“On a long enough timeline, the survival rate for everyone drops to zero.” – Fight Club

The Similarity Score apps will let you project 2013 stats for fantasy players by looking at the performance of similar historical players. The apps are great tools to use as you prepare for re-draft leagues and in fact I used the apps while I was drafting a recent dynasty team as well. But one of the considerations in dynasty formats is how much value the player you’re drafting has not just in this season, but also following seasons. Because of the short nature of a lot of football careers, I would argue that near term results should probably carry with them a premium, but in any event it’s possible to create rough approximations of near term and long term value.

For the exercise below I used the same methods used in the Similarity Score apps to create estimates. Except that instead of stopping at estimating Y1 fantasy points, I also looked at Y2, Y3 and so on. Let me use an example to show how this works.

Adrian Peterson just had a season that was remarkable, but was also similar to some past seasons by other running backs that might be rough approximations for AP. His 2012 was similar to the following seasons:

Barry Sanders199729
Ahman Green200326
Jamal Lewis200324
Barry Sanders199426
Willie Parker200626
Shaun Alexander200427
Curtis Martin200431
Barry Sanders199628
Rudi Johnson200526
Mike Anderson200027
Rudi Johnson200627
Corey Dillon200430
Edgerrin James200527
Clinton Portis200827
Curtis Martin200128
Terrell Davis199826
Jamal Anderson199826
Barry Foster199224
Barry Sanders199527
Average 26.9

I listed the average of the ages of the backs because in 2012, AP was 27 years old. So even though these backs range from 24 to 31 years old, on average they were really close to AP’s age.

Then I simply looked at what those players did after they were similar to AP-2012 in order to come up with multi-year projections. Before I get to the projections, let me just say a few things about what this analysis is good for, and where it might be limited. This is a good exercise to go through to get some sense as to the long term value of players. You’ll get a rough picture as to how many productive years a running back has left before the odds tip in favor of a decline. But it will also illustrate the idea that over time, the outlook for all players starts to head to zero. Football is a violent sport that is a little bit like a demolition derby. Players are lucky to just make it to the next season in a lot of cases and each season adds more years and mileage to a player.  Even young players are up against the fact that there is a fragility involved with playing football. A 23 year old player might have a number of years in front of them, but that also means they have 1,000 carries to sustain an ACL injury or broken bone.

Here’s where the analysis will have limits: this is all based on using the player’s most recent season to create probabilities of future success. The player’s most recent season will usually be the best data to look at, but it won’t always be the best data. For instance, this exercise can’t account for Shonn Greene’s move from feature back to back-up. I’ve tried to make an estimate for Chris Ivory, but it might not be accurate. Using a median expectation will generally get the most players right, but it could also miss on outliers (for instance, I know that people will say that AP is an outlier and not appropriate for comparison to historical players).

Additionally, it’s interesting to think about this in the context of “Back to the Future.”  Everything below is conditional upon what has already happened. But when the future plays out, the projections will also change. So if AP has the same kind of season he just had, but does it in 2013, then his future projection will change. In the same way that Marty McFly’s existence changed when George McFly punched Biff at the Enchantment Under the Sea dance, AP’s projection will change if his 2013 season deviates from the projection below. If AP runs for 2000 yards again, he’ll essentially buy himself another year of the model predicting him to be a top back.

There’s also another idea that I think is important to consider and that is the idea that until a player has actually created value, it’s not there yet. We could have all predicted Ryan Mathews to be a multi-year fantasy stud if we would have just speculated based on his 2011 season. But until a player actually becomes a fantasy stud (Mathews didn’t quite get there) it’s not the most accurate to project it on into infinity. You’ll notice that David Wilson’s name doesn’t appear in the table below and that’s because Wilson’s value is not the kind of thing that this exercise will be any good at predicting. He has to have value before we can start to project how long he’ll be able to maintain it. There are other projection systems that might be more suited for establishing Wilson’s value, including the WAG method (Wild Ass Guess).

Having offered those guiding thoughts, here is the table which shows a year by year projection of PPR points, along with two summary columns. In one column I’ve simply added all of the points together and in the other column titled “NPV” I’ve estimated a discounted value of the player’s total points. NPV is equal to Net Present Value, which is the idea that future production is worth less than today’s production. I used a discount rate of 25% in the NPV calculation.

