Jump to succeed.

   If a prospect meets these requirements, they have a very high chance of success. Here's an explanation of each measurement:

The 1000+ yard season means that the player is able to reach 1000+ yards and could be able to reach that benchmark in the NFL. If the player cannot reach the mark in college, they most likely will not in the NFL.
The 5.0+ yards/carry means that the player was reaching the 1000+ yards on their abilities and not because the coach called 300 running plays for under 3.5 yards a play. It shows that the player can make plays.
The sub 4.5s 40 Yard Dash means that the player can break free in the open field, as well as turn the corner. A player can succeed without the speed, but they won't be an elite player.
The 9-8+ Broad Jump means that the player has the explosion and burst to hit the hole and get to the second level, where they can use their 40 Yard Dash speed.

   I based my numbers of the CBS Draft website which, unfortunately has some incorrect numbers for this year's prospects. They're not completely up to date with Pro Day numbers and Mark Ingram's broad jump was nearly half a foot short of what he actually performed. Please let me know if the numbers are incorrect, or if you know a prospect who meets these qualifications!

   Players who meet the physical requirements of an NFL running back (not fully complete due to on-going Pro Days):

DeMarco Murray, Oklahoma, 4.37s, 10-10.
Jordan Todman, Connecticut, 4.40s, 10-6.
Derrick Locke, Kentucky, 4.37s, 9-10.
Roy Helu Jr., Nebraska, 4.40s, 9-11.
Ryan Williams, Virginia Tech, 4.45s, 10-3.
Kendall Hunter, Oklahoma State, 4.46s, 10-2.
Da'Rel Scott, Maryland, 4.40s, 9-9.
Jamie Harper, Clemson, 4.53s, 10-0.
Johnny White, North Carolina, 4.50s, 9-9.
Delone Carter, Syracuse, 4.54s, 10-0.
Mark Ingram, Alabama, 4.53s, 9-10.

   This is a pretty solid list of top prospects, mid-round hopefuls and late-round dreamers. Let's take a look at which players meet the production requirements: The players are put into production tiers:
Ryan Grant: Early performance, followed by lack of production.
Chris Johnson: No production, followed by a season of tremendous production.
Adrian Peterson: Consistent production throughout college.

DeMarco Murray, Oklahoma, 4.37s, 10-10. Ryan Grant/Adrian Peterson Production (mid-career injury, successful post-injury season).
Jordan Todman, Connecticut, 4.40s, 10-6. Adrian Peterson Production.
Roy Helu Jr., Nebraska, 4.40s, 9-11. Adrian Peterson Production.
Ryan Williams, Virginia Tech, 4.45s, 10-3. Ryan Grant Production.
Kendall Hunter, Oklahoma State, 4.46s, 10-2. Ryan Grant/Adrian Peterson Production (mid-career injury, successful post-injury season).
Da'Rel Scott, Maryland, 4.40s, 9-9. Ryan Grant Production.
Delone Carter, Syracuse, 4.54s, 10-0. Adrian Peterson Production.
Mark Ingram, Alabama, 4.53s, 9-10. Ryan Grant Production.
Derrick Locke, Kentucky, 4.37s, 9-10. Barely meets the production requirement as a Danny Woodhead-esque type player in the NFL. Not a feature back. Consistent all through college.
Johnny White, North Carolina, 4.50s, 9-9. Barely meets the production requirement as a Danny Woodhead-esque type player in the NFL. Not a feature back. One solid final season.

   Not surprisingly, most of the players who met the physical test matched it with on-field success. I will eliminate both Locke and White from the Patriots' conversation because they're both not every down backs. That leaves the following players as potential draft prospects:

First Round
Mark Ingram, Alabama, 4.53s, 9-10. Ryan Grant Production.
Second Round
Ryan Williams, Virginia Tech, 4.45s, 10-3. Ryan Grant Production.
Jordan Todman, Connecticut, 4.40s, 10-6. Adrian Peterson Production.
Kendall Hunter, Oklahoma State, 4.46s, 10-2. Ryan Grant/Adrian Peterson Production (mid-career injury, successful post-injury season).
DeMarco Murray, Oklahoma, 4.37s, 10-10. Ryan Grant/Adrian Peterson Production (mid-career injury, successful post-injury season).
Mid Rounds
Roy Helu Jr., Nebraska, 4.40s, 9-11. Adrian Peterson Production.
Delone Carter, Syracuse, 4.54s, 10-0. Adrian Peterson Production.
Late Rounds
Da'Rel Scott, Maryland, 4.40s, 9-9. Ryan Grant Production.

