Writers of Pro Football Prospectus 2008

25 Apr 2018

FiveThirtyEight: Barkley Isn't the Hero Your Team Needs

At FiveThirtyEight, we looked at the conundrum of drafting Penn State running back Saquon Barkley in the top five, especially if you're the Browns or Giants. While workhorse backs have been romanticized for decades of NFL history, the league has shifted away from them in favor of more passing and a by-committee approach to the position. We also have not seen first-round backs outperform those taken in later rounds. The rushing DVOA since 2002 is the second lowest in the first round, and there's essentially no difference in yards per carry among the seven rounds. Running backs truly can be found anywhere, and Super Bowl-winning teams this century have also proven that.

Posted by: Scott Kacsmar on 25 Apr 2018

24 comments, Last at 03 May 2018, 9:55pm by The Ninjalectual


by Aaron Brooks Go... :: Wed, 04/25/2018 - 2:50pm

And Barkley has drawn comparisons to Hall of Famer Barry Sanders, who can serve as Exhibit A for the inability of a back to carry a team alone — despite his 10 great seasons for the Lions, the team won just one playoff game.

On the other hand, Sanders won more playoff games in his career than the Lions won in the other 45 years of the Ford ownership.

by JoeyHarringtonsPiano :: Thu, 04/26/2018 - 1:09pm

Yea, I don't think Sanders is a good example, either. Look at the 10 years before, and the 10 years after Sanders. Utter garbage for the most part (the only non-garbage years were the brief career of Billy Sims). Sanders pulled up the franchise from cellar-dwellers to a consistent playoff team. Imagine if he was drafted by a well-run franchise.

That being said, I don't totally disagree with the premise of the article in today's NFL.

by Aaron Brooks Go... :: Wed, 04/25/2018 - 2:53pm

Incidentally, shouldn't 538 have used average DYAR, not DVOA, for their comparison?

You can't average a sum of per-carry metrics like that. Summed DYAR isn't quite correct either, but it's a lot closer.

by The Ninjalectual :: Wed, 04/25/2018 - 3:37pm

I don't think they make a very good argument here. All they do is point out the risks, and it's true that for every AP or LT, there are dozens of Trent Richardsons and Curtis Enises. But they don't even try to examine whether it's actually worth it drafting successes like Ezekiel Elliot or LaDanian Tomlinson in the top 5. I think there's a strong argument in favor of picking a back that high, in the right circumstances. Barkley is very likely worth that high pick, and the main question is whether Cleveland or NY are teams prepared to get the most out of him. Honestly, I think Cleveland might be.

by Steve in WI :: Wed, 04/25/2018 - 4:03pm

I think in general we do so much hand-wringing over where it's acceptable to pick a particular guy or a particular position that we sometimes lose sight of the simple fact that if your first round pick is great, it was probably a good idea to pick him (and if he's bad, no amount of justification for why it made perfect sense to pick him is going to mitigate that). Drafting players is certainly not easy, but in the first round at least it really is kind of simple that way.

Regarding this draft in particular, you have to consider that Cleveland has 2 picks in the top 5 and if they took Barkley at 4 after taking a QB at 1, that's quite different than a team with 1 first-round pick taking Barkley at 4. Plus you could argue that one of the best things for a rookie QB is a very good running game.

Also, as a bit of an aside, I read articles like this stating that the Giants would be crazy not to take a QB given Manning's age, and I think...what if they honestly don't like any of the QBs in this draft? (Or they zero in on one and that guy gets drafted first by the Browns?) It seems to me like drafting a QB at the top of the draft because you need one, not because you actually believe in him, is a recipe for disaster.

by The Ninjalectual :: Wed, 04/25/2018 - 5:21pm

Yes, unfortunately that's about what it comes down to. A pick is good if the guy is successful, regardless of his position on the field or in the draft. Not the most useful analysis I've ever thought up :D

by jtr :: Wed, 04/25/2018 - 5:28pm

>Also, as a bit of an aside, I read articles like this stating that the Giants would be crazy not to take a QB given Manning's age, and I think...what if they honestly don't like any of the QBs in this draft? (Or they zero in on one and that guy gets drafted first by the Browns?) It seems to me like drafting a QB at the top of the draft because you need one, not because you actually believe in him, is a recipe for disaster.