Name – Age20132014201520162017TotalNPV
Doug Martin – 2318.6215.5315.4616.321479.9344.02
Ray Rice – 2516.216.9314.7313.8311.1672.8540.66
Trent Richardson – 2213.4614.5715.5514.3610.1968.1337.28
C.J. Spiller – 2516.314.6914.7311.626.7964.1336.97
Alfred Morris – 2415.7812.729.689.055.9753.2031.38
Arian Foster – 2614.3812.859.187.332.9946.7328.41
Marshawn Lynch – 2615.8213.
Stevan Ridley – 2312.049.9711.5106.3149.8228.06
Mikel Leshoure – 2212.5611.3610.185.8039.9024.91
Adrian Peterson – 2715.1810.959.18.54035.8524.07
Jamaal Charles – 2613.1812.437.252.68035.5423.31
Chris Johnson – 2713.1410.596.463.762.7736.7223.04
Ahmad Bradshaw – 2613.3510.257.275.08035.9523.04
LeSean McCoy – 2412.1112.295.566.14036.1022.92
Matt Forte – 2711.910.986.676.52036.0722.63
DeMarco Murray – 2410.3511.688.225.96036.2122.41
Chris Ivory – 2411.529.966.963.55031.9920.61
Reggie Bush – 2712.767.914.164.171.2430.2419.51
Jonathan Dwyer – 238.568.735.625.841.6730.4218.25
Shonn Greene – 2710.967.795.532.8027.0817.73
Frank Gore – 2911.349.514.25.94026.0417.72
Vick Ballard – 229.736.637.281.82025.4616.50
Steven Jackson – 2910.637.564.250022.4415.52
BenJarvus Green-Ellis – 2711.417.342.25.7021.7015.26
Willis McGahee – 3110.115.463.820019.3913.54
Ryan Mathews – 258.295.414.332.01020.0413.13
Knowshon Moreno – 259.525.751.890017.1612.26
Bryce Brown – 216.844.216.111.02018.1811.71
Darren McFadden – 257.796.152.060016.0011.22
Michael Turner – 308.826.500015.3211.22
Andre Brown – 265.874.814.570015.2510.11
DeAngelo Williams – 297.684.400012.088.96
Alex Green – 245.064.563.290012.918.65
Bernard Pierce – 223.363.623.442.921.2714.618.38
Cedric Benson – 306.431.90008.336.36
Mark Ingram – 233.552.13.29008.945.87
Jonathan Stewart – 254.322.330006.654.95
Fred Jackson – 315.9300005.934.74
Kendall Hunter – 242.841.91.91006.654.47
Maurice Jones-Drew – 274.4800004.483.58
Michael Bush – 283.1700003.172.54
Donald Brown – 252.37.330002.702.11
DuJuan Harris – 241.7500001.751.40
James Starks – 261.1900001.190.95

Finally, some notes and thoughts:

  • In order to estimate Chris Ivory’s future years, I created a hypothetical 2012 season for him where he got 15 touches a game at 4.5 yards per carry and then also got Shonn Greene’s touchdowns/game. Those were the stats that the similar players were based on. Despite Ivory’s age, the future outlook doesn’t extend forever because Ivory hasn’t been involved in the passing game thus far in his career. If that changes in NY, it will change Ivory’s outlook.
  • The model doesn’t know that Ahmad Bradshaw’s career is probably ending because of injuries.
  • The model also doesn’t know that DET picked up Reggie Bush. But even with Bush, it is interesting that Leshoure is still going to be just 23, so he might be a dynasty “buy low” right now.
  • Both Ryan Mathews and Darren McFadden are hurt by poor 2012 seasons. If they bounce back this year, their future year expectations would also rise.
  • Mark Ingram’s outlook probably isn’t as bleak as it looks here, but it probably isn’t significantly better either.
  • This doesn’t mean that I think that Willis McGahee is actually more valuable than all of the guys below him. There’s obviously more information that needs to go into the analysis. This exercise just answers the question of “Conditional upon a back of this size having a season like he just had, what is the median expectation for the future irrespective of changing opportunity?”
  • If this analysis has biases, or preferences built in, here’s what they probably are:
    • In favor of pass catching backs
    • In favor of younger backs
    • In favor backs that had good 2012 seasons
    • In favor of bigger backs

Subscribe for a constant stream of league-beating articles available only with a Premium Pass.