Daily Fantasy Playbook: Running Back Strategy

Aug 08, 2014
Daily Fantasy Playbook: Running Back Strategy

One of the best things about playing daily fantasy football is that filling the running back position is less stressful. Unlike re-draft leagues, you don’t have to worry about spending a top pick on a back only to see him get hurt (Doug Martin), be on the wrong side of a committee (C.J.Spiller), or just plain suck (Trent Richardson). Well, your running backs can still plain suck, but not if this article has anything to do with it.

Read on to see which statistic has a surprisingly strong correlation to running back success, whether goal line work is actually important, and how you may have been thinking about upside all wrong.

 

Running Back Salary Allocation

In last week’s quarterback strategy guide, we calculated positional volatility and found that in terms of fantasy points per game (FP/G), RBs are more volatile than QBs, but less volatile than WRs and TEs. The positional coefficients of variation (CV; standard deviation divided by mean, where lower means less volatile) were:

  • QB: 45%
  • RB: 55%
  • WR: 65%
  • TE: 67%

Since we have already established that in cash games (H2Hs, 50/50s, etc.) it’s generally best to pay up for a QB because they are the least volatile, let’s focus on RB, WR, and TE. We will again calculate an approximation of each player’s floor (assuming a 20 point projection) by multiplying the positional CV by the projection, and then subtracting the result from the projection:

  • RB: 20 - (0.55*20) = 9
  • WR: 20 - (0.65*20) = 7
  • TE: 20 - (0.67*20) = 6.6

In cash games, you should generally allocate more of your cap to RB than WR/TE because there is less inherent variance at RB, keeping your floor higher.

We then calculate an approximate ceiling in the same manner as above, except this time we add instead of subtract:

  • RB: 20 + (0.55*20) = 31
  • WR: 20 + (0.65*20) = 33
  • TE: 20 + (0.67*20) = 33.4

In GPPs (larger fields), finding cheaper value at RB becomes more important, because we want to pay up for high variance WRs and TEs (think Calvin Johnson and Jimmy Graham).

Remember, this salary allocation method should only come into play when you are deciding between equal values at different positions. A great value at any position should take precedence over any allocation strategy. Below, I will dive into how to actually identify value at RB.

Also, note that the example above used 20 FP as a projection, but in reality you will rarely have a TE projected as high as a RB or WR. The 20 in the example should be treated as a number indicating value (like points/$ or $/point). WR and TE have the most volatility, meaning they are more likely to greatly overperform or underperform their mean. RBs generally don’t overperform or underperform their mean by as much, which is advantageous in cash games where you need a high floor, but not so much in GPPs, where you need to maximize positive variance.

 

What Are The Most Relevant Statistics When Analyzing Running Backs?

The table below shows how various statistics correlated to RB FP/G using two different scoring systems: 0.5 PPR and 1.0 PPR.

A quick refresher on correlation: The scale is -1 to 1. Positive values indicate variables move in the same direction, while negative values indicate variables move in opposite directions. An absolute value of 1 means perfect correlation. Absolute values above 0.7 are considered strong, those between 0.3 and 0.7 are moderate, and those under 0.3 are weak.

Statistical Correlations to FP/G, 2013
  CORRELATION TO 1 POINT PPR FP/G CORRELATION TO 0.5 POINT PPR FP/G

Total Yards/G

0.94

0.96

Total TD/G

0.83

0.86

Snaps/G

0.82

0.83

Touch/G

0.78

0.82

RZ TD/G

0.75

0.77

RZ Touches/G

0.74

0.77

Rush Yards/G

0.73

0.79

Rec Yards/G

0.71

0.63

Rec/G

0.68

0.58

Targ/G

0.66

0.58

Rush TD/G

0.66

0.72

Car/G

0.62

0.69

Rec TD/G

0.59

0.54

Yards/Touch

0.41

0.36

Team Points Scored

0.37

0.37

Yards/Car

0.36

0.37

Yards/Rec

0.24

0.24

 

Not surprisingly, total yardage and total TDs show the strongest positive correlation to fantasy points per game. The fact that we need yardage and TDs from our RBs is no secret, so how do we get an edge? It is the next few strongest correlations that are essential to grasp in order to develop the correct running back strategy.

