NFL Running Back Strategies

By Jake Fisher

Last NFL season, I looked at how single-back rushing systems (one primary running back) compared to multi-back rushing attacks. It turned out that multi-back teams ran the ball slightly better. 

In the 2008 NFL season, several teams had success with multi-back rushing systems. The Giants used the combination of Earth (Brandon Jacobs), Wind (Derrick Ward), and Fire (Ahmad Bradshaw). The Panthers used DeAngelo Williams and Jonathan Stewart. The Titans used Smash (LenDale White) and Dash (Chris Johnson). All three of these teams ranked in the top seven in yards per carry and each earned a first-round bye in the playoffs. For these squads, and others, splitting carries may have been a factor in their rushing achievements.

Some NFL analysts claim that having a system that uses multiple backs increases the explosiveness of a running game. I wanted to see if the stats backed this up.

I attempted to find the connection between explosiveness and rushing system in 2008 by analyzing every team game-by-game and seeing how the percentage of carries given to the primary running back (degree of single-back system) correlated to yards per carry for the team (running explosiveness).

From this analysis I could have gotten one of three possible outcomes.

1. Multi-back systems led to more rushing explosiveness in 2008.

2. The Giants, Panthers, and Titans were aberrations, and single-back systems were actually more explosive in 2008.

3. There was a weak correlation between the type of running system and the explosiveness of the ground attack in 2008. This would mean that the system employed might not have been a factor in a team’s yards per carry.

II. Data Collection

The two main variables utilized are rushing explosiveness and degree of single-back system. I used yards per carry to represent running explosiveness. I used the percentage of carries given to the No. 1 running back to represent the type of system.

To clarify, let’s say the Chargers rush the ball 30 times in a game for 120 yards. The yards per carry would be 4.0. Let’s assume LaDanian Tomlinson had the most carries for San Diego in that game with 20 rushes. The percentage of carries given to the primary back for that game would be 67%.

I used game-by-game statistics to be more exact. I didn’t use cumulative season statistics because the percentage of carries given to the primary back over the course of a year is not always representative of the team’s rushing system.

For instance, in 2008, the Colts appeared on paper to have more of a split-back system than they actually had. Over the course of the season, 42% of the carries went to the running back with the most carries overall (Joseph Addai). This percentage is much lower than the league average and makes Indianapolis look like it has a multi-back running strategy. In actuality, the Colts used much more of a single-back system. Addai was the starter, but when Addai got hurt, Rhodes stepped in as the No. 1 back. Upon Addai’s return, the two split carries, but only for a few games.

For the majority of the season, the Colts were a one-back team. But, at the end of the season, Addai had only three more carries than Rhodes, and the statistics make it seem like the system was perfectly split. When looking at the Colts game-by-game, you can see that they actually gave their primary running back the ball 71% of the time. This is higher than the league average.

In the NFL, teams gave their primary running back the ball an average of 64% of the time during a single game. I listed the statistics for every NFL team in 2008 below.

Team YPC Percent carry No. 1 back
Giants 5.02 0.55
Falcons 4.84 0.68
Panthers 4.81 0.57
Ravens 4.78 0.53
Vikings 4.75 0.71
Patriots 4.49 0.42
Titans 4.44 0.55
Redskins 4.38 0.72
Jets 4.36 0.70
Raiders 4.33 0.60
Dolphins 4.33 0.50
Broncos 4.30 0.52
Texans 4.27 0.64
Bills 4.24 0.64
Bucs 4.23 0.56
Chiefs 4.20 0.63
Packers 4.16 0.72
Jaguars 4.13 0.54
Seahawks 4.10 0.58
Chargers 4.07 0.72
Cowboys 4.03 0.72
Eagles 4.01 0.65
Steelers 4.00 0.66
Bears 3.97 0.74
Rams 3.95 0.74
Browns 3.92 0.68
49ers 3.85 0.70
Saints 3.78 0.56
Bengals 3.67 0.68
Lions 3.62 0.74
Colts 3.46 0.71
Cardinals 3.44 0.64

III. Results

After I compiled the data, I ran an OLS regression. There was a negative correlation between the percentage of carries given to the primary back (degree of single-back system) and running explosiveness (yards per carry). What this means is that the more a team split carries, the higher that team’s yards per carry was in 2008.

The  regression line turned out to be…

yards per carry= -.0192(percentage of carries to primary back) + 5.41.

A ten percent decrease in primary back carries, gave a team .192 more yards per carry. The trend is statistically significant with 95% confidence.

The correlation is -.42 and the R^2 is .176.

Out of the 16 teams with the highest yards per carry during the season, 12 were below the median for percent of carries given to the primary back.

It appears that multi-back teams did indeed have more rushing explosiveness in 2008.

About the author

harvardsports

View all posts

5 Comments

  • Jake, very interesting research. Now that you’re taking econometrics, it might be time to include some controls in that regression. Perhaps teams that use multiple backs do so because both of them are very talented…thus, we have reverse causality (having good RBs leads to using multi-backs, which leads to better yards per carry).

    Perhaps we need some type of instrument for split backs. Maybe something based on the coach’s history of using split-backs. This is a very interesting idea and something we should further investigate.

  • I think you really need to take into account down and distance when performing this sort of analysis. Generally, with a 2+ back system each RB is a specialist.

    I would expect Ahmad Bradshaw to have a higher YPC on 1st and 10 or 3rd and long than Jacobs due to the nature of the down. Similarly, I would expect Jacobs to have a higher YPC than Bradshaw on 3rd and 1.

    Gaining and extra .5 yards on each carry vs. a workhorse back really doesn’t measure explosiveness. I would be more curious to see if the 2+ back teams had a significantly greater number of 5+, 10+, 15+ yard carries. It would also be interesting, if they did have a significant increase longer carries, if this increase in was due to better performance later in the game (the lighter workload kept the 2+ system backs fresh).

    If 2+ back systems still prove to be superior then it would seem that Barry Sanders would be considered the most explosive RB in the history of the game. Not only did he have a high career YPC, consistently lead the league in 10+, 20+ yard runs, he also took the highest percentage of non-QB carries for his team in the history of the NFL. If I remember correctly, Sanders took about 80-85% of the Lions carries over the course of his career.

  • I’ll say this–the 2009 Saints have had better running from both Pierre Thomas and Mike Bell than they did last year with ANYBODY. I do seem to notice that, because PT & MB are similar style backs, there is a noticeable push on the pile in the second half, ESPECIALLY when that RB is the one who got less/no carries in the 1st half. (See Bills and Dolphins games for reference–ESP the Bills game.)

  • There’s a selection bias that is completely overlooked in this. Teams with a high percentage of RB carries by their #1 feature an elite running back, so naturally they would have a higher proportion of carries by that individual. Imagine if the team with the lowest carries by their #1 (New England – 42%) had Adrian Peterson – you still think they’d stick to a running back by committee?

    The only way to accurately measure this effect is if you compare the same team within the course of a season, and then see how they did on games in which a feature back monopolized most of the carries versus other games in which the feature back had a lower proportion of carries.

    • If that bias exists, then the conclusion of the study – that multi-back systems do better – actually becomes more powerful. It would mean that teams that don’t have elite backs actually do better on the ground than teams that do.

Leave a Reply

Your email address will not be published. Required fields are marked *