image source: wwltv.com
BLUF (Bottom Line Up Front)
Accounting for opportunity, supporting talent, individual player’s talent, and draft stock, the data shows:
- Saquon Barkley is head of the class with a projection of several ppg higher than the number 2 ranked player, Sony Michel.
- There isn’t much separation between Ronald Jones II, Rashaad Penny, and Derrius Guice at #3 through #5.
- Nick Chubb (#8) and Kerryon Johnson (#10) are currently overated due to a lack of opportunity.
- The rookie RB field is deep this year. The top 6 are projected to be useful in fantasy lineups while the #7 through #9 have a legitimate chance to be useful in fantasy football.
It’s been a while since the 2018 NFL Draft ended. While it’s way too early to be sure which rookies will lock down a large portion of the touches, it’s never too early to start preparing for the fantasy draft. At this time of year, we are often bombarded with rankings that are mostly just opinions. Well, FFinsights says opinions oschminions; let the data lead the way. The more we can remove opinion and insert data, the more consistently accurate the projections will be. You can see how inaccurate opinions were last year by reading the following FFinsights article by Scott Reida: link.
I’ve developed data driven single season (redraft) rankings for the top 10 projected RBs in PPR and Standard scoring. What do I mean by “data driven”? Well, in these rankings, I’ve minimized the opinions and let the data predict the scoring for the season. The opinions are kept to a minimum and I’ll explain where I’ve used opinion so you know exactly what you’re getting. Another advantage to these rankings is that you can see, not only why they are ranked where they are, but also the gap between them rather than just a numbered list of players.
If you don’t care to read about the how and why, scroll down to the large print saying “TOP 10 ROOKIE RB RANKINGS (REDRAFT)”. For those who are interested, simply continue reading.
Why Use Reference Data?
People tend to be more familiar with the players drafted higher in the NFL draft. This tends to form a bias when it comes to fantasy drafts. People tend to consider only talent and how much play time the player will get. While college players are analyzed and graded on metrics such as size and talent, these are not necessarily the biggest contributors to short term fantasy success. It is often the new circumstances surrounding the player, such as opportunity and supporting talent, that have a bigger impact. That’s why this ranking focuses on these two key components. Both opportunity and supporting talent are encompassed by using data from the NFL team’s previous season. Then, the estimated work load (snap percentage) accounts for the player’s talent, opportunity, and draft stock all in one because all three play a role in leading to increased snaps on the field.
This is, obviously, not a perfect science because we know things surrounding the team have changed since the previous season’s data. But, hey! At least it’s in the vicinity of science! In summary, using the previous season’s data with little to no tweaks is a much better way of projecting than mostly opinion.
About the reference data
The 2017 data used as a reference for each ranked rookie RB shows the following:
• Head coaches. This is an indicator that even the change in coaches that occurred for some teams has been accounted for.
• All the RBs who received snaps on that team last season.
• Snaps per game average that each played.
• Utility percent. This is the percentage of plays where they touched the ball or were the target of a pass while in the game.
• Points per snap in PPR scoring format.
• Points per snap in Standard scoring format.
• The yellow line and digits are the average of that given statistical category
Other 2017 reference data
Average total offensive snaps per game for each team (excluding special teams plays) are shown here:
Assumptions
These rankings operate under an estimation of the workload each RB will get, namely the percentage of snaps. There is also a small opinion injected into a figure used in calculation for a few of the RBs listed. Again, I’ll tell you where this occurs within the player explanations below.
Methodology and Calculations
First, I estimate the percentage of snaps the rookie will receive this year. Then, I use the average number of snaps for their new NFL team’s offense during last year to project the number of snaps the rookie will play. For you math types, the formula looks like this:
(% of est snaps) x (team snaps / gm avg '17) = proj snaps '18
Next, the projected player snaps 2018 are multiplied by the points per snap based on the reference data from players performing similar roles last year on their NFL team. Slight adjustments and assumptions are used here for perceived talent and having to determine comparisons:
(proj snaps '18) x (RB pts / snap '17) = proj ppg '18
After reading all of that, you finally get to the rankings. These rankings are based on single season projections and thus, are for redraft leagues. Although standard scoring is projected as well, they are ordered by PPR ranking.
