NFL

Fantasy Running Back Value, Part Four: Predictability

LeSean McCoy's been one of the most consistent running backs over the last two seasons, despite his injuries in 2012.

Fantasy owners who drafted Ray Rice, Doug Martin or Arian Foster this season in the first round will be more prone to shy away from the running back position next year. It’s how it works in our fake sport – we let personal experience change the way we select our future teams.

This isn’t necessarily a bad thing. For instance, a buddy of mine seems to always pick breakout wide receivers late in the draft. Always. He’s like the new-age Pittsburgh Steelers of fantasy football. To him, selecting Megatron in Round 1 doesn’t always yield the best results given his surplus of stud receivers once the season hits the midway point.

But in general, anecdotal evidence can – and often will – decrease your chance for fantasy success. Just because your league champion drafted a quarterback in the first round over the last two seasons doesn’t mean drafting a quarterback in the first round is the most effective draft strategy.

So far in this series, I’ve examined the supply and demand of fantasy football positions, as well as the general replaceability of the stud players at each of the main lineup slots. As a result, we’ve seen that running backs are in high demand given their supply, resulting in higher costs and value, and that stud backs are arguably less replaceable than even top-notch wide receivers. They certainly are more valuable than quarterbacks and tight ends in terms of replaceability.

However, the data provided hasn’t been on the individual players level. That is, I’ve essentially only looked at fantasy positions, not players.

To some, that may skew results. I can state that top running backs are more valuable than any other player at any other position in fantasy football, but knowing who that top running back is going to be is the challenge. Choosing the top running back is difficult.

I’m not denying the validity of that statement. Picking the “correct” running back – the one who will give you usable fantasy points week in and week out – isn’t a simple task. Ray Rice owners know what I’m talking about.

To make things worse, the running back position gets hurt more than the other slots in your fantasy football lineup. Quarterbacks seem to rarely go down, and wide receivers and tight ends appear to drop at a slower rate than backs. That’s the reality of the most bruising position in fantasy football.

This has translated into the perception that the running back position is the least predictable in the game. Running backs are tough to predict because of their yearly fluctuation, their short life span and their tendency to get injured. Running backs, to thousands of fantasy football owners, aren’t worth high draft picks because they’re simply too unpredictable to trust.

But that’s just not true.

The Misuse of Flop Rates

You may have read the title of this section and thought, “How is LeBron James going to make his way into a fantasy football article?” Here, however, I’m not speaking about on-the-court flopping. In this case, flopping is performing far below your original draft value.

Typically, websites and analysts will dub a player a flop if he was drafted as a starting-caliber player while finishing far outside of that initial draft spot (keep in mind, where a player is drafted will differ depending on where the data is being grabbed). Arian Foster, for instance, was drafted in the first round and finished as the 45th-best running back. Therefore, Arian Foster was a flop, as he didn’t even finish as a starter, or a top-24 running back.

This type of analysis makes general sense, and it’s not a poor way of showing how a player stacked up against his preliminary value. Though I dislike judging a fantasy season based on cumulative data, I can understand why people do it this way – it’s quick, easy and gives a general idea of how well a player performed versus how he should have performed.

Let’s run through a quick exercise to portray what I’m talking about here. From a high level, let’s consider a flop as any player who started the season as a starter (a QB1-12, a RB1-24, a WR1-24, a TE1-12) and finished outside of that spot. This is fairly basic, I know, but it’s a good start to a larger discussion.

If we follow this general guideline, the flopping results would yield the following for the 2013 season:

Quarterbacks: 4
Running Backs: 9
Wide Receivers: 7
Tight Ends: 3

Some may conclude that the running back position is the least reliable or predictable because they own the most flops. In fact, more than one-third of the top-24 running backs in 2013 ended up finishing outside of the top 24, allowing nine new running backs to be considered worthwhile or usable.

There’s one thing to keep in mind here: one flop at running back or receiver does not equate to one at quarterback or tight end. Why? Because you’re starting double the number of backs and receivers – you’re looking at the top 24 (in a 12-team league) versus the top 12.