Snaps per game correlate better to RB FP/G than any other statistic that is independent of FP/G. In fact, snaps per game correlate to FP/G is just as strongly as TDs/G! This is significant because unlike yardage and TDs, snaps do not get factored into a player’s salary. Players who get a lot of snaps but are coming off rough games may see their salary depressed and be undervalued. By the same token, players who are coming off great performances but who don’t play as many snaps may see their salary increase and be overvalued.

Snaps are also a very valuable statistic to pay attention to because while usage and efficiency are hard to predict and vary quite a bit, a player’s role within the offense-and therefore his snap count- tends to remain more stable from game to game.

Note that yards per touch, yards per carry, and yards per reception correlate only half as much to FP/G as snaps per game and touches per game do. Per game total yardage correlates most strongly with FP/G, and total yardage is amassed by getting touches. For RBs, it is clear that we should prioritize volume over efficiency.

A great example of volume trumping efficiency in 2013 is Le’Veon Bell. Bell was not an efficient runner, averaging only 3.5 yards per carry for the season. However, he played 53.2 snaps per game, good for fourth-best among all RBs. Furthermore, he ranked second among RBs with a 75% share of his team’s total rushing attempts, and seventh among RBs with 14% of his team’s total passing targets. Sure enough, Bell finished the season as a top-10 RB in terms of FP/G.

Speaking of volume, notice how important red zone production is. As we will see in a bit, the red zone is where the majority of touchdowns are scored, and by not selecting a running back who gets looks in close, you are putting yourself at a big disadvantage. Going back to the Bell example, he also led the NFL in red zone touches per game. To properly select RBs in DFS, you have to look way deeper than yards per carry.

How does a team’s offense effect RB fantasy production? Since red zone production is important, a RB’s offense will definitely influence his overall production because the better offenses get into the red zone more often. However, the amount of points a running back’s team scored only has a weak to moderate correlation to his FP/G. Again,  volume is a priority. Playing in a good offense will only help if RB gets significant volume within that offense in a given game.

A site’s scoring system does have a slight effect on how much importance should be given to pass catching backs. In 0.5 PPR, rushing attempts (0.69 correlation) should be emphasized a bit more than receptions (0.58). In 1.0 PPR, receptions (0.68) become more important than rushing attempts (0.62). On a 0.5 PPR site like FanDuel, you will get a little bit more out of a one-dimensional back like Frank Gore, while on a 1.0 PPR site like DraftKings, receiving specialists like Danny Woodhead and Shane Vereen become a bit more valuable.

Also note that since volume is more important than efficiency for RBs, a cheap, less-skilled backup who replaces an injured starter can often still be very valuable. Just make sure that this backup is slated to play the majority of snaps, rather than be part of a committee approach.

 

How Can We Maximize Rushing Touchdowns in Our Lineups?

Now that we know we need a RB that will get a decent amount of volume, how do we figure which of those backs are most likely to score? Let’s take a look at where rushing touchdowns are most likely to occur.

Rushing by Field Position, 2013
FIELD POSITION % OF RUSH PLAYS CALLED RUSH TD/ATT

Own 1-10

47.1%

0.002

Own 1-20

43.7%

0.002

Own 21-50

38.8%

0.002

Opp 49-20

42.0%

0.010

Red Zone

45.8%

0.181

Opp 1-10

48.6%

0.289

Notice how a runner has virtually no chance (one percent to be exact) to score a touchdown outside of the red zone! This means that just as with QBs, we want to target RBs on teams with all-around good matchups. Value may be had when a RB has an apparent neutral matchup with the opposing run defense, provided his team’s passing offense and it’s defense have good matchups that can lead to good field position.

Goal line work is also essential. Check out how much more likely rushing touchdowns are the closer you get to the end zone.