Key
Points per reception (PPR)
Standard scoring (Std)
Points per game (ppg)
TOP 10 ROOKIE RB RANKINGS (REDRAFT)
1. Saquon Barkley, RB, Penn State, Rd 1 pick 2 (2overall), NY Giants
Saquon Barkley is the easiest of any rookie to rank. He is an extremely talented RB expected to be a 3-down starter from day one. I estimate him at around 85% of the Giants’ 68 average snaps which puts him at 57.8 snaps per game. For perspective, Zeke Elliott got 59 snaps per game last year. The Giants’ offense will be improved this year under Pat Shurmur as the new head coach along with the addition of Barkley. That and Barkley’s talent led me to raise the per snap scoring to closer resemble Orleans Darkwa’s, the highest RB numbers from last year. That gave 0.4 and 0.34 points for PPR and Std respectively. But even with those, likely conservative, numbers– he still came out on top of rookie backs by far with projections of over 23 ppg in PPR and almost 19.7 in standard scoring. He’s a locked-in RB1.
Reference Data:
2. Sony Michel, RB, Georgia, Rd. 1 Pick 31 (31 overall), New England
Michel lands on a great, although crowded, offense in New England. But, this is the same offense that had 70 RB snaps per game last year. It also saw Dion Lewis get 25 snaps per game leading to 212 touches last year and he’s no longer there. Plus, New England just spent a 1st round pick on him. There must have been something they really liked there. I assigned Sony a 55% share of snaps good for about 38.5 of the 70 RB snaps that NE had per game. For PPR, inserting the average of a points per snap between James White and Dion Lewis gives 0.42 per snap. Knowing that Dion Lewis put up 0.41 points per snap in standard scoring where his receptions don’t gain him anything, it’s fair to say that Michel will score at least 0.31 per snap in Standard scoring. That projects to about 16.2 ppg (PPR) and 11.7 ppg (Std), good for solid RB2 usage.
Reference Data:
3. Ronald Jones II, RB, USC, Rd. 2 Pick 6 (38 overall) Tampa Bay
With Doug Martin gone, Jones will likely be the starting RB locking down the early-down RB job in TB. Sims, while not the runner Jones is, is quite adept at receiving and TB seems to like using him as a 3rd down back. Assigning a usage of roughly 67% of TB’s 61 RB snaps per game last year equals 40.9 snaps per game. Using a Peyton Barber’s points per snap of 0.34 (PPR) last year and 0.27 (Standard), the data shows Ronald Jones scoring about 13.9 and 11 ppg in PPR and Standard respectively. Those are good enough for an RB2.
Reference Data:
4. Rashaad Penny, RB, San Diego State, Rd. 1 pick 27 (27 overall), Seattle
Penny lands in a good spot for playing time but a bad spot in terms of the team’s recent production. Seattle’s struggles may be a blessing and a curse to Penny as he will likely get more playing time but on a team with an offensive line that could use some improvement. Seattle RBs last year had an average of 0.23 points per snap in PPR and 0.16 in Standard. Sure, you could say some of that may be attributed to play calling, but it’s still pretty bad. Even if it were due to play calling, do you think the same coaching staff will change philosophies that much in one year? Not likely. I’ve estimated that by the start of the season Penny’s snap percentage will be 70% of Seattle’s snaps. Using the reference data of 66 average snaps per game from last year, that puts Penny at 46.2 snaps per game. That may be optimistic but he is the most talented back on the team and has high draft stock invested in him. However, applying the highest points per snap scored by a non-3rd down type of RB on Seattle last year (Mike Davis at 0.3 on a small sample size) into the equation, the data projects his scoring is limited to 13.9 (PPR) and 9.7 (Std) ppg. Not bad, but it’s poor output for such a high percentage of snaps. Consider him a low end RB2.
Reference Data:
5. Derrius Guice, RB, LSU, Rd. 2 Pick 27 (59 overall), Washington
Guice slipped in the draft due to off-field question marks, but his talent is not a question mark at all. Nonetheless, we are looking at data driven stats, not just opinions of talent. Guice was drafted by a Washington team where RBs not named Chris Thompson did not perform consistently well last year and even though Perine had some decent games, I don’t think he will impede Guice’s playtime much. Based on a projection of 60% of the snaps for 38.4 snaps per game and using last year’s Washington RB average points per snap along with the NFL average of RB snaps per game for the team (64), the data projects Guice will score 13.8 (PPR) and 10.8 (Std). Consider him an RB2 even though he was the 7th RB off the board.
Reference Data:
6. Royce Freeman, RB, Oregon, Rd. 3 pick 7 (71 overall), Denver
Freeman could be a sleeper of sorts after being the 8th RB drafted this year. The Oregon product will be joining the Denver Broncos where I believe he’ll take the majority of work from Devontae Booker sooner than later. As you can see, Denver RBs had a very low variance in points per snap last year so inserting the average of the Anderson, Booker, and Charles from last year should be fairly accurate. So, if Freeman gets about 60% of the 62 average snaps Denver had last year, he is projected to score about 11.9 (PPR) and 8.1 (Std). He is a solid Flex player.
