A week or so back, while I was perusing Twitter, a couple of Tweets caught my attention. The first tweet that caught my attention was about selling running backs after their fourth season. Without an ounce of research, I initially agreed with this premise. It seems intuitive that running backs, who generally have shorter careers than other position players, should be sold after their four seasons.
This is why you sell RBs after their 4th season. Gurley went from the 1.01 in dynasty to virtually being tied with Kenyan Drake in value. In one league I traded Gurley for Chubb, Sutton, and a 2020 2nd. Also look at Gordon and DJ. Now it’s time to sell Zeke and Henry. https://t.co/3nXkqnDokp
— Austin DeWitt (@DeWitt_Dynasty) January 16, 2020
A couple of days later, I saw the Twitter thread below. This tweet caused me to re-think my initial stance on when to sell off running backs. This tweet thread also inspired me to look further into running back production based on age.
This is an interesting and plausible-sounding theory that I'd never heard before (sell all dynasty RBs entering year 5). Is it a good idea? I have no idea!
But you know what I like to do when confronted with a plausible-sounding hypothesis. I test it against the data. https://t.co/5GprGJV0s2
— Adam Harstad (@AdamHarstad) January 18, 2020
I gathered the top 24 running backs from 2012 through the 2019 season. I used data from weeks 1-13 of the NFL season, which encompasses the fantasy regular season. And I only included players who played at least 10 games. Running back finishes were based on a per-game average and on PPR scoring, which was downloaded from FantasyPros while player ages were gathered from Pro-Football-Reference.
A quick explanation of why I used per game averages, and a 10-game threshold. Overall scoring rewards players who play every game, but doesn’t necessarily reflect that player’s true value. Per game scoring gives a more accurate view of how valuable a player is on a week to week basis. Without setting a threshold of games played, per-game scoring can be too easily skewed by low or high scoring games. Setting the baseline of 10 games played was done to reduce the impact of that variance.
With that brief explanation behind us, let’s move on.
Below is the full data set over an eight-season period.
The age range for all top 24 running backs during this eight-year period was 21 – 34 years of age. I created a simple sheet listing the ages from 21-34 and recorded how many times a running back that age finished as either a top-six, 12, 18, or top 24 running back. This provides a nice snap-shot of the production provided by the specific ages.
To provide another visualization of what the above spreadsheet represents, I’ve added the chart below (mostly because I think it’s pretty).
Fun With Averages
Now let’s look at the averages ages for each RB finish (RB1 – RB24) over the past eight seasons, along with the average age for the top 24 running backs by season. Per game average scoring finish, from 1-24, is listed in the first row. The top column lists the season, and below that, the age of the running back when they finished as an RB1, RB2, RB3 through RB24 are listed. I also added average ages for all top 24 running backs for each particular season and the average age for each running back ranking.
Over the past eight seasons, the average age of the RB1 has been 25.13 years old. It should be noted the sample is skewed by the 2012-2014 seasons when the RB1 was between 27 and 29 years old. Since that time, the oldest an RB1 has been was 25 years old in the 2016 season. We can also see that the average age for the top 24 running backs has dropped from 26 to 24.79 over the past eight seasons.
The difference in production from the RB1 and RB24 in any given season can be anywhere from 10 to 16 points per game, so to refine this data let’s look at the averages for RB1-RB6, RB7-RB12, RB13-RB18, and RB19-24.
Top Six Running Backs
From 2012 through 2014, the RB1 was ancient when compared to the past five seasons. The 2012 and 2013 seasons saw 27-year-old running backs (Adrian Peterson and Jamaal Charles) lead the position in per-game scoring, and the relatively ancient 29-year-old Matt Forte was the leading scorer in 2014. In the five subsequent seasons, the leading scorer has been 23 years-old on three occasions, with a 24-year-old running back leading per-game scoring once and a 24-year-old leading in per-game scoring in another season.
