Yards per carry is a trap for evaluating running backs
Yards per carry is the most-cited running back stat in football and one of the most consistently misleading. A 5.0 average reads as obviously good. A 3.4 reads as obviously bad. Both numbers tell you almost nothing about the back unless you know what plays produced them, against which defenses, and behind which offensive line. The stat looks neutral. It is actually a confounded mess that has been driving bad personnel decisions for two decades and propping up a contract market with a clear systematic error in it.
The single-run problem
The most basic statistical problem with yards per carry is that the distribution is wildly skewed. A back with twenty carries in a game might have nineteen carries for between zero and seven yards, and one carry for 60 yards because a safety missed a tackle. That single broken run can move the back's YPC for the game from 3.4 to 6.5, and the back will be described in the postgame coverage as having "torched the defense" or "shown explosiveness" or "given the offense a spark." None of those phrases capture what actually happened, which was nineteen ordinary runs and one accidental long one.
This is not a hypothetical. Across an NFL season, roughly 10% of running back carries account for roughly 50% of total rushing yards. The long-run tail is so heavy that a stat averaging across all carries is essentially a coin flip on whether the back had one of those tail events that week. Running back YPC, especially in a single game or even a single month, is closer to a sampling lottery than a measurement of the back's quality. The averaging mathematics that work fine for shooting percentages or batting averages do not work for an outcome distribution this skewed.
The blocking confound
The most important variable in determining how many yards a running back gains on a given carry is whether the defender at the second level got blocked. Backs almost never break tackles for the first three or four yards of a run — they accept whatever the line and the play design give them. The difference between a 3-yard run and an 11-yard run is almost entirely determined by what happens in front of the back, not by what the back does with the ball. YPC is therefore mostly a measure of offensive line quality with a small running back contribution layered on top.
This is why veteran scouts have, for a long time, distrusted YPC for evaluating backs and preferred yards after contact, which strips out the blocking part of the equation. Yards after contact is also not a perfect stat — it depends on which defender qualifies as the "tackler" and where the contact happens — but it captures something more like the back's individual contribution. Backs who lead the league in YPC and backs who lead the league in yards after contact are usually different players, and the second list is more predictive of which back will look good in a different offense.
Game script and defense-faced effects
A back who plays for a team that gets ahead a lot will carry the ball against tired, conservative, eight-in-the-box defenses that are conceding the run to prevent the pass. A back who plays for a team that falls behind will get most of his carries against fresh defenses that know the offense has to throw and can crowd the line as part of pass-rush packages. The first back will post higher YPC numbers basically as a side effect of being on a good team. The second back will post lower YPC numbers basically as a side effect of being on a bad one.
The size of this effect is large. Studies of league-wide carry data have found that game script alone — whether the team was ahead, behind, or tied — accounts for roughly half a yard per carry difference between otherwise identical backs. A back going from a 4.0 YPC offense to a 4.5 YPC offense isn't necessarily a better player. He might be the same player who got drafted by the team that wins more. The stat that's supposed to evaluate him is instead evaluating the situation he plays in.
The contract-market trap
The cleanest illustration of the trap is how running back contracts get debated. A back is judged "elite" partly on YPC, gets a top-of- market deal, leaves for a new team, posts a worse YPC, and is declared washed. The vast majority of the time, the explanation is that the new team has a worse offensive line, a worse passing game, or a worse game script profile, all of which were doing most of the statistical work in the previous team's YPC number. The back hasn't gotten worse. The stat was always describing the team more than the player.
Front offices have started catching on. The market for running backs has compressed substantially because the smart-money teams figured out that paying for a YPC-based reputation is paying for blocking and play design rather than for the back himself. The teams that still pay top dollar for backs are mostly buying a label that the stat has constructed for them, and the rate at which those contracts end up looking bad is higher than for any other position in football.
The stats that actually predict
Three running back stats have stabilized as more predictive than YPC across multiple research lines. Success rate, which measures the percentage of carries that gained enough yardage to keep the offense ahead of the chains, captures something more like consistent gain quality and is far less skewed by long runs. Rushing DVOA, which adjusts for opponent defense and game situation, captures what YPC pretends to measure but actually doesn't. And yards after contact per carry strips out the blocking and isolates what the back contributed.
If you want a single number that tells you whether a back is good, success rate is the closest available substitute for YPC and is much harder to fool. A back with a 50% success rate is genuinely producing for the offense on most carries. A back with a 38% success rate is bleeding the offense even if his YPC looks fine because of one or two long runs. Pairing success rate with yards after contact gives you a much closer read on what kind of back you're actually evaluating.
How to read a stat line
None of this means YPC should be ignored. It means it should be read with context. A high YPC paired with a low success rate is a back having a few long runs in an otherwise unproductive game. A low YPC paired with a high success rate is a back grinding out consistent gains behind a line that doesn't open big holes. The two stats together tell you the kind of back you're watching. Either one alone tells you almost nothing.
This matters because YPC is the stat that contract negotiations and Pro Bowl ballots and trade evaluations still anchor on. The market for running backs has, for the last decade, been quietly distorted by the fact that the headline number used to evaluate them is one of the most context-dependent stats in football. Teams that have figured this out have started buying running backs the way other people sell them, and they keep being right. The stat survives anyway, because it's simple, intuitive, and on the broadcast graphic every Sunday. The teams winning with running backs have stopped trusting it.