Are Running Backs At Higher Risk of Injury Compared to Wide Receivers?

Welcome to the inaugural article of the Dynasty Ghost Map! I figured the best way to introduce myself would be to tell a horrible joke….. Ready?

“Last season, I won a dynasty championship with Bilal Powell and Tim Hightower as my starting running backs......”

For most you, this is far more frustrating than comical. However, this statement provides a perfect depiction for the growing frustrations around the running back position.

At the beginning of the 2015 season, running backs like Le'veon Bell and Mark Ingram had firmly seated themselves within the top 100 of startup ADP. Like many of you, I also drafted running backs early in hopes of them carrying me towards fantasy football glory. However, the story seems to always end with these stud backs being locked in an IR spot or providing minuscule production as they play through an injury.

In the current climate of dynasty fantasy football, wide receivers have largely monopolized the top 100 of startup drafts. This is largely due to both the growing frustrations around running back injuries and the overall devaluation of the position by the NFL. This frustration has led to many analysts to advocate for the early selection of wide receivers over their running back counterparts because they are deemed to be “safer” and “have more years of production”.

There are many great writers in the community that have done fantastic work to outline their conclusion that running backs are riskier, but I have yet to find one who has used the proper study design necessary for determining the numerical amount of risk for running back injury compared to wide receivers and how much risk can be eliminated by selecting a wide receiver instead of a running back in the top 100. The purpose of this study is to answer both questions and to determine if the precieved risk in running backs is real or if it is a product of out recall bias due to previous heartbreak.

Methodology

The first thing I did to approach this problem was collecting the names of all the Running Backs and Wide Receivers (from 2009-2015) in the top 100 of ADP. These 800 players served as the source population from which my study group will be drawn from. Next, I seperated the positions and assigned each player a study ID. I then completely blocked myself from seeing who represented each ID, because I did not want to be tempted into selecting players I knew got injured into my next step. ** Note: I know it sounds silly, but anyone who is looking to support their hypothesis will be tempted to over select the injured running backs. If a researcher is not blinding themselves in their selection, it should be assumed that their selection has been biased.**

Then, I selected 480 players from the source populations to represent my study population by using a number generator. In all honest, I could have used every player in my original player cohort, but I'm a one man analytics team and this would have increased my work significantly. ** Note: 60% of my source population was calculated through a required sample size equation that sought to maintain a 90% power.**

I then unblinded myself to those who had been randomly selected into my sample. For each player within my study population, I noted every game missed due to injury. If a player had started a game but did not complete the game due to injury, the game was counted as a missed game given that they did not play the following week. If a player missed four or more games, he was denoted as missing a significant amount of the season. (1/4 of the season is a larger portion of our shortened fantasy season.) This information was then analyzed using SAS (a statistical software commonly used by Epidemiologists and Statisticians). Below is a 2x2 table depicting the number of players that fell into each category and the output generated by the data collected.

Risk in Running Backs = 0.3833

*Risk = proportion of running backs who are injured in the running back population*

* RB significantly injured divided by total RBs in the sample. *(92/240)

Risk in Wide Receivers = 0.2917

*WRs significantly injured divided by total WRs in the sample. (70/240)

Relative Risk = 1.314

** The most important stat, because it denotes the amount of increased risk **

* Incidence of RB injury divided by Incidence of WR injury.

(92/240) / (70/240)*

Attributive Risk Percent = 43.5%

** Denotes the amount of risk that can be elimated by selecting a wide reciever instead of a running back**

*(92/240) - (70/240) / (92/240)*

Conclusions of the study

1. Early round running backs had about a 32% increased risk of missing significant playing time per season (4 games missed) compared to wide receivers.

2. By selecting a wide receiver, you eliminate about 44% of the risk of injury in early round draft pick per season.

Most other studies are concerned with the average amount of running back injuries per season, however these numbers are indepencent each season. The figures that I have shown above are both statistically significant and account for the independency of each season. By using the retrospective cohort study design, we are able to derive actual risk figures and association measures.

Based on the results demonstrated above, it appears that running backs are even riskier than we might have imagined. These numbers suggest that dynasty running backs have the same per season risk of injury (32%) as skin cancer in people who use the tanning bed daily. For example, let’s say that you drafted three running backs in the first eight rounds of your startup draft. My study suggests that it is probable that at least one of those backs will miss significant time each fantasy season. The more you draft early, the more risk you take on each season. To add to the argument, we would eliminate nearly 44% of in injuries in our top 8 picks by selecting wide receivers instead of running backs. I am not advocating for the dismissal of drafting talented running backs. However, it is clear that we must consider the amount of per season volatility we introduce to our team by the use of high capital draft picks. We need our studs to stay healthy!

Suggestions

Start up:

I would recommend drafting as few running back in the first 8 rounds as possible. It sounds ludicrous to some, but this study provides further support to the "Zero Running Back" strategy (Waiting until later rounds in a draft to select running backs). Additionally, the elimination of excess seasonal injury risk is vital to putting together a championship roster. By placing your bets on a wide receiver, you increase your chances of bringing home the trophy significantly.

Rookie Draft:

Obviously, this becomes the tricky aspect of my study. My study is not aimed at the rookie draft, therefore, I cannot say confidently that you are more or less safe in choosing Zeke at 1.01. I can state that his risk of injury each season is the same as the running backs in the startup cohort. However, there are instances when position need outweighs the risk of injury. If you are able to infuse the risk of a highly drafted running back with a stable core of wide receivers, then draft him. The one recommendation I would make in light of injury risk would be to use as little capital on running back as possible. Jeremy Langford, Karlos Williams, Matt Jones, Buck Allen, and Thomas Rawls were all selected after the second round in rookie drafts last year. All of these backs going forward will provide you both serviceable production for the coming season and were had reasonable prices.

**Notes**

1. The assumption that a player is “injury prone” was accounted for by the randomization of selection.

2. Each season was deemed to be completely independent and not affected by an injury from the previous year. (further testing could be done to determine the validity of this assumption)

3. A player was able to be represented more than once and as either a case of a control given they were within the top 100 of ADP. (sounds strange to many, but it is a common practice in Epidemiology)