Readers sometimes ask me for recommended fantasy football reading, and I always forward them to a book that has nothing to do with fantasy football: Nassim Nicholas Taleb’s Antifragile. In my opinion, the biggest leaps you can make as a fantasy owner aren’t in the area of football analysis, but rather risk analysis.
We always hear that we’re supposed to minimize risk and maximize upside, but few ever tell us how to do that. I want to use Taleb’s “barbell” investment strategy as a template for how I think that’s best accomplished. First, an excerpt from Antifragile:
What do we mean by barbell? The barbell (a bar with weights on both ends that weight lifters use) is meant to illustrate the idea of a combination of extremes kept separate, with avoidance of the middle. In our context it is not necessarily symmetric: it is just composed of two extremes, with nothing in the center.
I initially used the image of the barbell to describe a dual attitude of playing it safe in some areas and taking a lot of small risks in others, hence achieving antifragility. That is extreme risk aversion on one side and extreme risk loving on the other, rather than just the “medium” or the beastly “moderate” risk attitude that in fact is a sucker game (because medium risks can be subjected to huge measurement errors). But the barbell also results, because of its construction, in the reduction of downside risk—the elimination of the risk of ruin.
So basically what we’re looking at is an extreme approach to fantasy football drafting that involves patching up potentially disastrous leaks in our strategy (extreme risk aversion) while trying to hit home runs (extreme risk seeking), as opposed to a more moderate strategy that emphasizes both risk aversion and risk-seeking behavior with each pick.
This is how I’m planning to implement a barbell fantasy football approach this year.
For rankings, projections, exclusive columns, mock drafts and tons more, check out our jam-packed online Draft Guide or Draft Guide iOS app, or follow @BalesFootall and Rotoworld Football on Twitter for the latest news.
A Barbell Approach to Positions
When I first began my fantasy football analysis, one of the strategies I proposed was drafting a quarterback semi-early for the sole reason of it being safe. Aaron Rodgers, Peyton Manning, Drew Brees—unless those players get injured, we know what we’re getting, and we’re going to be comfortable with it. There’s value in that.
Are quarterbacks the shrewd early-round play given their scarcity? No, they’re not, but I think a pure value-based draft strategy assumes that player performances are governed by a deterministic set of laws and all we need to care about is median projections.
The entire philosophy behind the barbell approach is understanding where we might be fragile to estimation errors. On the position level, every position is more error-prone than quarterback. The key is thinking of players in terms of probabilities with a range of potential outcomes; most elite quarterbacks have a narrow range of possible outcomes and don’t warrant early selection in the strict value-based sense. But we’re not looking to nab a few extra expected points of VBD with a barbell drafting strategy (while simultaneously opening ourselves up to measurement errors).
In effect, the cautious end of the barbell strategy works as an insurance policy for your team. When you pay for insurance, there’s no value in the strict sense; insurance companies make money by charging you more than you’ll put in over the long run. But there’s still a ton of value in insurance because it limits your exposure to massive downside. That’s exactly what an elite quarterback does to your fantasy roster (at least at his position).
The other reason that the early-quarterback approach is underrated, in my book, is that the cost of securing these elite passers is now incredibly low. We’re seeing Brees & Co. fall into the fourth and even fifth round in expert drafts, which is absurd. Now, these quarterbacks are at least close to offering value in the strict projected point/scarcity sense, in addition to the barbell-based reason we want to draft them: they aren’t that susceptible to projection errors and thus limit our exposure to huge downside.
So which quarterback should you draft? The answer is ‘I don’t know,’ which kind of illustrates the whole point. You don’t need to be extremely accurate in your individual player assessments because there’s not an extremely high bust rate among elite passers; the value comes in the fact that you can pick an elite quarterback, any quarterback, and he’s likely to act as your team’s insurance policy.
When you draft other positions, even in the first round, you necessarily require a higher level of accuracy in your player evaluations—a whole lot of risk to assume for a few projected points, right?
Editor’s Note: Rotoworld’s partner FanDuel is hosting a one-week $100,000 Fantasy Football Contest for Week 1’s games. It’s only $10 to join and first prize is $10,000. Starts Sunday, September 7th at 1pm ET. Here’s the link.
