Finding value at Running Back in 2017 PPR and standard fantasy leagues
According to both MathBox™ and the Fantasy Outliers team, these Running Backs could return higher value than ADP or Expert Consensus Rankings.

Listen to our draft prep monster, all-you-need-to-know podcast at Fantasy Outliers Podcast Ep. 5 — The Big 2017 Predictions Episode!
It refers to our draft kit, which is available free here. (desktop friendly)
As a fantasy community, we are not so good at drafting Running Backs
The top end of the running back position can literally make or break your fantasy season. The thing is, it’s a risky position to pick. Even for us fantasy geeks. Yes, we’re pretty bad at it.
In Fantasy Outliers’ analysis of competitive fantasy leagues, we found that the value you typically get from the 1st running back taken to the 10th running back taken was about even. The 3rd and 4th picks have historically done well, but other than that, the 1st, 2nd, and 4th to 10th picks have all done about the same. At other positions, the top pick historically has returned the highest value, and then it goes down from there with each subsequent pick. Therefore, we can only conclude that…
…we are just not very good at drafting running backs.

MathBox™ can give you some measured confidence when drafting Running Backs in PPR
MathBox™ consists of multiple Gradient Boosting Regressor machine learning models that help us get a better understanding of the Running Back position, in terms of projected Points Per Opportunity, Opportunities Per Game, Week-to-Week Variability, and Availability Risk. To learn more about the models, how we built them, and their projected 2017 error rates, check out this article.
Now, let’s compare MathBox™’s projections with Expert Consensus Rankings (ECR). This is where we will really find some sneaky value in fantasy drafts.
We took the Points Per Opportunity and Opportunities Per Game predictions and extracted each player’s projected value versus the top 20 running backs. We call this the Value Vs. Tier 1 and 2 Players metric in the chart below:

Just to be clear, David Johnson’s +5.73 value means that he’s projected to net a fantasy team +5.73 points/week on average against the top 20 running backs this year (in PPR leagues).
Roundtable discussion — How we think you can find value at running back in PPR leagues
In order to get you a breadth of opinions, we’ll go roundtable across the Fantasy Outliers Team to share each of our key insights with you.
Dave Seal: The top end of MathBox™’s projections for running back are very much in line with ECR. In fact, 6 out of the top 7 players are rated in exactly the same order. That stood out to me as somewhat of a validation of MathBox™, because it doesn’t listen to podcasts and somehow came away with the same take as the masses. To add some depth to the rankings, David Johnson, Le’Veon Bell, Devonta Freeman are at approximately +6, +4, and +3 points/week on average, and the value projections fall off steeply from there.
There are differences from ECR. Bilal Powell is one of them. He sticks out like a sore thumb at #4 MathBox™ vs. ECR at #13. On one side, the fact that ECR is even at #13 suggests that people agree that Powell is going to get the share of the Jets’ carries this year. And his PPO projection is pretty high at 0.88 PPO, that’s +0.1 above the average for running back. On the other hand, the Jets are terrible, and it’s probably going to be a running-back-by-committee this year. Sorry, MathBox™, I’m not really buying this one.
C.J. Anderson is also popping at #8 MathBox™ vs. #22 in ECR. Honestly, I don’t see how that happens, either.
Mark Ingram is #9 MathBox™ and #17 ECR at running back, #33 ECR overall. MathBox™ is clearly favoring Ingram over Adrian Peterson who is at #21 in our ranks. Those guys have been neck-and-neck on a lot of people’s draft boards, and MathBox™ is saying, we should be favoring Ingram. I’ll believe that one.
Joe Mixon is MathBox™’s highest rated rookie. This seems like kind of a qualifying thing to see him this high, given all the preseason hype around his potential. MathBox™ likes him better than Todd Gurley, who’s at #9 in ECR but #23 MathBox™. He’s projected to score just over 12 points/game. Apparently, MathBox™ just isn’t too hot on the Rams this year.
Late round picks I like are: James White (why not get a piece of the Patriots?); Theo Riddick (he’s at #16 MathBox™ vs. #28 ECR and has a pretty good PPO); Duke Johnson; and Travis Cadet.
David Harbarger: The Isaiah Crowell rank was surprising one to me. He’s #10 in MathBox’s ranks with a #15 ECR. Sure, he’s going to get a lot of work, but is Cleveland going to be good enough to get him the 20 OPG he’s projected to have? Also, some people think #15 ECR is even too high for him, so I’d have to think long and hard about pulling the trigger on him even at the 15th RB spot. Also, you have Marshawn Lynch hanging around at that same point in the draft, and MathBox™ is high on him, too. So given the choice, I’d take someone like Lynch in a heartbeat.
CJ Anderson’s #8 MathBox™ rating stands out vs. the 22nd running back in ECR (and #57 ECR overall). I’m circling him in the 6th round. Maybe Denver gets their offense together this year and he lives up to the OPG MathBox™ has in the projection?
Tevin Coleman is MathBox™’s #14th best Running Back, compared to #27 ECR. The model is basically saying that both of Atlanta’s running backs this year are going to be useful in fantasy leagues. He’s not a role breaker, but maybe someone who’s not going to hurt you as much as other guys taken around his time in the draft.
MathBox™ is down on pretty much all of the rookies and second year running backs this year. We’re not sure if it’s a bias in the models due to limited sample size or maybe the algorithms know something we don’t. Jordan Howard is projected as the 6th running back in terms of ECR, but MathBox™ has him at #24. It could be model bias, but let’s face it, the Bears are going to be terrible this year. Also, Leonard Fournette #27 MB vs #12 ECR and Christian McCaffrey is #18 in ECR and #47 MB.
Also, MathBox™ is down on Ezekiel Elliott this year. After adjusting MathBox™’s OPG to 18–20 range, I see him as more of a replacement level back once he comes back from his 6-game suspension. So the question is: is it worth taking a half of a replacement level guy as the 10th RB off the board? I would think not.
I’m not buying the Bilal Powell and Isaiah Crowell projections. Two terrible running backs on two terrible teams. What more is there to say?
I’ll be targeting CJ Anderson and Mark Ingram in the early-middle rounds.

