The Fantasy Footballers Mythbusters: Zero RB Strategy
If you’ve played fantasy football for a while, you’ve probably heard about the ‘Zero RB strategy’; you might have even tried it yourself at some point. This latest entry in the Fantasy Mythbusters series is simple: is Zero RB a legitimate draft approach? Fantasy data is from nflfastR, while ADP data is from the fantasy football calculator.*
What is Zero RB?
It’s not controversial to say that running back is the most important position in fantasy, thanks to a combination of skew – the best RBs score so much more than the rest of the field – and scarcity – there simply aren’t enough great RBs to go around. They dominate the start of drafts for that very reason: if you can roster a room of elite RBs, you have a vastly improved chance of finishing atop your league.
We can see this clearly if we plot average fantasy PPG by positional rank (i.e., overall RB1, etc.) across the main fantasy positions (QB, RB, WR, and TE) since 2010. While the RB dots start out high – the RB1 scores nearly 25 points a game, not that far from the QB1 – look at how fast the RB line decays. Although the WR1 scores were much less than the RB1, it only takes to the WR8 to see the wideouts scoring more than running backs. What’s more, this gap widens: the green RB line gets farther and farther away from the purple WR, as the value of deeper RBs just falls off a cliff. The vertical lines denote the 12th and 24th positional spots; I’ve only drawn the top 12 QBs and TEs since, in non-Superflex leagues, you probably won’t be looking much deeper at these positions, whereas you will have to start depth RBs and WRs.
However, running backs can also be fickle. The position is extremely physically demanding, and backs are often (unfortunately) injured, ‘age out’ shockingly quickly (Todd Gurley, Le’Veon Bell, etc.), see value get diluted by the ‘committee’ approach, or just plain underperform. It can be very frustrating to invest so much draft capital (your first few picks) in running backs that don’t pan out.
This is exactly the type of uncertainty that the Zero RB strategy hopes to take advantage of. A manager using this strategy will, instead of using high draft picks on running backs, instead target the other roster positions: QBs, WRs and TEs. One potential advantage is more certainty in getting a return on investment, especially in the relatively more stable positions of WR and TE (compared to RBs). Another advantage is ‘zigging where everyone else zags’: since most managers will be targeting running backs early on, you can get the cream of the crop at all of the other positions.
Of course, ‘Zero RB’ doesn’t actually mean that you don’t pick any running backs: you just defer your RB picks to later rounds (ideally when the rest of the managers have pivoted to other positions). You won’t get the top backs off the board, but you’ll be able to build an extremely strong foundation at other positions and then play the game of trying to find values / breakouts at running backs. You might plan to be very active on the waiver wire in-season, spending high amounts of FAAB (or waiver priority) to target RBs that emerge.
This might seem like an unintuitive strategy, largely because, by definition, it goes against the grain. Indeed, there is a ton of discussion in the fantasy community as to the merits of this approach. In a good scenario, you build out a superstar non-RB roster and hit on some solid running backs late in the draft / pick up a young back on the rise as the season progresses. In a bad scenario, you get positionally dominated every week because your meager RB group barely puts up any points.
When it comes to Zero RB, is it genius, or just heinous?
Zero RB Simulation
In an ideal world, we would look at a large number of rosters that followed the Zero RB strategy in a draft and measure their performance in-season. Unfortunately, this is clearly infeasible: it’s hard to get a large dataset of actual drafts and, in any case, the ‘draft strategy’ that managers use isn’t really labeled!
To tackle this problem, then, we can simulate a large number of ‘Zero RB’ rosters. I did a mock draft with a Zero RB strategy on Sleeper (which, in general, is a great tool to prepare for your draft). This was for a Half PPR Superflex: one of your FLEX spots can be a QB and, in almost every case, it is indeed prudent to start a QB in your second FLEX spot, so you can really think of this as a 2-QB league!