If you look at flopping this way, you’d probably still conclude that the running back position was the least predictable in 2013. After all, double the amount of quarterbacks would be eight, and double the number of tight ends would be six, numbers less than that at running back. This is no different than what we traditionally see.

But this is where I start to hate this cumulative method of evaluation.

First, it assumes that players drop off in a linear, consistent fashion. You’re looking at rankings to judge whether or not a player flopped or not, versus actual player production. In other words, you’re assuming the gap between the first- and second-best running back is the same as the gap between the 15th- and 16th-ranked one. That’s rarely the case, especially when you consider my article on replaceability which showed just how elite a top running back can be.

Second, by looking at cumulative data, you’re not factoring in injuries very fairly. I’ve already noted that running backs get hurt more than any other position (I didn’t do a study for this – this is a general thought), so the position is already at a disadvantage when you look only at year-long numbers. But you have to look at what that particular running back did when he actually played, without just giving him a “zero” for every week that he was absent.

For instance, in 2012, LeSean McCoy finished the season as the 21st-ranked running back because he missed four games. But in those 12 games played, Shady finished in the weekly top 24 11 times. While 20 running backs scored more cumulative points than McCoy, only four actually had more weekly top 24 finishes.

Though McCoy technically still finished as a usable running back – an RB2 – the fact is that he was more than that in terms of adding value to a fantasy squad. To me, McCoy was easily a top-10 running back last season.

Le’Veon Bell’s another good example, though not as extreme. Bell finished this season as the 14th-best PPR back, which came after missing the first three games of the season. However, Bell still had 11 weekly top 24 finishes, which was a higher total than all but four runners.

All of this being said, you may look at season-long data and note that a player flops based on that data. However, given the examples above, it shouldn’t be evaluated only in that fashion. You need to factor in tiers, or how those players produced throughout the season. You need to factor in consistency.

A Different Way of Looking at Flopping

In my article on replaceability, I introduced a chart that looked at how many players at each position finished with what I called “usable” weeks. That is, how many instances did a player finish in the top 12 at quarterback, the top 24 at running back and receiver and the top 12 at tight end.

Instead of looking at individual players, I wanted to show trends in positional data. In the end, it was pretty clear that elite running backs finished as usable starters in fantasy more than elite wide receivers, tight ends and quarterbacks.

To remind you, below is the same chart I showed in the previous article, which looks at the usable week instances by position:

QuarterbackRunning BackWide ReceiverTight End
14 or more0100
13 or more0300
12 or more0401
11 or more2711
10 or more31322
9 or more41694
8 or more418135
7 or more823189
6 or more14252413
5 or more20353116
4 or more25434022
3 or more31535730
2 or more37617439
1 or more458111054

Note: These numbers reflect 15 games, as Week 17 is omitted due to it not being important in fantasy football.

While you could make the case that a consistent usable wide receiver is more scarce than any other position, the fact that the top running backs have so many more usable weeks could trump that value analysis. Think of it this way: If one wide receiver had, say, five instances where he was a top 24 wideout in a given week, while every other receiver in the league had just one, would you consider him a top draft selection? Probably not, as he’s only giving you a top-24 performance in 5 of his 15 relevant fantasy games.

I told everyone to just think of the chart from a positional standpoint, not to associate names to the table. Why? Well, I wanted to simply show that top running backs still matter a whole lot in fantasy football. They’re finishing in the top 24 more than any other position. And not all of them are – there’s a clear elite group at the running back position, and this happens every year.

Now it’s time to associate names to these numbers.

The question you should be asking is, “OK, I now know that there was just one running back who finished with 14 top-24 performances – was he a top running back at the beginning of the season?” Or you may think something like, “Of the four running backs who finished with 12 or more usable weeks, how many were actually top picks in the fantasy draft last August?”

The reason these types of questions are important is because they associate flop rates with production. Instead of simply looking at rank, we’re now looking at how many worthwhile weeks a player at the position had. Instead of Le’Veon Bell being the “14th-ranked” running back, he’s now in the “11 or more” tier. If it was 2012, instead of LeSean McCoy ranking as the 21st running back, he would be in the “11 or more” tier as well.