Wide Receivers
I want to quickly touch on wide receivers. Unlike running backs, whose production is very binary—they’re either producing or not based on their workload—wide receiver production is more skill-based; while maybe 80 percent of running back production comes via factors outside of their control (I made up that number but I wouldn’t be surprised if it’s near that), a much larger degree of wide receiver production is due the actual talent.
We know that because we can predict which wide receivers will be good with a higher degree of accuracy. Note that I’m not saying we can predict which ones produce more easily than at running back—running back isn’t terribly hard to predict because it’s mainly just a quick glance at anticipated workload—but just that we see the same wide receiver types do the same things over and over.
There’s an extremely strong correlation between wide receiver weight and red zone production, for example. There’s also a very strong correlation between catch rate and yards-per-target; wide receivers who see shorter targets (usually from the slot) are the safest types of players on a weekly basis.
This relates to the barbell investment approach because it seems smart to construct your team with the two different wide receiver extremes—big receivers who can score often and short-target players (typically slot men) who give you a really high floor of production in a given week—but nothing between those two extremes.
What’s between those extremes? Players like Mike Wallace. I often hear other fantasy owners equate speed to upside, arguing that Wallace and similar players are high-ceiling options because they can score from deep. I disagree completely because the way they thrive is based off of low-frequency events (long touchdowns) that are highly unlikely to be repeated multiple times in a single game.
Want evidence of that? Up to this point, Wallace has played 79 games during his NFL career, but he has a career-high of just 144 yards. One-hundred forty-four yards! For one of the game’s most explosive players? Can’t be. Wallace has also never caught more than two touchdowns in a game in his entire career, and he’s done that only five times.
Wallace doesn’t have a high ceiling, even though that’s fun to believe, but he also doesn’t have a very high floor. Because he’s dependent on big plays for production, he’ll turn in some stinker performances. He’s also extremely easy to take out of games. Want to stop Wallace? Just play press-bail; don’t let him catch quick screens, but bail after the snap so he can’t get deep. That’s what defenses have done in the playoffs—when Wallace has compiled an average of four catches for 35 yards and 0.25 touchdowns per gave over four contests.
Now let’s compare Wallace to a receiver like Vincent Jackson, whose ADP has been similar to Wallace’s over the years. Since Wallace came into the league, V-Jax has topped Wallace’s career-high single-game receiving total on five occasions (with three more that have come within three yards), while also scoring three touchdowns a couple times. Or how about Josh Gordon, who beat Wallace’s career-high for receiving yards four times in 2013 alone!
The point here is that we can classify wide receivers fairly easily and make accurate assessments regarding their upside and safety. If we’re going to create a general barbell-based rule-of-thumb for receiver drafting, it’s this: draft wide receivers who can consistently dominate in the red zone, giving them high week-to-week ceilings, but if you side with a wide receiver who doesn’t score a lot, he better see a bunch of short targets (giving him high week-to-week consistency), a la Percy Harvin, Kendall Wright, and so on.
If a wide receiver isn’t going to score a lot of touchdowns but also isn’t likely to limit our downside on a weekly basis, he’s probably not going to fit the barbell-based wide receiver prototype that we’re seeking.
A Barbell Approach to Age
I’ve recently started obsessing over player ages a bit and I think there’s a potentially barbell-based inefficiency we can exploit. Namely, it seems like most fantasy owners are obsessing over the next big thing—the players who are on the verge of breaking out. That’s logical and the public hits on those players pretty well.
But what’s the cost? Every year, we see up-and-coming players rise up draft boards to the point that they no longer offer value. Recent examples include Lamar Miller and David Wilson.
Meanwhile, the players at the extremes of the age spectrum—rookies and old-ass veterans—can get overlooked. Let me be clear that some rookies get hyped up and are overvalued, but because fantasy football ADP mirrors NFL draft order (and because the NFL draft is perhaps inefficient), we can get value on certain types of rookies: big mid-round wide receivers (Keenan Allen), late-round running backs (Zac Stacy), and so on.
At the other end, we have fading players like Andre Johnson. Heading into 2013, Johnson was coming off a season with 112 catches for 1,598 yards and four touchdowns, yet getting drafted in the back of the third round. That had everything to do with Johnson being 31-years old. He went on to catch 109 passes for 1,407 yards and five touchdowns.
