Three years ago, I wrote this post titled “Take Away His X Best Carries and He’s Average.” I did the same thing last year, too. The idea was simple: Suppose you sort each running back’s carries in descending order by yards gained. How many carries would we need to take away from him to drop his production to at or below average?
There were 47 running backs last season who had at least 100 carries in 2017, those players averaged 4.08 yards per carry last season. The best running back by yards per carry last season, by far, was Alvin Kamara, with a remarkable 6.07 average gain. Kamara had just 120 carries last year, but he was remarkable at producing big runs. He had 13 carries of 15+ yards last season, an unfathomable (and unsustainable) rate of one 15+ yard run every 9.2 carries. Kamara’s lofty YPC number wasn’t the result of one big run, which is often the case for a player with a high YPC average on a low number of runs: sure, Kamara had a 74-yard run, but no other run went for more than 25 yards.
In fact, if you removed Kamara’s 12 best runs, he would have still averaged 4.083 YPC, a hair above the 4.078 average among all running backs with 100+ carries. So in order to bring Kamara’s YPC below-league average, you need to remove his 13 best runs of the season.
The next best player by this metric was Patriots RB Dion Lewis. You might be surprised that Lewis — who had a 4.98 YPC on 180 carries — would be higher than Chiefs RB Kareem Hunt, who averaged 4.88 YPC on 272 carries. All else being equal, this metric rewards players with higher averages on a larger number of carries. But that’s what made Lewis’s season so impressive: he averaged 4.98 YPC despite his longest run being just 44 yards, and his second-longest topping out at 31 yards. Hunt had runs of 69, 58, and 53 yards. Take away Lewis’s best 5 runs, and his YPC drops to 4.26; take away Hunt’s best five, and he drops below average to 4.04.
The table below shows for each RB how many of their best carries you need to take away to bring their 2017 YPC average below 4.08.
| Player | Team | Rush | Yards | YPD | # |
|---|---|---|---|---|---|
| Alvin Kamara | NOR | 120 | 728 | 6.07 | 13 |
| Dion Lewis | NWE | 180 | 896 | 4.98 | 7 |
| Kareem Hunt | KAN | 272 | 1328 | 4.88 | 5 |
| Todd Gurley | RAM | 279 | 1303 | 4.67 | 5 |
| Mark Ingram | NOR | 230 | 1124 | 4.89 | 4 |
| Alex Collins | BAL | 212 | 973 | 4.59 | 4 |
| Kenyan Drake | MIA | 133 | 648 | 4.87 | 3 |
| Alfred Morris | DAL | 116 | 549 | 4.73 | 2 |
| Matt Breida | SFO | 105 | 465 | 4.43 | 2 |
| Devonta Freeman | ATL | 196 | 863 | 4.40 | 2 |
| Giovani Bernard | CIN | 106 | 459 | 4.33 | 2 |
| Wayne Gallman | NYG | 111 | 476 | 4.29 | 2 |
| LeGarrette Blount | PHI | 173 | 765 | 4.42 | 1 |
| Orleans Darkwa | NYG | 171 | 751 | 4.39 | 1 |
| Bilal Powell | NYJ | 178 | 772 | 4.34 | 1 |
| Marshawn Lynch | RAI | 207 | 890 | 4.30 | 1 |
| Derrick Henry | TEN | 176 | 744 | 4.23 | 1 |
| Jay Ajayi | 2TM | 208 | 872 | 4.19 | 1 |
| Isaiah Crowell | CLE | 206 | 853 | 4.14 | 1 |
| C.J. Anderson | DEN | 245 | 1007 | 4.11 | 1 |
I’m no fan of yards per carry as a metric, and one reason why is that it is heavily skewed by big outliers. Kamara’s teammate, Mark Ingram, is a good example. He had a full starter’s workload with 230 carries, and averaged a very impressive 4.89 YPC. But without his four best runs, his YPC would have been just 4.05. If less than 2% of a player’s volume swings him from below average to All-Pro caliber, that’s a sign that the metric is heavily influenced by outliers, and it’s why YPC is not very sticky from year to year.