Using ADP data (I chose not to take individual rankings because those are the opinion of only, at most, a few people), I took each of the top-12 quarterbacks, top-24 running backs and receivers and top-12 tight ends and associated them into their correct tier. To put this another way, I replaced the numbers in the chart above only with the number of QB1s, RB1s, RB2s, WR1s, WR2s and TE1s that fit the criteria. In addition, instead of having “or more” at the end of each tier, I simply left the number, noting how many preseason start-worthy players finished with that many usable weeks.

Clearly the data used can shift the results here a little, but in general, ADP numbers are usually fairly similar from one another, as long as you’re not pulling from a random source.

The results are below:

QuarterbackRunning BackWide ReceiverTight End
140100
130200
120101
112110
101310
91362
80121
71121
64152
53421
40403
30131
20020
10000

As you can see, the number of quarterbacks that were ranked in the top 12 to start the season (which, if you remember, was a very clear-cut group), finished with at least five usable starts, and at most, 11 of them. At running back, there was one complete flop, David Wilson, who didn’t even contribute a single usable week (I apologize, Wilson owners). Darren McFadden gave us three, and the rest gave us between four and 14.

If you drafted an ADP top 24 receiver, chances are he gave you between five and nine usable weeks. However, five guys gave between two and three, and two (Antonio Brown and A.J. Green) gave you more than nine. Again, this is based on PPR leagues.

Lastly, the tight end position is a mess. This is what happens every year, and why I tend to just tell people to stream your tight ends (well, one of the reasons). I understand the allure with someone like Jimmy Graham, but that’s for another discussion.

It should be noted that I’ve done this exercise many times in the past, and the numbers are nearly identical every season. There’s a small variance at quarterback, which is natural given how easy it is to predict how good a passer will be, give or take a Peyton Manning-like outlier performance. And while some running backs completely flop, most of the top-24 backs that you select will still be, at the very least, usable for 4-10 weeks of the season. Moreover, of the elite ones, the majority are all drafted as RB1s or RB2s. While we may assume there are Knowshon Morenos every year, that’s not exactly the case.

This is no different than at the wide receiver position. In fact, an argument could be made that having an elite wide receiver is advantageous because of scarcity, but obtaining that wide receiver is a little more risky given general flop rates among top-24 players. To show why this is the case, notice the “floor” at the wide receiver position – five wideouts selected as WR1s and WR2s ended up failing to give you just four usable weeks. Five!

With this chart in particular, I wanted to point out the second overall pick in many fantasy drafts this season, Arian Foster. I noted earlier that he finished as the 45th-ranked running back this year (in PPR leagues), but so much of that had to do with the fact that he missed eight games. In fact, if you followed reports on him throughout the season, you probably only started him in his first six, as injuries began creeping up on him.

In those six contests, Foster finished with five usable weeks. Only 27 running backs had six or more, meaning there's a case to be made that Arian Foster wasn't even close to the 45th-best running back this season.

Now, I know the problem with just looking at start-worthy players (top-12 quarterbacks, top-24 running backs and receivers and top-12 tight ends) is that you’re not always just concerned about the entire group. Sometimes it’s more important to see how RB1s (top-12 running backs) flopped, and how top-half quarterbacks finished. Below is a chart displaying these instances:

QuarterbackRunning BackWide ReceiverTight End
140100
130200
120001
112000
100110
90241
80010
70100
62122
52211
40201
30020
20010
10000

Looking at this chart shows a very similar story. Essentially, we’re looking at the first two rounds of the draft for running backs, the first three or so for wide receivers, the first five for quarterbacks and about the first seven at tight end. I’ll get more into these market values (draft positions) with my next article though.

If you can take something away here, it’s that choosing an elite wide receiver in the draft can bring you great success, but it can also bring you pretty significant failure. That being said, this season especially, those top-notch wideouts in the chart were the ones who were drafted fairly early (e.g. Calvin Johnson and A.J. Green).

But the point to be driven home here is that preseason top running backs aren’t worthless. While we often point to players like Trent Richardson or C.J. Spiller, we fail to compare those players to guys who play other positions. What about Roddy White? What about Rob Gronkowski? Julio Jones? Tom Brady?

