Visualizing the Best Rookie RB Class in Years

Is the 2017 rookie running back class really that good, or am I just buying into the hype? That’s the question I’ve been asking myself as I’ve been consistently taking more RBs than usual in rookie drafts this spring. 

Sure, underwhelming results at the NFL scouting combine dimmed some of the shine for guys like Leonard Fournette and Dalvin Cook, but at least it’s still a deep class, right?  I mean, I feel a helluva lot better about the RBs I’m seeing in the second round this year than I did back in 2016. Or am I just succumbing to good old fashioned rookie derangement syndrome? With a few rookie drafts still to come, I decided to analyze whether I was justified in thinking that this RB class was better than recent seasons, or if the perception really does match reality.


Quantifying the quality of a prospect is a tricky thing, but fortunately, we have some tools at our disposal. The first is the NFL draft. Draft position correlates well to NFL success and is a good starting point for any analysis.1

As fantasy players though, we’re always looking beyond the draft to find the value that’s lurking beneath the surface. Enter RotoViz’s RB Prospect Lab. The Prospect Lab gives us a draft-agnostic method of comparing prospects’ athletic and production profiles to get a rough idea of how they project to the NFL.

By looking at draft classes through the draft and the Prospect Lab we can get a clear picture of the relative strength of each year’s draft class. NFL draft position incorporates teams’ knowledge of factors that can’t be included in a statistical model, while the Prospect Lab can cut through the biases that are part of traditional scouting and highlight players that may be getting either overlooked or overvalued.

Since we’re looking at this in the context of fantasy, I took the top-20 RBs in positional ADP from each class.2 As you’ll see in the charts this also works out to a cutoff right around the end of the fourth round of rookie drafts.3


We’ll start with the NFL draft, looking at both the average draft position of the 20 prospects in the sample, as well as a visualization of the way players were distributed in each class:


2017 looks like a clear winner, and my perception that it was way better than 2016 seems to be validated. 2016 experienced the steepest dropoff at the top of the draft. In 2016 the sixth RB, Kenneth Dixon, was drafted with pick 134, whereas in 2017, by the time Jamaal Williams was taken with pick 134, he was the 13th RB off the board.

2015 was also considered a fairly strong RB class, headlined by first-rounders Todd Gurley and Melvin Gordon – after no RBs were selected in the first round the previous two seasons – and the results seem to bear that out. 2014 does edge out 2015 on average, but most of that comes from the very tail end of the draft with few studs at the top. It’s worth noting when looking at the averages that a difference of 30 picks at the top of the draft is a lot more significant than a difference of 30 picks at the bottom of the draft, so 2014 is probably not as strong as it appears.

And now, let’s look at the Prospect Lab scores using the same visualizations we used for draft position:


2017 still comes out on top. This year’s draft class features five prospects with a score of 70 or higher, and seven of 60 or higher, whereas every other class has no more than three and five, respectively. This class also appears to be deep towards the bottom. All four of the classes from 2013 through 2016 take a pretty steep drop towards the end, while 2017 maintains some prospects with decent profiles. In 2017, Wayne Gallman checks in with the lowest score (25), while previous years bottomed out with guys like Andre Ellington (0), Alfred Blue (0), Michael Dyer (3), and Kelvin Taylor (21).

2016 performs much better in the Prospect Lab than it did in NFL draft position. The class of ’16 benefits from the two studs at the top and the drop-off after those two is jarring, which probably contributed to my perception that 2016 was a poor class.4 Overall, 2016 looks better than the 2013 and 2014 classes. It’s easy to see how having this class sandwiched in between the strong 2015 and 2017 classes could make it look worse than it actually was.

2015 overtakes 2014 by a decent margin, suggesting that the distribution of draft picks in 2014 probably was indicative of a weaker class than it appeared, with no elite options at the top. Further analysis supports this, as the 2014 class has mostly underwhelmed since arriving in the NFL, while the 2015 class has four RBs in the top-30 of dynasty startup ADP and several other solid contributors.

Finally, if we standardize the NFL draft position and Prospect Lab score for each player using z-scores, average each class, and then combine them, we get a rough ranking of where each class stands using both criteria:

Draft Class Avg. PLab z-score Avg. Draft z-score Combined
2013 -0.27 0.09 -0.36
2014 -0.18 -0.10 -0.08
2015 0.05 -0.03 0.08
2016 0.09 0.35 -0.26
2017 0.31 -0.31 0.62

The 2017 RB draft class has the best Prospect Lab scores and NFL draft positions over the last five years by a relatively large margin and indeed appears to be a strong class for RBs when looked at through two different lenses of prospect evaluation.


So now let’s answer the next logical question: “Does it matter?” If a stronger draft class doesn’t actually translate to fantasy success, then these results, while interesting, won’t really help us when deciding how to approach our rookie drafts. To test this, I averaged the yearly PPR points from the same 20 players used above and compared how they performed in each season they’ve played so far.

If a stronger draft class doesn’t actually translate to fantasy success, then these results, while interesting, won’t really help us. To test this, I averaged the yearly PPR points from the same players and compared how they performed in each season.