Reference Data:
7. Nyheim Hines, RB, Rd. 4 Pick 4 (104 overall), Indianapolis
Hines is a tricky one to project because he’s expected to get some slot receiving duties and RB snaps. Although an extremely fast runner (4.38 40 yd time), Hines seems to be destined for more of a change-of-pace RB due to his 5’8, 198 lb. frame. He will make some big plays, but how often and how productive will he be otherwise? As mentioned for Jordan Wilkins, I’ve assigned the 2017 NFL average of 64 RB snaps per game and increased their 2017 team RB points per snap by 20% to give 0.34 (PPR) and 0.26 (Std). After plugging in an estimated 20% of the RB snaps this year and around 20 plays at slot receiver per game, the data projects around 8.0 (PPR) and 5.8 (Std).
Reference Data:
8. Nick Chubb, Georgia, Rd. 2 Pick 3 (35 overall), Cleveland
Chubb is a great runner. Unfortunately, he lands in CLE where he’ll have stiff competition from Carlos Hyde and Duke Johnson for snaps. A projected 37% of the snaps in CLE last year would have garnered 21.8 snaps per game. I’m expecting more points per snap in CLE this year so I’ve increased the team average for PPR by 10% and Std scoring by 20% since Chubb is more likely to excel in Standard scoring with Duke Johnson having locked down 3rd down duties. Given those slight bumps to the snaps per game, the data projects Chubb at about 7.9 in PPR and 5.7 in Standard scoring per game. Something the data doesn’t show is that if the oft-injured Hyde goes down, Chubb could become a high RB2 and that is something to keep in mind when drafting your fantasy team.
Reference Data:
9. Jordan Wilkins, RB, Rd. 5 Pick 32 (169 overall), Indianapolis
I believe Jordan Wilkins winds up in a good situation in Indianapolis as Gore left town and took his 35 snaps per game along with him. Although, the early reports from Indy’s local papers are that Marlon Mack maintains his starting role, the Colts still spent a 4th and 5th round pick on the position. Of those two RBs drafted, it is the 5th round pick, Jordan Wilkins (6’1, 216 lb), who’s size is more likely to pressure Mack for the majority of snaps while the other pick, Hines gets moved around and used more for receiving. For now, I project Mack to keep a majority of the snaps so I estimate 35% of snaps go to Wilkins. With the head coaching change and improvements Indy made by drafting a tremendous blocker in guard Quenton Nelson and the potential of Andrew Luck coming back, I’ve assigned the 2017 NFL average of 64 RB snaps per game and increased their 2017 team RB points per snap by 20% to give 0.34 (PPR) and 0.26 (Std). Using those numbers, the data projects Wilkins at about 7.5 ppg (PPR) and 5.9 (Std). Not starting numbers in most any league.
Reference Data:
10. Kerryon Johnson, RB, Rd. 2 Pick 11 (43 overall), Detroit
Johnson is probably the most difficult rookie RB to place. He has the skills to be a featured back, but he will have to push Ameer Abdullah and there is also Theo Riddick playing on obvious passing downs. Plus, he’ll have LeGarrette Blount most likely vulturing short-yardage and TDs. So, I understand that Abdullah hasn’t really seized the opportunity when given the chance, but, there’s just too much uncertainty surrounding the Lions’ backfield. If Abdullah ends up not being on the Lions roster, then Kerryon gets a boost to perhaps weak RB2 territory. But, as for now, I’m estimating a very conservative 30% of the Lions’ 64 snaps per game. Using the combined average points per snap from Theo Riddick and Ameer Abdullah as the average for Johnson, the data projects him with a mere 6.3 (PPR) and 4.6 (Std). Although, he is worth drafting as a backup with upside, those numbers aren’t someone to count on in your lineup. That makes him, perhaps, the biggest surprise of these rankings.
Reference Data:
BONUS RANKING #11
11. Kalen Ballage, RB, Rd. 4 Pick 31 (131 overall), Miami
Don’t you love bonuses? This is like one of those Easter eggs during the credits of a movie. Thanks for reading. Here’s the bonus ranking. Ballage has a chance to push Kenyon Drake, but I believe Drake has earned at least a starter’s share with his performance last year. That said, Drake won’t play every snap. I can see Ballage actually taking about 20% of the snaps especially since he shows promise in the passing game. Inserting the estimated 20% of Miami’s 55 RB snaps still only nets him around 11 snaps which as you can imagine doesn’t do much in the way of average ppg. He’s not likely to be consistent in scoring as the data puts his average at under 4 ppg (PPR) and 3 ppg (Std). But, if you’re drafting him in a redraft league, it’s as a handcuff to Drake anyway.
Reference Data:
The dashboard used to populate the reference data images is below if you would like to tinker with it.
