Print

RB Collapses

You are viewing free content provided by FantasyGuru.com. Why not consider subscribing today?

by Mike Horn, Staff Writer

Published, 5/20/14

 

I’ve been playing around with estimating the dynasty values of players. I haven’t quite got that article ready, but I thought I’d share some of the research by focusing on Top-10 RBs whose future performance collapsed.

 

By “collapsed,” I mean that a player went from being a top 10 RB one season to having almost no future fantasy value. So first, I have to define value.

 

We all can agree that owning the #1 RB (RB1) is better than owning the #11 RB (RB11). The gap between them is typically large in fantasy points, certainly larger than between RB10 and RB11. The issue is how much larger. I decided that for me, if the value of RB1 = 300, then the value of RB11 was one-third of that and the value of RB21 was only one third of RB11. Not everyone may agree, but I believe it’s far easier to plug together RB11 performance from lesser RBs by playing matchups than it is to find RB1 scoring week-in-week-out by patching together a lineup. Plus, if you have RB1, you’re only scrambling to find help at the next tiers down, and you probably need less of it, whereas if you’re scuffling to find RB1 scoring, you’re probably in trouble throughout the position group. The same holds true for RB11 vs. RB21 value, and on down through the RB ranks.

 

In line with that logic, I came up with these approximate values:

RB Rank

Value

1

300

11

100

21

33.3

31

11.1

41

3.7

51

1.2

61

0.4

71

0.1

 

You’ll notice I just kept dividing by three. On a graph, those values look like this (the blue diamonds correspond to the numbers in the table):

 

 

The black curve is the trendline Excel fits to the blue diamond, with the curve having the equation shown. Using that formula, I’m able to calculate or interpolate all the other ranks, so RB10 has a value of 161 (x=10 in the equation shown). RB11 is 99.9 using the formula.

 

I then gathered up the stats and fantasy rankings of all RBs from 1988 to 2013 (all 16-game seasons), eliminating those backs still active in 2013, so I could measure RBs through the end of their career. I assigned values to each RB’s seasons using the formula from the curve (not the table, so an RB11 season is worth 99.9 not 100).

 

Fantasy scoring used: 10 yards rushing/receiving = 1 FP; rushing/receiving TDs = 6 FP. Rankings are based on total FP.

 

For each running back, after each season I added up how much value they had left in the remainder of their career. For example, following the 2004 season, Shaun Alexander (the RB1 that year) played three years and ranked RB1, RB28, and RB35. Those ranks correspond to values of 300, 15.4, and 7.1, so at that point he had 323 total value left (rounding up). But after finishing #1 again in 2005 (value =300), he only had 23 total value left in his career.

 

That brings me to the idea of a collapse: when an RB’s past (just-finished season) value is worth far more than his future value. Those are guys you want to avoid like the plague for dynasty, as it’s likely will produce nothing like their past season ever again. Even in a re-draft format, you want to avoid those RBs: they are going to be overdrafted based on last year’s numbers, even if owners discount them a little for aging, and they will be disappointments.

 

I defined “collapse” numerically as such: past year value minus total future value is greater than 100. If you think about the table and curve, to be a “collapse” under this formula, a player has to have at least a past year Top-10 finish (value of greater than 100) and score very little in any future season. For Alexander, who was an extreme case since he finished as RB1 in 2005, his past year value at that point was 300. Subtracting 23 from that gives you 277. The only bigger collapse in the period 1988-2013 was Terrell Davis, who was RB1 in 1998 (value = 300) and had subsequent seasons as RB77, RB58, and RB46 (total value = 3). At least Davis had what basically amounted to a career-ending injury, while Alexander seemed to just fall apart.

 

Some Top-10 RBs never collapse. Take Emmitt Smith for example. He had nine Top-10 finishes in the 90s. In 1999, he was RB5 (value = 193). Because he had a remaining career value of 105, he avoided being classified as a “collapse” even though he fell to RB20 in 2000 and never broke the Top 20 again. His redraft owners in 2000 were probably disappointed; his dynasty owners did get four years of RB20-26 performance in from 2000 to 2004, which wasn’t great but had some value.

