While the RotoViz staff is all pretty much in agreement that the best long-term approach to fantasy football is scientific in nature—flexible enough to change in the face of new evidence—we still have different specific ways that such a foundation manifests itself in the way we draft.
One big difference that I’ve noticed among some of the writers is the willingness to embrace volatility, particularly early in the draft. Jon Moore falls in the same camp as me as someone who seeks safety early in the draft—the first couple rounds, at least—and then pure upside later. Others—and I think the Douche falls into this group—don’t necessarily specifically seek safety early on.
When I was writing the content for my new fantasy football books this year, I really started moving away from thinking about a player’s value in terms of a comparison of his projected points to his draft slot. One reason for that is because player projections are inherently fragile—very susceptible to small alterations. If you need to project a running back within 10 yards of his actual output to have him accurately ranked, for example, you’re going to be in trouble. For the record, I’m not against creating projections because I think there are other uses, but I don’t think you should blindly use them to rank players.
Another reason I’ve moved away from value-based drafting is because I’m thinking about a player’s value more in terms of what he can offer me in regards to his range of potential outcomes. Frank mentioned a couple weeks ago that Aaron Rodgers is probably the safest quarterback in the league because he’s not dependent on any of his receivers for production; he could legitimately lose any one of them and probably be in a spot not much worse than before.
The way I’m thinking about player value is more probabilistic than ever before, and I’m implementing a Taleb-inspired barbell approach to drafting in which I’m generally seeking extremes in players. One extreme is that I’m looking to target either rookies (or second-year players who underperformed) or aging players at any spot other than running back. The reason is that I think the middle area—players in their prime—see their draft stock artificially inflated. Everyone wants those guys, and while they’re optimal in a vacuum, they probably aren’t in real life because there’s a price to pay. I think I can find value on certain young players who are underpriced due to an inefficient draft market and some older players—guys like Andre Johnson—who still have enough meat left on the bone to give me a nice return.
Another barbell-themed approach is to pursue pure safety or pure upside in picks, as opposed to a more risk-balanced approach. It’s like putting some of your money into an ultra-conservative investment and then going super-risky with a smaller portion, as opposed to dumping all of your cash into a moderately risky investment with only moderate upside to accompany it.
Before I go into this idea too much more, I want to show you some data I just compiled on the variance in the RotoViz composite re-draft ranks for quarterbacks, running backs, wide receivers, and tight ends. I charted this because I think the RotoViz rankings—an aggregate of multiple individual player rankings—are probably a decent proxy for inherent player/position volatility. When the writers are all over the place on a particular player or group of players, it probably means he’s more of a risk/reward guy with a wide range of outcomes than those on which there’s a consensus.
I sorted the players into three buckets. To me, it’s surprising to see that the top five running backs and receivers had less deviation than the top five quarterbacks. That fits with my completely-anecdotal-early-season-mock-draft evidence that we should be jumping on only the first few elite backs—for me, it’s Charles, McCoy, and Forte—before moving to wide receiver in the first round. Personally, I’ll be gunning for a late pick in just about every league because I think it more easily allows for an antigrafile draft approach, while also ensuring you get two elite wide receivers (if that’s how you want to play it). It’s a Dez/Demaryius combo versus maybe Charles/Keenan Allen with an early pick, and I prefer the former duo.
Note that while quarterback deviation is higher than that for backs and receivers, passers also have much lower ADPs. I’m not a proponent of a true early-quarterback strategy this year, so I probably won’t own Peyton Manning because I think the cost is too rich, but I’m probably one of the few writers on this site who is going to be drafting a quarterback in the middle rounds—quite frequently landing Rodgers or Luck, in all likelihood.
I understand all of the arguments against such an approach, and I’ll be the first to admit that it doesn’t work out in terms of strict VBD. I also don’t care because, like I said, examining players in such a manner is fragile and too deterministic. There’s value in just knowing what you’re getting. We see this all of the time in daily fantasy football; most weeks, the top-priced options are “overpriced” in terms of a strict value sense and the bargain bin players offer “sensational” value, but there’s also value in just getting the points you expect to get. I’m not necessarily anti-streaming, although I do think it’s highly contingent on your ability to actually play matchups in an effective manner—easier said than done.
The other issue is that these elite quarterbacks are just falling too far. I’m going to take Aaron Rodgers in the fifth round of every single draft that he falls, and I’m going to be happy about it. It’s not that I’m doubting that his production can be replaced later. It’s that I think the probability of actually doing it is lower than people believe and increases the value of Rodgers’ extremely high floor.
Anyway, there’s some other cool stuff going on with this data, so let me know your thoughts in the comments.