Rushing TD Success Rate by Yard Line, 2004-2013
YARD LINE RUSH TD SUCCESS RATE (TD/ATT)

1

54.6%

2

37.9%

3

33.0%

4

26.4%

5

18.9%

6

17.4%

7

13.9%

8

12.4%

9

7.8%

10

7.2%

As you can see, it is crucial for a running back to get those high-percentage chances inside the 4 yard line. While we cannot predict which yard line(s) a team will run goal line plays from each game, we do need to make sure that there’s a good chance the RB we’ve selected is on the field.

 

Redefining Upside

Many DFS players tend to define upside mostly as big play ability, which is wrong. To understand why, take a look at the likelihood of big plays compared to all rushing TDs, regardless of length.

Per Game Averages of Various Types of Rushing Plays, 2013
  W/O TD, PER GAME WITH TD, PER GAME

All Rushing Touchdowns

N/A

0.865

20+ Yd Rush

0.648

0.105

30+ Yd Rush

0.281

0.082

40+ Yd Rush

0.137

0.051

50+ Yd Rush

0.072

0.031

60+ Yd Rush

0.031

0.018

70+ Yd Rush

0.012

0.010

80+ Yd Rush

0.008

0.008

90+ Yd Rush

0.002

0.002

A rushing TD is much more likely to occur than a big play of any distance. Twenty-plus yard rushes have a decent chance to occur during a game, but only 16 percent of those go for touchdowns. The equivalent 6 point fantasy output of a rushing TD-a 60+ yard run-is 28 times less likely to occur than a rushing TD itself!

Rather than thinking of RB upside as big-play potential, it should instead be thought of as TD potential.

To further confirm this, I took a look at last season’s game logs of the top 36 RBs using 1 point PPR scoring. After eliminating a few games due to RBs leaving with an injury, there were 465 total games in my sample. I defined ceiling as any game where a RB scored 20 or more FP, and floor where a RB scored less than 13 FP.

2013 Ceiling and Floor Games, Top 36 RBs, 1 Point PPR Scoring
  CEILING GAMES (20+ FP) PCT FLOOR GAMES (Under 13 FP) PCT

3 Rush TD

4

4%

0

0%

2 Rush TD

29

25%

0

0%

1 Rush TD

68

60%

14

7%

0 Rush TD

13

11%

190

93%

Eighty nine percent of RBs who had games in the ceiling range scored at least one TD, and 29 percent had multiple touchdowns. Contrast this with the floor games, where 93 percent of RBs did not score a TD at all.

Recent big plays may cause a RB to be overvalued. If a player has inflated stats over the last game or two due to big plays, his salary may be artifically inflated, even though the liklihood of repeating the big plays isn't high. Be careful not to overpay for a player because he's had a few big plays lately.

I have a theory that many DFS players overemphasize the importance of big-play potential due to the availability heuristic, as defined below:

The tendency to overestimate the likelihood of events with greater "availability" in memory, which can be influenced by how recent the memories are or how unusual or emotionally charged they may be.

Big plays are exciting, and can even be “emotionally charged” for someone if they happen to own the player, are a fan of his team, or simply love great football plays. But do not let the availability heuristic prevent you from realizing that those big plays are also “unusual”, especially compared to TDs. Big plays can be a tremendous boost to a player’s fantasy output, but simply happen too infrequently for us to predict with much accuracy. Thus, Big play potential should still be taken into consideration when assessing upside, but should be deemphasized in favor of TD potential.

But doesn’t yardage have a higher correlation to FP/G than TDs, thereby making big plays more important, you ask? That’s a good question, but the answer remains no. Because big plays are so unlikely, volume is the best way accumulate yardage. Instead, target TDs for upside and think of big plays as a bonus.

By viewing upside in terms of TD potential, you can gain an advantage over owners who only consider big play potential. This can be a valuable contrarian strategy in tournaments.

For example, Spiller is frequently a popular upside pick, mostly because people remember his stellar 2012 season. That year, he had an unsustainable amount of big plays. But recall how much taking a handoff from inside the 4 yard line increases the likelihood of a rushing TD. In Spiller’s career, he has had exactly two carries inside the 4 yard line. Two. And those both came in 2011. Unless things change in Buffalo, by picking Spiller for upside you are betting on an event with significantly less odds of returning value than many other alternatives. On the other hand, Fred Jackson was generally a less popular pick, but outscored Spiller in almost every game because he was able to notch nine rushing TDs.