 

Well, I don’t think I can predict a career-ending injury. And I’ll miss some declines like Smith’s. But is there anything that can warn us that a collapse is imminent?

 

Age is an obvious place to start. Since 1988, there have been 205 RBs who at one point finished in the Top 10 and has since retired (at least they didn’t play in 2013). Did those players collapse after their Top 10 finish? Let’s look at this table:

Did the RB Collapse?

Age

Yes

No

Pct Yes

21

 

6

0%

22

3

10

23%

23

3

11

21%

24

6

27

18%

25

1

19

5%

26

8

21

28%

27

10

23

30%

28

9

10

47%

29

7

10

41%

30

4

7

36%

31

4

1

80%

32

4

 

100%

33

 

1

0%

Overall

59

146

29% 

Age is as of December 31st in the “past year.” Remember, collapses are in the years after that. There were 59 collapses overall, or 29% of the total Top 10 RBs studied. Age 26 is a clear break point: at that age or younger, RBs have a below-average rate of collapse. From age 27 up, the rate of collapse is above average. If we just break the data at that age:

Did the RB Collapse?

Age

Yes

No

Pct Yes

21-26

21

94

18%

27+

38

52

42%

This would tell you that if you own an RB who was 26 last year and ranked in the Top 10, you’re safe holding him and expecting a reasonable amount of future value, but from age 27 on you’re playing with fire. Hold a fire sale and trade him, right?

 

Let’s slice the data a little finer:

Did the RB Collapse?

Age

Yes

No

Pct Yes

21-25

13

73

15%

26-27

18

44

29%

28+

28

29

49%

I think the true message is that RBs under 25 are in little danger of a collapse. Some of their future value may be diluted, but they are pretty safe. At age 26-27 the collapse rate is equal to the overall average of 29%. That is the danger area. From 28 years old on, it’s about 50-50 whether an RB has much value left. And remember, this misses guys like Emmitt Smith, who had a gradual decline.

 

With Shaun Alexander, at age 27 and as RB1, you might be inclined to hold on for a year and be rewarded with his RB1 finish in 2005 at age 28. Now with two straight RB1 seasons and five years in a row of RB6 or better, most of us would ride him out and get burned. But if you traded him at that point, you probably were happy with the outcome.

 

Is there another way to spot the “collapsers?” I once read that a decline in pass-catching was a “tell” when it came to imminent decline for RBs. I tested this. I calculated how much of an RB’s fantasy points came from catching the ball (receiving yards). Receiving TDs are very hit-or-miss, so I didn’t include them. RBs who scored fewer than 10% of their FP as receivers I deemed as non-pass catchers (NPC); all other were pass catchers (PC).

Did the RB Collapse? (Pct Yes)

Age

PC

NPC

21-25

16%

12%

26-27

21%

53%

28+

50%

47%

A bit surprisingly, the young and old RBs (21 to 25 and 28+) non-pass catchers actually collapsed at a slightly lower rate. Those differences are probably barely meaningful though. The important number, although perhaps not statistically significant, is the 53% collapse rate of the non-pass catching RBs aged 26 and 27. Those would be the guys I’d worry about most in that group.

 

I looked further into how all these RBs produced their fantasy points. Besides receiving yards, I also looked at rushing yards and TDs (rushing and receiving). I calculated the percentage of FP that came from each of these three categories, as well as the averages and standard deviations for those percentages within each category:

 

·         Average of past season FP from Rushing Yards: 54%, standard deviation +/- 8%

·         Average of past season FP from TDs: 31%, standard deviation +/- 7%

·         Average of past season FP from Receiving Yards: 15%, standard deviation +/- 8%

 

Not a lot jumped out at me, and the smaller the data the less reliable the conclusions, but for those cases in which I had at least 30 RBs in a subcategory (out of the 205 total RBs in this study):

 

·         RBs who had fewer than 46% of their past season FP from rushing yards (i.e. more than one standard deviation below average) collapsed 41% of the time (vs. 29% overall).

·         RBs who had more than 38% of FP from TDs collapsed 40% of the time (vs. 29% overall).

·         RBs who had more than 23% of FP from receiving yards collapsed 39% of the time (vs. 29% overall).