IF all of that is true, then yeah, they shouldn't force themselves to pick a QB they don't like. I just find it unlikely that there aren't at least two quarterbacks that they like this year. They have the second overall pick in a draft that is generally agreed to have at least 3 good quarterback prospects. They're not likely to have a pick this high next year, and who knows what the QB class will look like.

by Lebo :: Fri, 04/27/2018 - 9:40am

Even if next year's QB class isn't great, surely NYG will be the destination of choice for any free agent QB, since they've got OBJ, Sterling Shepard, Evan Engram, and Saquon Barkely.

by Sixknots :: Thu, 04/26/2018 - 11:51am

>Plus you could argue that one of the best things for a rookie QB is a very good running game.

Yes, but the Browns have a veteran QB and a couple of good RBs already on the roster. If the Browns draft at #4, I'd take the best lineman available.

by mehllageman56 :: Wed, 04/25/2018 - 4:15pm

Haven't read the article yet, but I'm not a fan of their sports coverage. Don't trust ELO, certainly not as much as DYAR or DVOA. The article claiming the Giants drafting a new QB would be insulting to Eli Manning was absolutely terrible. Barnwell's much better at ESPN. Have to say, though, 538's political coverage is really good.

by dbostedo :: Wed, 04/25/2018 - 11:07pm

Go ahead and read it... it doesn't reference ELO and does reference DVOA.

by Mike B. In Va :: Thu, 04/26/2018 - 10:44am

Ok, so no-one's answering
Well can't you just let it ring a little longer?

by theslothook :: Wed, 04/25/2018 - 9:44pm

Idk why comments here seem to disagree about the premise. The NFL draft is all about opportunity costs. If AP and Von Miller are in the same draft, you can only take one. Which would it be? Position absolutely should factor into where a player is drafted and rbs are, in my eyes, far less valuable than pass rushers or great receivers.

by MC2 :: Thu, 04/26/2018 - 7:30am

I posted this on another thread, but it seems (mostly) relevant here, too:

AV is a good tool for comparing players at the same position, but I don't think it does as well comparing players from different positions, because it doesn't account for supply and demand. For example, I don't think a Curt Warner clone would be worth a Top 5 pick (let alone the #1 pick) in the modern NFL. He was a good-but-not-great runner, who contributed almost nothing as a receiver. RBs like that have very little value nowadays. I think for a RB to be worth a Top 5 pick in today's game, he has to either be a HOF-caliber runner, or else a very good runner with excellent receiving skills, like Roger Craig. For what it's worth, Craig had 65 AV in his first 5 years, and 115 for his career.

If Barkley were a lock to be as good as Craig, then yes, he'd be worth a very high pick. But he's not. He's not a lock to even be as good as Warner. Remember, back when Reggie Bush came out, most pundits were even higher on him than they are on Barkley now. Bush ended up with 39 AV in his first five years (72 career). I'd say Barkley has about a 75% chance of being as good as Warner, and about a 25% chance of being as good as Craig. That's just not enough to warrant a Top 5, or even Top 10 pick, as I see it. Is he worth a Top 20 pick? Maybe, depending on the team's needs and who else is available at that point.

by Aaron Brooks Go... :: Thu, 04/26/2018 - 8:31am

You're missing the point.

Most picks bust. The average AV for a given draft slot is *really low*.

Even at #1, on average, for every Peyton Manning, there are seven Ryan Leafs. Is Curt Warner worth more than Ryan Leaf?

by MC2 :: Thu, 04/26/2018 - 9:12am

I think you're missing the point. The top of the first round is pretty much the only place to find a Peyton Manning. So, if you want one, you're going to have to gamble, and hope you don't get a Ryan Leaf. It's worth the gamble, because a Peyton Manning is so valuable. But a Curt Warner, or even a Roger Craig, is not nearly so valuable, and can be found elsewhere. So, there's no need to gamble to obtain one.