My rule was to only select running backs after Round 5. Here were the first seven rounds of my draft (from the 8th overall pick):
Zero RB draft:
As expected, this lineup is stacked at QB and TE and has a solid WR core (if I had prioritized WRs before QBs, which could certainly be shifted) and dubious starting running backs. Using this and a few other mocks, I sketched out the position ranks that you can expect with the Zero RB strategy:
Quarterback 1: QB1 – QB6
Quarterback 2: QB6 – QB12
Running Back 1: RB20 – RB25
Running Back 2: RB24 – RB30
Wide Receiver 1: WR1 – WR12
Wide Receiver 2: WR10 – WR20
Tight end: TE1 – TE3
So, for example, you can expect a top-6 option as your QB1, a top 3 option as your TE, etc. Naturally, your expected starting RBs will be lower ranked! We can do a similar exercise but with an RB-heavy approach (taking RBs with the first two picks). Here is my mock from the 1.04:
As expected, I got an elite RB room and sacrificed elsewhere (especially at TE). Using this and a few other mocks, I sketched out these ranges for what you can expect with this strategy:
Quarterback 1: QB5 – QB10
Quarterback 2: QB11 – QB15
Running Back 1: RB1 – RB4
Running Back 2: RB5 – RB12
Wide Receiver 1: WR8 – WR14
Wide Receiver 2: WR13 – WR24
Tight end: TE9 – TE12
Again, as expected, you will get elite RB talent – two RB1s – and generally take lower options at the other positions. With these ranges, our algorithm is now simple:
- Randomly select a year from 2015-2020.
- Randomly select players within each range (i.e., for the Zero RB strategy, select a QB1 – QB6 as your first quarterback).
- Find the average PPG for each player through Week 5 of that season (after Week 5 injuries / bye weeks start to pile up), and calculate the average score for the whole roster.
- Do steps 1-3 thousands of times.
We can do this for both the Zero RB and ‘RB heavy’ strategy and get simulated rosters for many of these types of drafts over many different years. Here’s the resulting player distribution for the RB heavy strategy:
And for the Zero RB strategy:
- Right off the bat, we should note that the Heavy RB strategy scores about two more points on average than the Zero RB strategy (with a simulated win rate of about 53%). This difference is small, but highly statistically significant.
- When we compare the two charts, the player distributions are as expected. The green RB ‘mound’ is shifted to the right (with a longer right tail), indicating higher performance, in the first chart, whereas the QB and TE mounds are shifted to the right, indicating higher performance, in the Zero RB charts.
- Interestingly, the WR mound is relatively similar between the two charts. This could have to do with more depth at the position, or it could just be the ‘ranges’ I set up above, which prioritize QB and TE in the Zero RB simulation as opposed to WR. If we targeted WRs early instead of TE or QB, it’s likely that the WR mound would shift right and the TE or QB mound would take a hit.
PLAUSIBLE. Are you surprised? We saw that the Zero RB strategy did worse in our simulations!
Although technically, in our simulation a Zero RB draft strategy yields a lower score by about two points on average, this is a pretty small difference. There are plenty of factors that could bridge this gap: adjusting the player ranges I set, added depth at RB later in the draft, spending a lot on FAAB early to pick up a breakout RB after a Zero RB draft, etc. In my opinion, the fact that the Zero RB simulation was relatively close to the RB heavy draft means that, in certain situations, it can be a viable strategy.
What are those situations? Well, you’ll probably want to be drafting later in the first: Zero RB likely isn’t a good move with the 1.01 (in non-Superflex leagues), since you should just take Christian McCaffrey (this could lead to a ‘Hero RB’ draft, where you just select one back early)! The strategy is more effective in leagues where other positions are emphasized, like Half PPR or (especially) full PPR, Superflex or 2-QB, or TE-premium leagues (TEs get 1 point per reception, everyone else is Half PPR). Finally, it’s most effective when no one else is doing Zero RB, or you will be targeting the same players! All of these situations serve to slightly devalue the running back position and make the Zero RB approach more likely to work.
Questions? Tried a Zero RB out and loved it? Message me on Twitter.
*Technically, I’m using Half PPR points, but ADP using standard scoring. This isn’t ideal: the reason is that I can’t get Half PPR ADP data going back very far, which I need for the simulation to work. Anyways, it won’t matter too much: ADP won’t vary an extreme amount in between Half-PPR and standard scoring and, in any case, we’re looking at ranges here.