Most Predictable Running Back Stats (2021 Update)
In fantasy football, preseason opinions are often shaped by how players performed in the previous season. Sometimes, data points from the previous season can be predictive but there are other stats that may be misleading. This analysis will dig into common running back stats and explore which ones often carry over from one year to the next and how to use those that don't.
This article has been updated to reflect data through 2020.
Running Back Predictability Methodology
For this study, I considered all running backs over the last 10 years who saw 100 touches in a season and compared their numbers to the following season, running a simple correlation for nine common per-game running back stats. This resulted in a sample of 437 running backs. The decision to use per-game stats is based on the fact that total games played are extremely unpredictable for all positions, especially running backs—those with an ADP in the top 200 over the last 10 years saw a year-to-year correlation of -.02 for games played. With the huge variance in games played, year-long volume stats showed very little year-over-year correlation over the entire sample.
The table below lists the year-to-year correlations for those nine running back stats using the methodology described above:
|Half PPR Points/Game||0.65|
|Yards Per Touch||0.37|
|Yards Per Rush Attempt||0.18|
Running Backs Need the Football
A running back’s usage is the most reliable data when trying to project his current season based on the previous year's stats. Running back production is mostly volume-driven—the year-to-year correlation for fantasy scoring is moderately strong, despite the fact that touchdowns don't often carry over from one year to the next. Compared to other positions, running backs compile so many touches and accumulate so many yards that as long as they have decent scoring equity, their predictable volume leads to fairly predictable fantasy production.
Per-Play Efficiency Can Fool You
The unreliability of efficiency metrics (yards per touch and yards per carry) is perhaps the biggest data point to note.
Touchdowns are generally known to be the most volatile stat (and therefore the hardest to predict), but we should also exercise caution when using a back's previous-year per-touch data as a reference point for projections.
Consider Christian McCaffrey—this year's consensus 1.01—and his yearly yards per carry:
Christian McCaffrey Year-to-Year Yards Per Carry
|Year||Yards Per Carry|
|2020 (3 Games)||3.81|
McCaffrey's range of outcomes may seem relatively small, but assuming a 300-carry season, the difference between his best and worst yards per touch figure would result in a gap of roughly 400 yards—in fantasy terms, this could be the difference between four to five spots in end-of-season ranks and could have an even bigger impact in leagues that offer yardage bonuses. Even the most consistent fantasy scorers are susceptible to large swings in efficiency from season to season.
Efficiency metrics, as I've mentioned in my previous work, are perhaps best used to compare a player's end-of-season numbers to the league average, or to his individual average (if the sample is large enough). You can then decide to what extent regression to the mean should be expected.
The Bottom Line
When analyzing the previous season's running back stats, here's what to keep in mind:
- The most predictable year-to-year stats for running backs are their usage numbers (touches, rushing attempts, targets, receptions).
- A running back's yardage is much more likely to carry over from year to year than his touchdowns.
- A running back's previous-year efficiency is a poor predictor of his future efficiency—don't put much stock into yards per touch and yards per carry.