Chris Seal: So there’s some hatin’ on the Bilal Powell and Isaiah Crowell MathBox™ picks. I totally hear that. But let’s just play Devil’s Advocate. Who knows, maybe it’s one of those Moneyball situations where the model sees something that we‘re biased towards (like: Jets=BAD, Browns=BAD). The Points Per Opportunity model for veterans goes back to 1994 and the Opportunities Per Game model starts in 2009 with a 3-years trail. There are a lot of individual and team-wide statistics that go into MathBox™. It knows the Browns and Jets are bad teams, yet it still projected them high. I’ll admit, I am hesitant, but Bilal Powell did well last year when he had opportunities. If he falls a little bit and you draft him as the 15th running back in the 4th round, what if he ends up being a top 10 back? It just might be a way to get sneaky points.
Other than that, I pretty much agree with Dave and Dave. Follow what both ECR and MathBox™ say early on, then Marshawn Lynch, CJ Anderson, and Mark Ingram all seem like pretty solid picks at about the 3rd or 4th round. Maaybe Crowell, too, if he slides down in the draft a bit.
I also think Joe Mixon will be a star at some point. Just due to potential alone, he might be worth taking a shot on in Round 7 or 8. Not many three-headed monsters produce positively-valued fantasy running backs, but the Bengals are in win-now mode — and you never know how the season will unfold.
Also, word to the wise, MathBox™ had Theo Riddick, Doug Martin, and Matt Forte all at higher than normal Availability Risk this year. Just sayin’.
Okay, here’s my favorite graph. It sums up a lot of what the Dave’s are saying above. Basically, if you trust MathBox™, you should be targeting the guys on the top-left-hand portion of this graph. That’s because MathBox™ has them rated higher than ECR. If they’re green and big, they’re above replacement, if they’re gray they’re at replacement, and if they’re red and small, they’re below replacement. Simple right? Thank goodness I have the Dave’s to keep me in check. It takes a bit to wrap one’s head around, but it’s what I’m going to use when drafting — so I figured you‘d want to see it too:

To get access to a unique draft kit that will give you a competitive advantage, go to our projections here. (desktop friendly)
Finding Running Back draft value in Standard Scoring leagues
You get the hang of it by now. Here are MathBox’s running back projection ranks versus ECR in standard scoring leagues:

You can connect with us by:
- Listening to our podcast, reading our articles, signing up for our newsletter, following us on Facebook, Twitter, Instagram, or emailing us at: podcast@fantasyoutliers.com