   Since we're looking at the Patriots, we're looking for the high value prospect. Which player will exceed their draft value? Which player will be considered a miss? Let's consider the following chart (that I just made up). On a scale of 1-10, you want the following breakdown:

Top Half 1st Round
Minimum Acceptable: 8.5/10
Bottom Half 1st Round
Minimum Acceptable: 7.5/10
Top Half 2nd Round
Minimum Acceptable: 7/10
Bottom Half 2nd Round
Minimum Acceptable: 6.5/10
3rd Round
Minimum Acceptable: 6/10
4th Round->UDFA
Minimum Acceptable: 5.5/10

   I believe that's a pretty fair evaluation of expected return.

   Running through the regressions, we're able to come across formula. This roughly means is that the faster they run their 40 and the farther they jump their broad, the greater their expected physical value in the NFL. As a result, I predict the following values with the aforementioned running backs:

DeMarco Murray - 7.24
Jordan Todman - 6.78
Roy Helu Jr. - 6.29
Ryan Williams - 6.29
Kendall Hunter - 6.16
De'Rel Scott - 6.14
Delone Carter - 5.56
Mark Ingram - 5.48

   Keep in mind that the standard error with these predictions is 1.88. The expected values from these predictions are a range of values because they represent the possible outcomes of a player with a set physical nature. It's up to the analysis of their college statistics to see whether the prospect deserves to be in the upper portion of the prediction, or the lower part.

   For example:

Delone Carter, Syracuse - 3.68 -> 7.44 (one standard deviation range, represents nearly 70% of the possible outcomes)

    After examining Carter's college production, it seems as if he's still in the process of improving. As a result, he belongs in the upper portion of expected values. I feel comfortable giving him a grade of 6.3 in the NFL; that means that Carter will be a reliable committee back and possible a solid tandem back, but he won't be an elite feature back.

   Here are my scores following the college production adjustment:

Jordan Todman - 7.8
DeMarco Murray - 7.5
Mark Ingram - 7.4
Kendall Hunter - 7.2
Roy Helu Jr. - 7.1
Ryan Williams - 6.6
Delone Carter - 6.3
De'Rel Scott - 6.1

    Now I'm sure these numbers are debatable, and I certainly haven't refined them (they're a quick evaluation on my part), but I would say that these numbers (in my mind) are a fair representation of expected value from these running backs after combining their physical abilities with their actual production. I believe that Todman has the most upside in the draft and that Williams is overrated with his first round grade. I think it will be interesting to compare these scores with the physical scores at the end of a couple seasons, in order to see if production adjustments are needed to make a more accurate score.

   This represents where the players deserve to be drafted:

Early First Round - None
Late First Round - Jordan Todman
Early Second Round - DeMarco Murray, Mark Ingram
Middle Second Round - Kendall Hunter, Roy Helu Jr.
Late Second Round - Ryan Williams
Third Round - Delone Carter, Da'Rel Scott

   And we can see a few players who are overrated (Ingram, Williams), a few who are underrated (Todman, Helu, Scott), and a few who are fairly rated (Murray, Hunter, Carter). As a result, the Patriots should remove Ingram and Williams from their list of potential running backs because they would have to pay more than their expected return in order to draft them. The question remains: which players should the Patriots draft?

   The Patriots could get a first round talent in Jordan Todman in the second round, and they could get a second round talent like Roy Helu Jr. in the fourth or fifth round. They could even go for the third round talent in Da'Rel Scott in the sixth or seventh round. Why pay a second for an RB like Hunter when an equivalent RB like Helu is available in the later rounds? There are clear differences since Hunter can catch the ball and is a solid blocker (which are factors not evaluated in this piece that focuses purely on the player's ability to run the ball), but from a running perspective, they're expected to have similar outputs.

   If I were the Patriots, I would take Jordan Todman in the middle of the second round (trade down or up, just grab him) and then take Da'Rel Scott in the sixth round. This provides two running backs with potential to the stable and it gives the Patriots the opportunity to have a feature back, or at least an elite tandem back, in Todman.