The App is Back! Don’t forget to download the NBC Sports EDGE app to receive real-time player news, mobile alerts and track your favorite players. Plus, now you can check out articles and player cards. Get it here!
In 2015, Le'Veon Bell was suspended for the first two games of the season. He came back in Week 3, played six games, and tore his MCL in Week 8. DeAngelo Williams took over as the workhorse and finished as the RB4 in PPR formats. His preseason best ball ADP was 169.6. In games Bell missed, Williams averaged a whopping 21.5 points.
Here's the thing: Williams was the exception rather than the rule. In fact, it's those situations that lead fantasy players to make the suboptimal decision of drafting a running back who is purely an insurance policy. It's easy to draft two RBs on the same team and think you have that backfield locked up for the entire season, but it usually doesn't work out that way in practice – even if the RB1 actually does get hurt.
With that in mind, let's dive into the data to see why it's usually a poor decision to draft this type of back.
First, we have to define what it means to be an RB1 so we know which RBs actually qualify as insurance backs.
Over the last six seasons, the average difference in ADP between an NFL team's RB1 and RB2 was 98.5 spots. For this exercise, we'll say a gap smaller than that was a committee, while a disparity larger than that was a starter-and-backup situation. We're not interested in committees today, so we're just going to look at the second group.
We're also only going to hone in on situations where the starter had an ADP in the first four rounds. Matt Jones was a Round 6 pick in 2016, but you probably didn't feel the need to draft his backup to lock up the Washington backfield. By limiting our sample to the first four rounds, we make sure we're only looking at situations where you might actually want a backup RB solely as insurance. After all of this, we're left with a sample of 79 RBs since 2015.
We'll assign each of them an insurance RB based on their most-drafted backup. The reason we're not using ADP is that it can be skewed in a small sample of drafts. For example, Mike Davis had an ever-so-slightly earlier ADP than Reggie Bonnafon last season, but Davis wasn't selected in 98.6% of best ball drafts.
A Deep Dive on Insurance RBs
Our sample of 79 starting RBs missed an average of 2.9 games between Week 1 and Week 16. 30 of them played all 15 games (38.0%). Another 12 missed just one game (15.2%). Right away, that's more than half the sample that missed one or fewer games.
16 of them missed six or more games (20.2%). Let's go one at a time to see how the preseason RB2 performed in those cases.
|Player||Year||Games Played||Preseason RB2||RB2 ADP||RB2 Best Ball Win Rate||RB2 Points Over ADP-Based Expectation||Notes|
|David Johnson||2017||1||Andre Ellington||222.0||5.8%||42.3||The market barely leaned toward Andre Ellington over Chris Johnson as the Cardinals' RB2 in 2017. It didn't end up mattering, as Arizona traded for Adrian Peterson after Week 4. Ellington and Johnson both got waived midseason. Peterson and Kerwynn Williams were their two leading rushers.|
|Saquon Barkley||2020||2||Dion Lewis||224.1||7.9%||17.4||Dion Lewis was picked in six times as many drafts as Wayne Gallman, who ended up leading the team in rushing. The Giants also signed Devonta Freeman and relied on him for a few games.|
|Christian McCaffrey||2020||3||Reggie Bonnafon||224.1||5.2%||-26.4||Fantasy players picked Reggie Bonnafon more than four times as often as Mike Davis, who finished as a low-end RB1.|
|Adrian Peterson||2016||3||Jerick McKinnon||143.4||6.3%||22.5||After Peterson's injury, McKinnon shared the load with Matt Asiata. McKinnon got the pass-catching work, while Asiata handled short-yardage situations and goal-line carries.|
|Jamaal Charles||2016||3||Spencer Ware||178.2||11.5%||118.7||Ware was the Week 1 starter for the Chiefs. Charles struggled to return from a season-ending injury suffered in 2015 and reaggravated it upon his eventual return, allowing Ware to finish as a high-end RB2.|
|Arian Foster||2015||4||Alfred Blue||147.7||7.6%||3.9||Blue was the Texans' leading rusher in 2015, but he ceded passing-down work to Chris Polk and Jonathan Grimes. Between a one-dimensional workload, a Brian Hoyer-led offense, and subpar efficiency, Blue couldn't capitalize on his opportunity.|
|Jamaal Charles||2015||5||Knile Davis||125.0||4.2%||-87.5||It was Charcandrick West and the aforementioned Ware who emerged once Charles went down. Davis finished with just 28 carries in 2015.|
|Eddie Lacy||2016||5||James Starks||168.8||7.2%||-28.3||Starks was next in line after Lacy's injury but underwent surgery for a torn meniscus, which opened the door for Ty Montgomery to take over.|
