Since releasing Breakaway Rush Score (BRS) in 2020, it’s been one of our favorite metrics to use for evaluating running back prospects. Not only does it have one of the stronger relationships with fantasy points earned by a player in their first three seasons of all the metrics we look at, but it also helps us to confirm if a player truly is “explosive” or “seems to have lost some of their big play ability post-injury.”

I’d recommend reading the article linked above, but for a brief summary, BRS counts the percentage of a player’s rushing attempts that went for 15, 20, and 40 yards or more, blends these percentages together, and then multiplies by a normalized rushing attempt number to arrive at a score for each prospect. (The normalized rushing number simply takes a player’s rushing attempts per game and multiplies by 48 to set all players on a level playing field. This allows the metric to capture efficiency while also incorporating volume in a way that isn’t skewed by seasons or games played.)

While BRS provides a useful context for better understanding a prospect as a rusher, it doesn’t account for receiving ability — a major component of the modern-day running back. When working on the original article, I quickly explored the relationship between breakaway receptions and fantasy scoring. At the time, I had much less data available than I do now, so I wasn’t able to firmly determine if it should be included in the metric.

With a lot more play-by-play data now at my disposal, I revisited including breakaway receptions into the metric and confirmed that BRS is improved by capturing an element of a player’s receiving production. (I recently made a slight change to BRS given the availability of expanded play-by-play data. This updated the scores presented in the 2023 version of the article, but didn’t change the majority of player’s relative positions when compared to each other.)

Breakaway Score

The relationship between breakaway receptions and fantasy scoring is pretty weak. In a sample of 211 RBs drafted between 2014 and 2021, the count of a player’s career receptions that went for 10 or more yards explained somewhere between 3 and 5% of the variance in PPG. Of course, in this context, explaining more than 20% of the variance would be impressive. After all, draft position only explained approximately 30% of the variance in PPG in the data used in this article, and career rushing yards, for example, explained less than  13% of the variance in PPG.

However, there is an important caveat when discussing R² or correlations among prospect metrics — correlations between prospect metrics and fantasy scoring tend to change in material ways depending on the players included in the sample they are derived from. For this reason, many of the models one might create while attempting to maximize R² tend to overfit. (A model is overfit when it is built too heavily on the specifics of the data it is being fed. An overfit model might look like it does a really good job of predicting something, but when transferred to data that wasn’t included in the sample it does a very poor job with prediction.) As a result, recognize that the numbers presented in this article could fall outside of these ranges if calculated using a different sample.