Again, I understand you used a first-round pick on Richardson and Spiller, but I’ll get into that in the next part of this series. For now, I’m simply noting that top players at other positions flop just as much as running backs do, average draft position (other than using it for consensus rankings) aside. And running backs, even if you think they flop, still can provide usable weeks.

What I now like to do is the same exercise, but this time looking at elite performances. While I defined "usable" as a player who finishes in the top 12 at quarterback, top 24 at running back and receiver and top 12 at tight end, an "elite" performance is one where these parameters are cut in half. I showed how this impacts things in my last article on replaceability, but for a refresher, below is the table with the number of elite performances by tier.

QuarterbackRunning BackWide ReceiverTight End
13 or more0100
12 or more0100
11 or more0100
10 or more0200
9 or more1201
8 or more1411
7 or more2651
6 or more21174
5 or more515117
4 or more1121168
3 or more17292816
2 or more24384522
1 or more33618036

Even more than the usable chart, this one shows us a clear advantage when having an elite running back. The question, again, we should ask is, "Of the elite running back instances, were those running backs highly sought after at the beginning of the season?"

QuarterbackRunning BackWide ReceiverTight End
130100
120000
110000
100100
91001
80200
70130
60132
52533
45330
32442
21242
11132
00310

We already knew that top running backs were the most valuable when looking at things from this angle, but what we didn't know was that the running backs who were elite this season were actually drafted as either an RB1 or an RB2. Now we do, thanks to guys like Jamaal Charles and Matt Forte.

When you compare running back to wide receiver, you see a little more separation with the elite players. In fact, if you drafted a wide receiver this year in the top 24, the most "elite" weeks he gave you was seven, the lowest of any of the main positions in fantasy.

This chart also shows us the true value of Peyton Manning this year. He gave nine elite performances at quarterback, four more than any other quarterback, which was the highest discrepancy among the positions. However, he also was a massive outlier in 2013, so keep that in mind when you evaluate his draft position next year.

The three running backs who were drafted in the top 24 and finished without a single elite performance in 2013 were Doug Martin, Lamar Miller and David Wilson - one RB1 and two RB2s. At wide receiver, the player was Dwayne Bowe. However, Reggie Wayne, Randall Cobb and Hakeem Nicks all finished with just one.

Let me go back to my example before, Arian Foster. While we're all up in arms about him being a bust (the 45th-ranked running back!), it should be noted that only 17 running backs finished with more elite weekly performances than Foster this season. And Foster, as I said, played in only eight (six if you paid attention to news) fantasy games.

What I'm attempting to show you, if you haven't noticed, is that running backs are not unpredictable. The reason we view them as unpredictable is because we use cumulative data to judge how good or bad a player was over a given season. Just because Arian Foster played half the year and finished as the 45th-ranked running back doesn't mean you should automatically assume Foster returned 45th-ranked running back value. He didn't - you weren't taking a zero in his spot when he wasn't in your lineup.

This, I think, is where people have things backwards. It's not so much the understanding that elite running backs are important, but the post-season evaluation done to judge whether or not a player was valuable in fantasy.

Bringing it Together

We’ve already seen that, in fantasy, running backs have a natural higher value due to supply and demand. And based on replaceability, we know that top running backs mean more to the game than any other players at any other positions.

Now we’ve seen that not only are those top backs traditionally top picks at the beginning of the season, but even if your early-round running back doesn’t finish as a top back, he’ll still give you usable weeks to work with. Injuries, in other words, are overrated.

I’ll continue to say this as I write this series of articles: The goal here is not to prove that you need to draft a running back with your early picks in next year’s draft. Rather, it’s to show you that the perception of running backs holding less value in fantasy football is off. People seem to think that they’re becoming more and more useless with the changes in today's NFL, when this is simply not the case.

In the next portion of this series, I’m going to take a look at where these top-12 quarterbacks, top-24 running backs and receivers and top-12 tight ends are being drafted. We know they’re elite players within their position once the season starts, but knowing where they fall in a traditional snake draft is actually more important than knowing where they fall in rankings.