|Le'Veon Bell||2015||6||DeAngelo Williams||169.6||16.6%||144.5||Like Ware, Williams opened the season as the starting running back while Bell served a two-game suspension. After Bell got hurt in Week 8, Williams took the starting role once again and ran with the opportunity.|
|Joe Mixon||2020||6||Giovani Bernard||208.2||10.9%||101.7||Bernard became a fantasy-viable asset without Mixon. Samaje Perine got a solid snap share in the second half of the year, but it was still clearly Bernard's job.|
|Marshawn Lynch||2015||7||Christine Michael||196.2||5.7%||-49.1||This one was tough because Christine Michael was the presumed backup all offseason but got cut before Week 1. The Seahawks then signed Fred Jackson, who was also drafted in almost every league. It didn't end up mattering because Thomas Rawls emerged in Lynch's absence.|
|Kerryon Johnson||2019||7||C.J. Anderson||186.7||6.4%||-70.7||Anderson got cut after two games with the Lions and retired soon thereafter. It was a three-back committee between Bo Scarbrough, Ty Johnson, and J.D. McKissic for the rest of the season.|
|Leonard Fournette||2018||8||T.J. Yeldon||195.2||11.5%||106.6||Yeldon paid off in a big way for those who didn't trust Fournette's health, carrying the rock 104 times and notching 78 targets.|
|Thomas Rawls||2016||8||C.J. Prosise||142.3||5.9%||-39.2||Prosise was dominant in his one healthy start (153 total yards on 24 touches), but he fractured his scapula in the next game and missed the rest of the season.|
|Ezekiel Elliott||2017||9||Darren McFadden||158.1||5.4%||-93.8||McFadden was a healthy scratch in most games and ended the season with one carry. Alfred Morris and Rod Smith carried the load when Elliott missed time.|
|Austin Ekeler||2020||9||Justin Jackson||161.4||8.5%||-39.8||Jackson got hurt and missed some of the same games as Ekeler. When he was healthy, he was productive in conjunction with Joshua Kelley.|
For those unfamiliar with best ball, the average win rate is 8.3% (100 percentage points divided by 12 teams per league). Since win rate data can be noisy, I also used local regression to estimate how many points you'd expect an RB to score based on their ADP. The RBs listed above averaged a 7.9% win rate and 7.7 points over ADP-based expectation. Let that sink in: Even if you only look at cases where a starting RB missed significant time, insurance RBs still posted a below-average win rate and barely returned value at ADP.
It's clear we are overconfident in predicting who is going to emerge in the event of an injury. Some of these guys – Ellington, Bonnafon, Lewis – went so late that role uncertainty was baked into their ADP. Others – Davis, Michael, Anderson, McFadden – were being picked as the clear backup in every single draft and ended up as an afterthought.
We think we know more than we do. It's a human thing. We can capitalize on that because there's an edge in knowing the market is overconfident. Even when it looks like an RB is next in line, so much can happen throughout an NFL season to prevent that player from actually producing. Maybe their team opts for a committee approach if the starter goes down. Maybe they sign or trade for someone else. Maybe the RB2 gets hurt at the same time as the RB1. The opportunity of insurance backs even when an injury occurs is often much more fragile than ADP indicates.
Zooming out further, 49 of the 79 RBs in our sample sat out at least one game with an average of 4.7 missed games, which freed up 240 potential productive weeks for RB2s. In those games, the expected output out of an insurance RB was 9.4 PPR points. Think about that: If an RB1 missed a game, his preseason backup's expected value was just 9.4 points.
It's important to realize this is best ball-centric analysis. The process changes in season-long formats. In best ball, the cost of being wrong is much higher because it means you have to take a zero on your roster every week. If you picked Bonnafon in best ball last season, he probably never contributed to your team and there was absolutely nothing you could do about it. In season-long, you could have just dropped him as soon as the Panthers released him and picked up Davis instead. This is also dependent on when you're drafting. Bonnafon got cut before Week 1. If you were drafting the week before the season, you would have known to draft Davis instead.
In best ball, the risks of taking an insurance RB usually outweigh the benefits. First of all, we are overconfident in predicting who would emerge in the event of an injury. If a starting RB goes down, history indicates NFL teams rarely rely exclusively on the next RB in line – and we're not especially good at predicting who is the next RB in line. It's slightly different in redraft formats because you can manage your roster in-season, but this still tells us to be wary of holding an insurance back all season waiting for an injury to occur.