·         RBs age 21-25 who had a “normal” percentage of FP from rushing yards (within one standard deviation of the average only collapsed 10% of the time (vs. 15% overall).

·         RBs age 21-25 who had a “normal” percentage of FP from all three categories (+/-1 SD from average) only collapsed 7% of the time (vs. 15% overall).

 

By far the RB group least likely to collapse of any meaningful size are RBs age 21-25 with a typical or normal distribution of FP scoring. Put more simply, the safest RBs to own are young backs who didn’t have extreme amounts of FP from any one area of scoring (rushing, receiving, or TDs).

 

I don’t think that’s a surprise. In fact, was more surprised by the relative instability of the future of 26-27 year-old backs, particularly those who don’t catch a lot of passes.

 

So what does all this mean for our current crop of Top 10 backs? All statistics are for last season:

Name

Rk

Age

Rush
PCT

Rec
PCT

TD
PCT

Risk Factors for Collapse

Jamaal Charles

1

27

41%

22%

37%

Age = Medium; Rush PCT<46% adds risk

LeSean McCoy

2

25

57%

19%

24%

Age = Low; Normal amount of Rush PCT = low risk

Matt Forte

3

28

50%

22%

27%

Age = High

Marshawn Lynch

4

27

52%

13%

35%

Age = Medium

Knowshon Moreno

5

26

44%

23%

33%

Age = Medium; Rush PCT < 46% adds risk

Adrian Peterson

6

28

60%

8%

31%

Age= High

Eddie Lacy

7

22

56%

12%

32%

Age = Low; Normal FP distro = very low risk

DeMarco Murray

8

25

54%

17%

29%

Age = Low; Normal FP distro = very low risk

Chris Johnson

9

28

53%

17%

30%

Age = High

Reggie Bush

10

28

52%

26%

22%

Age = High; Rec PCT > 23% adds risk

The table identifies the key factors that put each back at risk for collapse (or make it less likely he’ll collapse). For example, Jamaal Charles has a medium risk factor for age (worse than the 21-25 year olds, better than the 28+ year olds) and since his percentage of FP from rushing yards was less than the average minus one standard deviation, that is an additional risk factor for him.

 

Eddie Lacy and DeMarco Murray are the RBs from last year’s Top 10 who are least likely to collapse, based on both their age and the fact that their FP scoring is distributed with +/- one standard deviation of the average for all three categories.

 

LeSean McCoy is the other young back in last year’s Top 10, but he was a little outside the norm for TD PCT so he only has one of the lowered risk factors for young backs (normal amount of FP from rushing yards), so he’s a tad riskier than Lacy or Murray.

 

Of the middle-aged backs besides Charles, Knowshon Moreno also has elevated risk because of the large percentage of his FP that came from receiving yards. Marshawn Lynch surprisingly avoids the 10% level as a non-pass catcher.

 

I know people will scoff at Matt Forte or Adrian Peterson as having high risk. Some notable RBs who collapsed after their age 28 seasons: Marshall Faulk, Eric Dickerson, Roger Craig, Brian Westbrook, and Jamal Lewis. I mentioned Shaun Alexander already. On the other hand, some who did not: Priest Holmes, Barry Sanders, Thurman Thomas, and LaDainian Tomlinson. Tomlinson didn’t collapse after two straight years as RB1 but did slip to RB7, then RB19, RB17, and RB40, so his dominance was coming to an end faster than most expected. Emmitt Smith isn’t on the list because he only ranked 18th at age 28, then bounced back to RB6 at 29 and RB5 at 30 before his slower decline. If I owned either Forte or Peterson in dynasty, I’d sell high if I could, but otherwise would ride them some more.

 

Reggie Bush has the highest collapse risk of last year’s Top 10. Both age and overreliance on receiving yards are risk factors for him. It’s possible that PPR scoring could affect this some but I wouldn’t count on it.

 

Of course this study doesn’t account for other risk factors like injury (Murray) or leaving Peyton Manning (Moreno). So don’t forget the other things you know. But when evaluating risk for collapse, don’t ignore age. Eventually, it gets us all.

 

As an addendum to this article, Horn looks at how workload affects running backs.

 

3,897 people wish they still made Flutie Flakes.

Back to the top