This season, a RB to beware is Andre Ellington. Last season, only one of his 118 carries came inside the 4 yard line, while Rashard Mendenhall got seven such opportunities. Mendenhall has since retired, but the Cardinals have two ‘big backs’ behind the 203 pound Ellington that may potentially inherit goal line work: 229 pound Jonathan Dwyer and 216 pound Stepfan Taylor.

 

How Game Scripts Affect RB Value

Whether a team is leading, tied, or trailing in a game-otherwise known as the game script-can have a drastic impact on a RB’s output. This is because game scripts usually have a huge impact on the amount of rushing volume a RB sees.

Play-calling by Game Script, 2013
  RUN % PASS %

Leading

51%

49%

Tied

44%

56%

Trailing

34%

66%

Teams generally heavily favor the pass unless they are leading in a game. You may have noticed I have yet to touch on the topic of strength of an opposing run defense. This is because that is only part of the puzzle-assessing a matchup should be more focused on how the RB’s opponent will affect his game script.

The goal is always to target high volume runners on teams that project to have a positive game script. However, game scripts are largely unpredictable, reinforcing why snaps are so important. If a RB’s team gets caught in a bad game script, you need him to still be on the field so that he has the opportunity to catch passes to make up for his lost rushing volume.

In my earlier study of game logs, I found that 66 percent of 20+ FP ceiling games occurred in games that the RB’s team won. While it is sometimes the case that a strong performance by a RB single handedly drags his team to victory, more often than not, winning begets good rushing performances. This is due to the increased volume that a RB sees in those situations, reinforcing the role that game script plays and suggesting that a RB’s upside increases the more his team plays with a lead.

I wouldn’t spend too much time trying to project game scripts, but do have an idea of how you think each game will play out and adjust your RB values accordingly. Check the Vegas point spread- large favorites or underdogs can give you a heads up as to what type of game script is most likely to play out. Teams with elite QBs or defenses are always a good bet to hold more leads.

 

How Home/Road Splits Affect RB Value

Due to the home field advantage present in the NFL, home teams generally find themselves in better game scripts, which results in better production for RBs at home.

Home/Road Rushing Splits, 2013
  RUSH % YDS/G TD/G

Home

43%

122.5

0.98

Away

40%

113.8

0.75

There is a significant increase in rushing TDs at home, so tend to favor RBs at home when choosing between players that are projected closely to one another. Going back to my study of 114 ceiling games, 56 percent of them occurred at home.

 

How Injuries Affect RB Value

Because NFL teams and players are not very forthcoming about the true extent of injuries, we have to take extra care in assessing how injuries will affect a player’s value. Past injuries can play a role weeks-even months- later. In 2013, Spiller and Ray Rice both played in 15 games, but played through injuries that did not allow them to be as effective as they normally are.

Injuries also affect different players in distinct ways. Spiller’s 2013 ankle injury limited him because he is a back that relies on quick cuts and lateral movements to break tackles. Conversely, Eddie Lacy’s ankle injury did not hamper him because he is more of a north-south runner who relies on power to break tackles.

 

Recap

  1. Allocate more salary per RB than per WR in cash games, but try to save on at least one RB slot in GPPs.
  2. Volume trumps efficiency for RBs- favor RBs who play a lot of snaps and get a lot of touches.
  3. A backup RB replacing an injured starter is very valuable so long as he will play most of the snaps.
  4. Almost all rushing TDs occur in the red zone or closer, so target RBs who get goal line work.
  5. Think of upside as TD potential rather than big play potential.
  6. Target RBs on teams likely to win the game-they will tend to see increased volume if their team is leading.
  7. Target RBs at home if possible.
  8. Injuries will affect each RB differently. 

 

Chris Raybon will be covering FanDuel and DraftDay value/optimal plays on a weekly basis each week this season. To learn more about Daily Fantasy Sports (DFS) click here.

 

Latest Articles
Most Popular