What can the difference in average draft position between two running backs on the same team tell us?
This post is an update to some work I did a few years ago; they’re here if you’d like a deeper look at the thought process behind this player screen, or just a laugh at some of the hot fantasy RBs of seasons past.
- Mind the Gap: Using ADP to Create a Running Back Typology
- Using ADP to Identify Running Back Traps and Values
- Running Back Performance by Team ADP
Like my most recent articles, which used average team dynasty rankings to find low-cost acquisition targets, RB ADP differential is really just a way to screen players, and separate them into different buckets. Once we know which buckets have performed better in aggregate, we’ll know which players to analyze more closely.
First, I’ll explain my methodology and then we’ll identify some 2018 RB targets.
Here’s how it works. A player’s ADP represents the market’s best assessment of his individual value. What if the difference in ADP between two RBs on the same team also tells us something about each player’s relative value? Consider the Arizona Cardinals and Atlanta Falcons. Each team has a pair of RBs being drafted in MFL10 best ball leagues.
Since ADP reflects the collective prediction of the crowd of drafters, we should expect these two players to perform differently. David Johnson is being taken roughly 234 picks ahead of Chase Edmonds. That’s about as far apart as you can get in a 240 pick draft. Clearly, these are different types of RBs, and it doesn’t seem as if drafters expect Edmonds to have any sort of effect on Johnson’s role.
Not so in Atlanta. Devonta Freeman and Tevin Coleman are separated by just 42 picks. Here, the community of drafters is sending a different message. It’s less clear which will be the most productive, or by how much, and Coleman carries an expectation of value on his own. Since there’s less confidence about this backfield, there’s more opportunity to get an ADP-beating return on investment.
When I first did this work in 2014, best ball leagues weren’t a thing, so I simply looked at PPR redraft leagues. This time around, I’m looking at MFL10 performance by RB type, but the conclusions are the same. All data comes from FFPages, and includes RBs appearing in more than 10 percent of drafts, August ADP, for the 2015-2017 seasons.
Relative ADP-Based Running Back Types
To test for the idea that drafters expect a team’s second RB to have value, I divided same-team pairs of RBs into four groups, based on whether or not the difference in ADP between them was greater or lesser than the average ADP gap for all same-team RBs.1 This screen gives us four distinct types of backs.
|TEAM RB||1st Drafted||2nd Drafted|
|Big ADP Gap||B1||B2|
|Small ADP Gap||S1||S2|
In this example, David Johnson is a B1 back, Chase Edmonds is a B2, and Freeman and Coleman are S1 and S2 respectively. Johnson is expected to be a stud, Edmonds is a straight handcuff, and the Atlanta backs are in a committee. These aren’t new concepts, but looking at them in these buckets reveals some interesting insights.
RB Performance by Type
|RB Type||Pct N||Ave ADP||Pct Top 6||Ave Applied Pts||Ave Val Add||Win Rt > Teammate?||Win Rt > AveRB?||Ave Win Rt|
S2 backs are the most likely to outperform their teammate in MFL10 win rate.
The B-type rows make intuitive sense. The top row tells us that B1 backs — think David Johnson and Le’Veon Bell — comprise just 14 percent of the data set. However, they account for over 24 percent of the top-six league finishes.4 Meanwhile, B2 backs — think Chase Edmonds or James Conner — appear on only half as many top-six rosters, making them a poor investment. They add far less than average RB value to your team.5 The B2 win rate vs. their B1 teammates is misleading since these backs tend to appear on very few rosters, as evidenced by the low rate of top-six finishes.
The most interesting takeaway is that the S type, or committee backs, look an awful lot like the B1 backs. The S backs feature much lower ADPs yet clear this basic threshold of success (top-six finish) at the same rate as the more highly regarded B1 backs. S2 backs look particularly nice; their average ADP is about five rounds later than their backfield teammates, yet the applied points and value added averages6 reveal a similar difference between RB types.
This table also tells us how often an RB has an MFL10 win rate greater than their teammate, and how often he has a win rate greater than the average win rate (0.074) for all RBs. The poor performance of B1 backs reflects their higher risk. They add more points and more value to your lineup than any other type of back — when they have good seasons. When they don’t, the loss of production from a high-value draft pick is devastating. They will definitely help you win (David Johnson, 2016), but they’ll also help you lose (David Johnson, 2017).
Again, the best performance comes from the S2 backs. They have a lower ADP than either of the “lead back” (B1 and S1) categories, so the risk of poor performance is mitigated, but the upside of good performance is captured in your lineup. The S2 backs outperform the average back at a better rate than any of the other three types, and are the most-likely cohort to outperform their teammate.
The presumptive stud B1 backs deserve full consideration when you’re on the clock — but you didn’t need me to tell you that. What I’m suggesting is that the next set of backs you should target are the S2s, while fading the B2s and S1s.7 With that in mind, here are my fades and targets.
A couple of notes. First, my analysis is based on August ADP, when drafters have the most knowledge about backfield roles. It’s only the beginning of July, so this list may change. Second, I’ll identify which backs fit into this disproportionately successfully cohort, but that just gives us a starting point for further analysis; at the level of the individual team and player, some situations will be better than others.
|Kamara, Alvin||NO||6.0||S1||PROB B1 BY AUG|
Filter for “TAR” or “FADE” to see my list, but also use the rest of the table to look for your own opportunities.
- Devonta Freeman, Jordan Howard, Joe Mixon, Marshawn Lynch, Kerryon Johnson, Isaiah Crowell, and Marlon Mack are among the priciest S1 backs. In other words, these are the “lead” backs about whom drafters have the least confidence. Therefore, their battery mates make great S2 targets.
- Devonta Freeman (S1) – In two of the past three seasons, Atlanta’s RBs have been of the small ADP gap variety, and in both cases, the later-drafted S2 back had the better MFL10 win rate.
- C.J. Anderson (B2) – Based on ADP gap, drafters see him as nothing but a handcuff, however, he’s being drafted ahead of many backs that have a more desirable (S1 or S2) profile.
- Jerick McKinnon, Kenyan Drake, Rashaad Penny, Alex Collins, and Ronald Jones represent the cheapest B1 backs. That mitigates as much as possible the impact of a bust, while keeping alive the hope of a dominant RB1 season.
- Theo Riddick, Devontae Booker, D’Onta Foreman, Giovani Bernard, Nyheim Hines, Bilal Powell, and Doug Martin represent the cheapest S2 backs, making them my preferred RB targets in their ADP ranges.
Knowing which backs look vulnerable (e.g. expensive S1 and B2 backs) or poised for outsized returns (cheap B1s and S2s) can help set your draft boards. For those who play in many leagues, acquiring exposure to the least expensive members of the most desirable cohorts makes sense.
Stay tuned for more in-depth analysis of these RB types. As ADP data become more informed, and our team and player projections more detailed, we’ll fine-tune our list of targets.
- The average difference is about 101 picks. (back)
- Two teams had just a single, highly-drafted RB; I included those as B1 backs. (back)
- I’ll write about the other types of backs at a later time. (back)
- I’m using top-six finish as a basic proxy for success. It’s good enough to cash in 2X leagues and in a normal redraft would get you to the playoffs. (back)
- The average for all 330 RBs is 75 applied points and 46 value added. (back)
- Defined by FFPages creator Gentan Schulteis: “Applied points is the total points the player had in weekly starting lineups. That display is the average of all teams. Value added is kind of similar, but I subtract the ‘next best starter’ for each week if he started. That leaves how many extra points a player had above what would have been scored if he had zeros across the board.” (back)
- Or the first set, if you’re going ZeroRB. (back)