2023 Weighted Opportunity for RBs

There are a lot of bad, worthless, noisy, and overused stats in fantasy football. “Touches,” for running backs, might not be at the top of the list… but it’s close.

Why? Because not all touches are created equal. A running back can touch the ball as a runner (rushing attempt) or as a receiver (reception). Rushing attempts are far more common, but receptions are far more valuable. Over the past five seasons, rushing attempts have comprised 81% of all running back touches, though only 57% of all running back fantasy points scored have come on the ground. How can that be true? Because a reception is 3.2 times as valuable as a rushing attempt in PPR leagues.

Why is “touches” such a bad stat? Because I’ve come up with something much better.

Volume is far more important than efficiency for fantasy running backs, and far more important for running backs than for any other position in fantasy. But “raw touches” doesn't do a good job of measuring volume. “Raw touches” is inferior to “raw opportunities (carries + targets),” and “raw opportunities” is inferior to our stat: weighted opportunity.

Weighted opportunity is what it sounds like. It measures a running back’s opportunity, weighted appropriately for the worth of each unit of opportunity (a carry or a target).

On average (over the past five seasons), a single rushing attempt has been worth about 0.62 fantasy points. A target has been worth roughly 1.57 fantasy points in PPR leagues. So, broadly speaking, a target is worth 2.53 times as much as a carry in PPR leagues. In half-point PRR leagues, targets are 1.90 times as valuable. And even in standard leagues, targets are still 1.26 times as valuable.

The methodology to calculate this was simple. In each year, we totaled running back rushing fantasy points and then divided that number by total carries. For targets, we totaled all running back receiving fantasy points and then divided that number by total targets.

However, we can improve upon this further by incorporating red-zone usage. (For simplicity, we’re using PPR scoring in the next two charts.)

BROADLY SPEAKING, OUTSIDE OF THE RED ZONE, A TARGET IS WORTH 3.0 TIMES AS MUCH AS A CARRY IN PPR LEAGUES.

With this knowledge at our disposal, we can better approximate the value of a player’s role than through raw touches. By multiplying a running back’s red-zone carries by 1.33, red-zone targets by 2.27, outside-the-red-zone carries by 0.49, and outside-the-red-zone targets by 1.48, we can sum up these numbers to create what I’ve been calling a running back’s “weighted opportunity.”

Using the methodology I outlined above, here were 2022’s top 30 running backs by weighted opportunity per game:

Weighted Opportunity Points per Game (2022)

What does the differential represent?

Should we be avoiding players with a negative differential (assuming this represents a measure of efficiency, and thus, skill) or should we be targeting these players (assuming this is a number that tends to regress to the mean)? This isn’t an easy question to answer, but it’s mostly the latter, while still being some of the former.

Last season, Nick Chubb ranked only 17th in weighted opportunity points per game (13.8). By scoring 16.6 fantasy points per game (8th-most), he produced a +2.8-point differential. But it shouldn’t at all be surprising that Chubb scored more fantasy points than his weighted opportunity would suggest, just like it’s not surprising Chubb ranks highly in yards-per-carry every season. Because Chubb is really freaking good. A player’s weighted opportunity is based on the average of all running backs over the past five seasons, and Chubb is much better than a perfectly average running back.

So, is Chubb a regression candidate? I’d say probably not. A +2.8 differential isn’t inordinately high, and is actually significantly less than Chubb’s career average (+3.9).

But, typically, we should expect a heavy regression to the mean. And the higher the differential, the greater the regression to the mean. Of the top 40 seasons this past decade (by positive differential), 34 regressed the following season (85%). On average, the fall was from a differential of +3.46 to +0.48. Of the bottom 40 seasons this past decade, 39 regressed (positively) in efficiency the following year. On average, the climb was from -3.08 to -0.68.

Therefore I’d wager that Tony Pollard’s league-high +2.9 differential is going to regress to the mean. Or at least his mean (+1.4 since entering the league). This doesn’t mean that he’ll necessarily regress in fantasy points per game or weighted opportunity points per game, only that the differential between those two numbers is very likely to be much closer to zero (though still likely a little above average) next season.

To help aid in our analysis, here are the top 20 RBs by ADP alongside their differential over the past three seasons combined:

Notes: All RBs who failed to play in at least 18 games over this span were removed from our sample.

Player Analysis

1) Last season, there weren’t any players with a positive differential ranking top 40 this past decade. In fact, there wasn’t even a single RB in the top 100 this past decade. Tony Pollard (+2.9), Nick Chubb (+2.8), Derrick Henry (+2.3), and Christian McCaffrey (+2.2) would stand out as the top positive regression candidates. But these are fairly tame differentials historically, and these numbers are all fairly in line with their career norms.

2) The following players all had seasons ranking bottom 40 by differential this past decade: Jonathan Taylor (-3.2), Alvin Kamara (-2.5), Brian Robinson (-2.1), Dameon Pierce (-2.1), Michael Carter (-2.0), and Najee Harris (-1.6). It appears this was a uniquely and outlier-ishly inefficient season for fantasy RBs.

For all of these players, we should be expecting a (positive) regression to the mean next season. But then again, that doesn’t necessarily mean we should be expecting more fantasy production – only a positive increase by differential. And it’s entirely possible, and I think probable, that many of these RBs will see a massive reduction in volume next season (due in part to their abysmal efficiency last season) and thus a significant drop in fantasy production.

3) Tony Pollard (+2.9) is technically a top regression candidate – he’s almost certain to be less efficient in the touchdown department next season – but that doesn’t mean he’s a bad value at cost (ADP: RB7). Often enough, a hyper-efficient season will lead to more volume in the following year. And that certainly appears to be an inevitability this year, with Ezekiel Elliott no longer with the team. If Pollard captures just 50% of Elliott’s now vacated XFP and if he’s only half as efficient as he was last season, we should expect him to score 20.0 FPG, which would have ranked 3rd-best last season.