We shouldn’t necessarily have expected that kind of season from Johnson. Actually, there was reason to expect him to regress. But at a late-third-round ADP, Johnson was falling way, way too far. This example is one of three age-related inefficiencies I think can be exploited:
1. Players on the way down (veterans like Johnson) who are falling too far in drafts
2. Players on the way up (“prime of career” players like Wilson) who are rising too far in drafts
3. The unknowns (rookies like Allen/Stacy) whose price doesn’t reflect true risk/reward
Remember, we’re not only looking for good players, but good players at the right cost. Yes, ideally we want players who are in the prime of their careers, but guess what? So does everyone else. That causes this vicious cycle where the player’s ADP rises, then owners think they need to draft him even earlier, so his ADP rises more, so they draft him even higher, to the point that now all of his value has been sucked away. And the opposite happens for veterans who owners want to avoid.
Basically, I think we can use a barbell approach to age, going super young (basically just rookies or second-year players who disappointed in their first season) or “post-prime.” That’s true for every position except for running back. There are undervalued rookies there, so that’s fine, but I simply don’t want older running backs. Those players—Maurice Jones-Drew, Marshawn Lynch, whoever—tend to just fall off of a cliff such that when they drop, there’s just no meat left on the bone. That’s super-risky because we don’t know for sure when that decline is going to occur.
Compare that to the aging curves for every other position, where the drop isn’t so dramatic.
Quarterbacks can stick around for a long time. Wide receivers and tight ends see a dip in efficiency in their mid-20s, but their production isn’t as binary as running backs and we can usually spot signs of their demise. Meanwhile, running backs are basically the equivalent of a guy who died in a hang gliding accident. What an idiot. “Ahhh I’m hang gliding, honey take a picture I’M DEAD!!”
No old running backs!
As I was thinking about this age-barbell approach, I considered how it might affect week-to-week scoring, and I think it’s for the best. Too often, fantasy owners think about the season as one single entity when it’s really composed of 16 weeks (or however many you play in your league). Unless you’re in a total-points or best-ball league, your goal is to win each individual week.
The rookie/oldie strategy I’ve proposed maximizes your ability to do that because when one group is playing poorly, the other tends to do well. Specifically, there’s evidence that veterans begin the season hot and fade near the end, whereas rookies are at their best late in the year.
That last point flies in the face of conventional wisdom, which suggests that rookies fade down the stretch because they aren’t conditioned to playing so many games in a season. In real life, that might be the case, but rookies also get so many opportunities late in the year that, for fantasy purposes, they’re far more useful late in the season. Take a look at rookie receptions, for example.
Rookies see a huge boost in chances to make plays near the end of the season. You might feel uncomfortable starting an unproven rookie in your fantasy playoffs, but the numbers suggest it could be the best bet.
To give you an idea of how the distribution of points throughout the course of a season can affect your fantasy success, consider these two groups of quarterbacks.
Team A
QB1: 320 points (20 points every game)
QB2: 240 points (15 points every game)
Team B
QB1: 320 points (30 PPG for first eight games, 10 PPG for final eight games)
QB2: 240 points (8 PPG for first eight games, 22 PPG for final eight games)
In terms of total points, these two pairings are exactly the same; the QB1 scores an average of 20 PPG for each team, and the QB2 has an average of 15 PPG. The result for your fantasy team could be far different, however.
If you were to manage Team A perfectly, you’d just start QB1 every game, and he’d provide you with 320 points on the year. If you were to manage Team B properly, however, you’d start QB1 for the first eight games to accrue 240 points, then QB2 for the final half of the season and 176 fantasy points—a season-long total of 416 fantasy points.
The results wouldn’t be this extreme and we also don’t have perfect knowledge of fantasy outcomes, but the idea is that a barbell approach to age can help you more effectively maximize your team’s week-to-week output.
There are lots of potential barbell situations out there. If you’re considering a specific course of action, ask yourself ‘Does this plan either significantly limit my downside or significantly enhance my upside?’ If it’s neither—if you have the Mike Wallace of fantasy football strategies—it’s best avoided.