While acknowledging that we’re working with a sample of just 10 total seasons and that there are many factors that affect PPR scoring, the results are promising and correlate well with the rough rankings from above.5

The next step is to see if fantasy drafters recognize this by way of ADP. There are a lot of factors that go into rookie ADP. Perceived opportunity plays a huge role, and rightfully so. Overall ADP will also be affected by the relative strength of other positions. With this in mind, here’s how drafters have approached these RB draft classes:


Drafters do recognize that the 2017 class is strong, but perhaps not to the extent that they should. While rookie ADP correlates closely to NFL draft position, there may be some values to be found using the scores from the Prospect Lab.

Who Are the Best Rookie RB Draft Targets?


The two obvious values here from the 2017 class are D’Onta Foreman and Jeremy McNichols. There’s reason to question whether the Prospect Lab model is overrating them as they both come from college systems that may have inflated their raw numbers. But they’re both so far above other players taken at the same ADP in past years that they make great values where they’re currently being picked. Foreman became a RotoViz favorite after he killed it at his pro day, and McNichols quietly has one of the better all-around production and athletic profiles we’ve seen from a fifth-round NFL draft pick.

Foreman, in particular, is interesting because his 40 time and three-cone virtually max out the Prospect Lab.6 The other players who scored at the top of the lab have all been stars when healthy, however, they were also all drafted much earlier than Foreman and caught more passes in college.

Even if we assume Foreman isn’t quite as good as the model suggests, it’s a giant leap to believe that he’s not at least a great upside play at the top of the second round. Kevin Cole’s RB success model also loves Foreman, and if the lack of receptions in college were more of a scheme issue than a lack of ability, Foreman might actually be underrated.

Bolstering the optimism for Foreman and McNichols are two later-round prospects from previous years with scores above 80: Jay Ajayi and Latavius Murray. Both have proven successful, and though Murray gets a lot of hate for his mediocre efficiency, I’ll take a RB10 and RB13 season from a late-third-round rookie any day of the week.7


Players being selected in the fourth round generally don’t have a great track record of success, but there are reasons for optimism this year. Had Brian Hill been drafted to almost any other team he’d likely be getting drafted much higher. While there are scenarios where he could eventually find work, it’s very hard to see how he becomes relevant in 2017 without injuries to both Devonta Freeman and Tevin Coleman.8

Jordan Howard appears to have a stranglehold on the workhorse role in Chicago, however none of the backs behind him have impressed, so Tarik Cohen may be able to carve out a roleElijah McGuire isn’t a spectacular prospect and his path to playing time will require at least one injury, but it’s not unthinkable that he could see work as part of a committee with Bilal Powell. Finally, Elijah Hood is one of my favorite late round targets. I’ve already covered the reasons to like him , and his price is still dirt cheap.


The 2017 RB class does appear to be the best class out of the last five years. Fantasy drafters seem to be aware of this, but there are still some opportunities for RB value in rookie drafts.  I personally stick to a mostly Zero-RB philosophy in dynasty leagues,9 but you need to get RBs to fill your roster from somewhere, and this year’s rookie draft presents some cheap opportunities to do so.


Subscribe for a constant stream of league-beating articles available only with a Premium Pass.

  1. In large part because the army of professionals evaluating a player’s talent usually get it right, but also because teams are likely to provide opportunity to higher draft picks.  (back)
  2. ADP was taken from May 1 to May 31, except in the case of 2014 where the draft didn’t take place until early May, and 2017 where we don’t have ADP data from the end of the month yet. The cutoff was set at 20 players to avoid diving too deep into undrafted free agents and because ADP data past this point begins to get a lot less robust. Across the five years of the study, 80 percent of the players past an ADP of 20 were UDFAs. A few players who snuck into the 19th or 20th slot due to only being picked a handful of times were removed from the sample. I’m looking at you Terron Beckham.  (back)
  3. The end of the fourth round is an ADP of 48 and the average ADP of the 20th player in this analysis is 47.7 with a maximum or 49.1 and a minimum of 46.  (back)
  4. Darius Jackson checks in with a score of 75, but he’s also a very late draft pick from a smaller school.  (back)
  5. Team situation, health, and opportunity are clearly very important here. David Johnson would be good wherever he went, but he might not have broken out to the same extent on a team other than the Cardinals, or if Chris Johnson had stayed healthy.  (back)
  6. Even with the 0.03 Pro Day adjustment to his 40-time.  (back)
  7. There are some cautionary tales here too though with players in the 60-70 range. David Johnson (63) was a steal in the second round of 2015, but David Cobb (66) went just one pick after Johnson in rookie drafts and isn’t even currently on an NFL roster.  (back)
  8. I’m only targeting Hill in leagues with deeper benches where roster space isn’t at a premium and I can hold a guy for more than a year with no production. I don’t see him having much value in shallower leagues this year.  (back)
  9. In the leagues I own David Johnson though you’ll have to pry him from my cold, dead roster.  (back)
By John Lapinski | @FF_SkiBall | Archive

Comments   Add comment