4) Dameon Pierce was technically one of the least efficient RBs in fantasy last year, falling short of his volume-based expectation by 2.1 FPG. That said, I think Pierce was actually quite good last year, it’s just that his offensive line and supporting cast were league-low levels of bad — he ranked above average in both YAC/A (3.21) and MTF/A (0.22), but (thanks to poor offensive line play) a league-high 75% of Pierce’s yards came after contact. And this raises an important point – it’s extremely difficult to separate a RB’s skill and efficiency from that of the offense as a whole. So, although I do firmly believe in Pierce’s talent, just as I did when he was drafted, it’s still hard to get too excited about him this year. The team added more competition, upgrading Rex Burkhead with Devin Singletary, and Houston’s offensive line still projects to be bottom-10.

5) I think there’s a compelling upside argument for J.K. Dobbins. Sure, he’s missed more games than he’s played across his career, but he’s also been the single most efficient RB in fantasy over this stretch, out-scoring his expectation by +53%. Even last year, once he was activated off injured reserve in Week 14, he rounded out the season by averaging 100.0 YFS/G and 6.9 YPC. With a new offensive playcaller, we could see less of a committee this year, as well as more targets for Dobbins. That is, so long as he can stay healthy.

6) Last season, Joe Mixon finished 3rd in WO/G (18.3), behind only Austin Ekeler and Christian McCaffrey. Based on his -1.2 differential and RB16 ADP, he’s both a glaring positive regression candidate and a clear top value. Across his previous four seasons, Mixon outscored his expectation by +2.4 FPG, with a low of just +1.1. Apparently, there’s still some risk that the Bengals could cut Mixon, although I find that hard to believe. And instead, Mixon stands out to me as the best RB value in current drafts (for reasons outlined in more detail here).

7) I think James Conner is being severely underrated (ADP: RB28). He ranked 11th in WO/G last year (15.2) and ranks as one of the most efficient RBs in fantasy over the last three seasons (+23%). As I explored in more detail elsewhere, Conner currently rivals Mixon as the best mid-round value at current ADP.

8) I’m having an impossible time trying to get a good read on Najee Harris. He’s been one of the least-efficient RBs in fantasy over the past two seasons (-1.5). But then again, his volume has been great, his Round 1 draft capital typically means his team will be slow to give up on him even if he struggles, and (like with Pierce) his supporting cast wasn’t doing him any favors. Plus, I’m not sure Harris was actually “bad” last year, rather than merely “injured.” Harris suffered a Lisfranc sprain in training camp, but from Week 11 on, Harris averaged 15.8 FPG (8th-most among RBs) with an improved to a -0.4 differential over this span. That’s encouraging, but also, Harris only played on 64% of the team’s snaps from Week 11-on (down from 83% in 2021). Noted bell-cow advocate Mike Tomlin is no longer the sole arbiter of RB playing time, and Pittsburgh’s coaches (see last link) and beat writers can’t stop raving about backup Jaylen Warren… All this to say, his RB12 ADP isn’t terrible, but he’s not quite a RB I’m actively targeting.

9) It’s hard for me to get excited about Alvin Kamara, even at his depressed RB31 ADP. And, no, not just because he could be imprisoned for up to five years, nor because the Saints just drafted a RB I was very high on (Kendre Miller in Round 3). Mostly it’s because Kamara has struggled to stay healthy for most of his career, is now entering his age 28 season, and because he’s showing signs of a player in serious decline. (In fantasy football, it’s always better to bail on an aging RB a year early than a year late.) Through his first four seasons, Kamara exceeded his expectation by +5.6 FPG (3rd-most among all RBs). Since then, he’s fallen short of his expectation by 2.3 FPG (ahead of only Sony Michel and Mark Ingram).

10) Jonathan Taylor, meanwhile, is one of the most obvious positive regression candidates we’ve seen in quite some time. He fell short of his expectation by 3.2 FPG last year, but rose above his expectation by 4.8 FPG across his first two seasons. Some of this is due to Taylor’s ankle sprain (suffered in Week 4 and reaggravated in Week 15), as well as Jim Irsay’s embarrassing decision to let Jeff Saturday coach the franchise. If excluding the Week 15 game in which Taylor only played on 2 snaps, he averaged 18.0 WO/G, which would have ranked 4th-best last year, and just 0.4 points shy of his career-high.

11) It’s worth pointing out a few things from the above chart. First, that Austin Ekeler averaged 20.3 WO/G last year, which ranks 3rd-best by any RB this past half-decade. Outside of the red zone, Ekeler ranked only 30th in carries (153), but he ranked 2nd in red zone carries (51) and 1st in total targets (123). As I’ve been trying to articulate, that sort of volume – these high-value-opportunities – are what really matters in fantasy. Given that Ekeler is on the last year of his contract and is facing minimal competition for backfield touches, I see no reason why Ekeler shouldn’t be in play as a top-5 pick in fantasy drafts… The second thing I’d like to point out is that Chrisitan McCaffrey saw the 2nd-best workload of his career last year. And that’s no small feat for a RB who owns 6 top-20 seasons (in 6 career seasons) this past decade. Better yet, McCaffrey’s volume was also even better once he joined the 49ers – from Week 8 on, McCaffrey averaged 19.5 WO/G. Granted, Elijah Mitchell’s absence surely played a role. But still, like Ekeler, McCaffrey is well deserving of a top-5 fantasy pick to me.