Look at it this way. You're offered a chance to buy one of two lottery tickets, A or B. They cost the same, and each one gives you a 1% chance of winning. But the prize for winning with A is $100 (QB), while the prize for winning with B is $10 (RB). Which one are you going to buy?

by Aaron Brooks Go... :: Thu, 04/26/2018 - 10:17am

The difficulty with your analogy is that I suspect the odds are better for lottery B.

by jtr :: Thu, 04/26/2018 - 11:40am

I think you're right. Plus, I suspect that the running backs who don't turn into superstars at least turn into useful players at a higher rate than QB's. This article tries to sell Darren McFadden and Reggie Bush as failures, when both were good players with a few excellent seasons even if they never made it to sustained superstardom. Whereas top QB's who bust tend to bust all the way--in the past six drafts you have Johnny Manziel, EJ Manuel, Brandon Weeden, Jake Locker, Paxton Lynch and Blaine Gabbert as complete failures as first round QB's. Only Trent Richardson and David Wilson, who played well before a freak career-ending injury, have been failed to be productive players as first round tailbacks.

So for the lottery metaphor to be useful, you have to consider that the odds are different for the two lotteries, and that one lottery is more likely to offer a consolation prize if it doesn't hit than the other one. Which means, as we pretty much knew from the start, drafting the right player takes a complicated evaluation of risk, upside, and fit, rather than just grabbing a guy who plays the position considered most valuable.

by MC2 :: Thu, 04/26/2018 - 1:04pm

Well, of course there are a lot more first round busts at QB than RB in recent years. That's because there have been so many more QBs than RBs drafted in the first round. In fact, during the time frame you cite, there have been twice as many first round QBs as RBs -- 16 to 8. So, that's a 37.5% bust rate for QBs vs. a 25% bust rate for RBs. Given the small sample size, that could easily be random statistical noise.

But even if it's real, I think the difference would have to be a lot bigger to make up for the difference in value between a stud QB and a stud RB. If you doubt that difference, consider that a good-but-not-great QB like Kirk Cousins (a borderline "stud") got paid way more than any RB -- almost 4 times as much, on a per-year basis, and almost 7 times as much guaranteed money per year, as Devonta Freeman, the highest paid RB.

Finally, I never said that teams should just "grab any guy who plays QB" instead of Barkley. In the case of a guy with a very high bust potential, like Josh Allen, I think Barkley would be a better pick. But I would take any of the other "Big 5" QBs ahead of Barkley, and in the cases of Mayfield and Rosen, I don't think it's close.

by The Ninjalectual :: Thu, 05/03/2018 - 9:55pm

Geez, people STILL think Cousins is less than a great QB? What does the guy have to do?

by RobotBoy :: Sat, 04/28/2018 - 3:03am

The premise that 'most picks bust', especially when it comes to top five picks, seems dubious. Of course, it depends how you define 'bust' but the claim that 'for every Peyton Manning, there are seven Ryan Leafs' isn't a particularly useful. What does that claim say, really? That you are seven times more likely to draft the worst bust in NFL history than you are to draft the greatest QB in NFL history? You're talking about extreme outliers in both cases.
I've read somewhere (maybe here), that draft picks in the first half of the first round, on average, return decent value. Most aren't bust, in any case.

by ncuba :: Thu, 04/26/2018 - 12:11pm

You make a good point about the RB "consolation prize" since QBs either start and put up stats or don't.

But seems like the odds of the QB providing net benefit are better than with RB. I thought Barnwell's recent Barkley-to-NYG? rundown was very interesting in how it dealt with salary, basically rookie contracts are locked in according to position-in-draft, but that the salary benefit to team of any given rookie contract (i.e. relative to free agency) varies a ton according to position-on-the-field.

by Dan :: Thu, 04/26/2018 - 6:48pm

I suspect that things would look pretty similar at other positions, except maybe quarterback. e.g., Teams with a first round LG probably don't have a notably better rushing offense than teams with a 5th round left guard. Teams with a first round RCB probably don't have a notably better passing defense than teams with a 5th round RCB.

The main reasons are that 1) it's just one player out of the 11 on the field and 2) there is lots of filtering in terms of who gets playing time - 1st rounders who get playing time aren't much better than 5th rounders who get playing time, but 1st rounders are much more likely than 5th rounders to be good enough to earn playing time.

by Dan :: Thu, 04/26/2018 - 8:36pm

Taking Barkley at pick 2 is still a mistake. Giants would've been better off trading down with the Jets for the deal that Indy took. The only way it makes sense to stay put instead of taking that deal, IMO, is if you're taking a QB.