Nassim Nicholas Taleb, author of The Black Swan, has a new book out, and, if anything, it’s probably even more potent (if also self-contradictory in many places).
The book is called Antifragile and this is his definition of the term:
Some things benefit from shocks; they thrive and grow when exposed to volatility, randomness, disorder, and stressors and love adventure. Yet in spite of the ubiquity of the phenomenon, there is no word for the exact opposite of fragile. Let us call it antifragile. Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the antifragile gets better.
Although Taleb is adamant that the real world is nothing like the world of games – that game theory works fine to address the contrived experience of the game – I was still struck by the way his theory of antifragility seemed to dovetail with my fantasy football philosophy.
In just the past four years I’ve played well over 200 leagues, many of them for substantial amounts of money. And here’s the thing: value-based drafting, the default method used by almost all supposed fantasy “experts,” does not work.
I’ve spent much of the last two offseasons working on a book explaining why this is the case. It includes graphs and formulas and long-winded explanations and will hopefully be out next summer. In the meantime, Taleb’s simple framework may be enough to grasp the crux of the issue.
Taleb divides things into three groups: fragile, resilient/robust, and antifragile. While plenty of potential nuance exists in how you might approach a fantasy draft, let’s take a look at three simplified strategies using the matrix suggested by Taleb.
|Fantasy Strategy||Value-based drafting||Triple RB or RBx5||Zero RB|
One of the key issues in VBD focuses on creating replacement baselines. In his groundbreaking 2012 Rotoworld article, RotoViz creator Frank DuPont demonstrates how most analyses generate inaccurate and misleading baselines. The essential issue in fantasy football deals with running backs and their relative scarcity. FD demonstrates how the injury problem with running backs pushes their baseline even lower than traditionally understood and demonstrates the necessity for having even more redundancy at the position than you might otherwise think.
This need for redundancy was one of the reasons behind my suggestion that RB-RB-RB should be the default strategy for 2013. Or at least the default strategy in many formats. I wrote a separate column specifically to detail where it would work and why, but I’m starting to think I may have included too many formats in a desire to advocate resiliency when I should have stuck to my guns and just recommended antifragility.
Before the season, I explained some of the reasons why a RB-heavy approach may not yield positive results in Flex Wins Championships, the article I consider to be my most valuable RotoViz contribution so far. But the Flex-based argument may be fairly tangential to the real issue. And perhaps I was wrong to recommend RB-RB-RB in any format other than the most basic standard league.
The Fragility of VBD Lineups
VBD lineups are fragile for two main reasons.
1) They are overwhelmingly fragile in the face of prediction errors. If the prediction you are using to generate your VBD number is wrong, you’ll end up targeting both the wrong players and wrong positions. Since specific predictions – instead of the range of potential outcomes we provide with the Sim Scores – are almost always wrong, a VBD team is incredibly fragile. The most obvious way these prediction errors cause problems is in the way they can trick you into drafting a mediocre running back over an elite wide receiver simply because you’re projecting the RB to score more points above replacement.
2) They are overwhelmingly fragile in the face of injuries. Since VBD teams tend to end up fairly balanced and with an emphasis on the starting lineup, they have nowhere to go but down when injuries strike. This point can be easily overlooked because humans are incredible narrative-generators. When injuries inevitably occur, the most obvious conclusion to the VBDer is that “I was unlucky.”
The Resilience of RB-heavy Lineups
The types of lineups that Frank recommended in his article, I recommended in RB-RB-RB, and Matthew Freedman recommended in the stylishly hilarious RBx5 are very resilient in the face of both the uncertainty of predictions and the certainty of injuries. The RB position has undergone a ridiculous amount of upheaval since the beginning of the season, but if you began RB-RB-RB, there’s an excellent chance you are now in position to weather the storm. Many RotoViz favorites have experienced injury, underperformance, or both – C.J. Spiller, Doug Martin, Trent Richardson, for example – but many other players we praised have excelled. Jamaal Charles, LeSean McCoy, Matt Forte, Reggie Bush, and DeMarco Murray were all among our preferred targets.
The central idea behind Triple RB is that not all of your predictions will come true and some players will sustain injury. That’s been very much the case this year. Unfortunately, the Triple RB approach is merely resilient. It gives you redundancy in the face of both injuries and bad predictions, but your team will not benefit from those injuries or bad predictions.
The Antifragility of Zero RB
Zero RB is basically what it sounds like. You simply don’t draft running backs in the high leverage rounds. Depending on how a draft is progressing, I will draft either one high upside running back in Round 4 or 5, or I will draft none at all. My preferred lineup after five rounds is to own one tight end (Graham/Gronk) and four wide receivers. I then focus on selecting potential breakout players, the receiving back in timeshares, and backups in good offenses.
You can see fairly easily how Zero RB benefits from randomness. Whenever a starting RB gets hurt, my lineup gets better. It gets better in relation to my opponents because I didn’t have the player in question, and it gets better in the sense that I either own the backup or I have a shot to acquire the backup in free agency.
A quick note about ethics: It’s easy to see the above paragraph as parasitic, but I don’t believe it is. I’m not causing anyone to get injured, and in fact I’m expressly not rooting for it. I own Kendall Hunter in a lot of leagues and believe he would immediately be a Top 10 back if Frank Gore were injured. I am not rooting for this. I’m a big believer in karma. (I believe it’s a good concept whether or not it’s a true concept.) I tend to think of this in the following terms: if I root for someone to get injured, it’s more likely one of my foundational players will be injured instead. These are real people with real careers that are far more important than my fantasy results. I’m rooting for Adrian Peterson and Marshawn Lynch to get stuffed at the goal line each week, but I am not rooting for them to get hurt.
The historical results suggest running backs get hurt at a higher rate than other positions. The decision to structure rosters to take advantage of this fact doesn’t cause it to happen and is ethically neutral. When it naturally occurs, my roster gets better and everyone else’s gets worse. Right now on many of my teams I own Andre Ellington or Zac Stacy. A lot of my high stakes teams have both players. Without having spent any draft capital at RB, I’m actually stronger at the position than many of my opponents.
It’s also important to simply remember the power of wide receivers in ppr formats. I used a Zero RB approach in this year’s PFF Draft Guide Mock even though it’s essentially a draft champions format with no free agent pickups. My team was the leading scorer in Week 8 despite getting only 1 point from the running back position. Update: FD has an excellent follow-up piece that examines this crucial point in more detail: Winning the Flex; aka: Score Points, What’s Your Plan?
In Part 2 I’ll look at some examples of this strategy in action: I use this approach in the leagues where I have large sums of money on the line. In one of these leagues, I was heckled relentlessly during the draft itself for the decision to eschew running backs. The results in that league so far have been surprising, counterintuitive, and perhaps antifragile.Author’s Note: Part 2 – Antifragile Rosters in Real Events has been updated to provide season-long results for those Zero RB squads. Part 3 – Zero RB, Breakout Stars, and Having More Fun Drafting was penned this spring and looks at a variety of strategies for consistently hitting on breakout running backs.