Looking at the ages of the RB2, RB3, RB4, RB5, and RB6 from 2012-2015 seasons vs. the 2016-2019 seasons, there appeared to be a clear demarcation between the two time periods. So I diced up the dataset one more time.
The above chart confirms what I thought I was seeing. Excluding the RB5, the average age of the top six scoring running backs has decreased over the past four seasons. The average age of the RB1 has fallen 2.75 years over the past four seasons compared to the four seasons before the 2016 season. This is a consistent trend with the RB2 – RB6 tiers as well.
Only four times in the past four seasons has a top-six running back been 26 or older, with three of those instances occurring in 2016. Over the past two seasons, every top-six running back has been aged 25 or younger.
Looking at the next six running back finishes finds that the average age for every scoring finish (RB7, RB8, etc.) has also decreased. The most significant drop coming at RB12. The RB12’s average age was 27.75 years old from 2012-2015 and dropped to 23.5 years old the following four seasons. The past four seasons have found a tad more success for this tier of running backs, with eight over the age of 26 finishing as an RB12 or better. There were even a couple of 30-year-old running backs sprinkled in this range.
The same trend largely continues from RB13-RB24. The running backs finishing in these ranges continue to drop in age for the most part.
Below is one more sheet that looks at just the age differential from the first four year period (2012-2015) to the next (2016-2019). I’ve italicized the age differential in the instances that the average age of the running backs increased or had no change. The average age for the RB5, RB17, and RB24 increased from one four year period to the following four year period, while there was no change for the RB16.
- The past eight years does show that you can find production from running backs ranging in age from 21 to 34. The number of running backs still providing the top 24 seasons does begin to crater from age 29. The sweet spot for a running back to provide top-end production is 22 – 28. When I started this process, I expected running backs aged 22 – 26 to lead the way. They did so with 60% of all top 24 seasons. But running backs in the 27-29 age range provided 26% of the top 24 seasons during this period. I’ve always thought of 26 as the cutoff for running back production in my mind. Except for certain outliers, I usually try to sell running backs entering their age 26 seasons. Based on this data set I may have been selling too early.
- When you break up the eight-year sample into distinct eras (2012-2015 and 2016-2019), the average age of the top 24 running backs is decreasing at nearly every tier. But in 12 cases the difference in age is just a year or less.
Data is Needed
- There is no disputing that running back production is coming from younger players. No matter how you look at it, that is evident. If you look at just RB1’s or the top six running backs, etc.; top-24 production is coming from younger players. But that may be caused by other factors. It’s possible that running backs have started entering the NFL at a younger age. That would cause the player pool itself to be younger. If the player pool is younger, it should follow that the most productive running backs would be younger. The next iteration of this research will include how many seasons a running back has played in the NFL at the time of their top-24 season.
- Running back production should also account for total touches. One of the more popular narratives when determining to sell running backs has to do with the number of touches they’ve accrued in their career. Is it possible that we’ve been looking at age incorrectly? The older a running back is the more touches they will have had. Could total touches provide a better cut-off than age? Adding this data will provide another data point to help determine exactly when we should move on from running backs.
- One thing caught my eye while I was compiling this data. If you look at the very first sheet at the top of this page, it seems that a lot of the “older” running backs producing top-24 seasons were of the pass-catching variety. Do pass-catching backs age better? One of the parameters I’ll add to the data set is going to be receptions.
Looking at the age of running backs has been highly instructive, and has also been informative. But more work needs to be done. Yes, top-24 running back production is coming from younger players. But are there other factors that contribute to this trend? I’ll continue to dig deeper into this and provide you with more insights in my next article in this series.
Thank you for reading. If you have any questions or comments, please just let me know. If there is a data point you’d like added let me know that as well. You can find me on Twitter @DFF_Shane. If you’re a Factory member go ahead and reach out to me in our Slack chat as well. If you’re not a Factory Sports member, become one today!