I'm a dynasty league junkie with a problem. While I love the NFL, I've never had any real interest in the high school or college games. For 364 days every year, this didn't matter... but on the day of my rookie draft, I always felt like I was operating at a disadvantage. I had only a passing familiarity with most of the names I would be considering, which meant that I was forced to base my player valuations on the opinions of others, of the fantasy experts and the armchair scouts who work so hard all year long to generate player profiles and talent evaluations on all of the incoming rookies. Unfortunately, I don't know who many of these experts are. I don't know how rigorous their processes are, and I certainly don't know how good their track record is. I have a very hands-on and analytical approach to fantasy, so this uncertainty has never sat well with me, but it was difficult to see any alternative. How could a lay person like myself with no actual expertise make high-stakes predictions about a very complex system that would be at least as accurate- and preferably moreso- than the predictions of experienced and skilled experts in the field? It seemed like an impossibility. Eventually, I received inspiration from an unlikely source; in the world of finance, there is a method whereby a fund manager with absolutely no underlying knowledge of the companies in which he invests can consistently receive returns that equal or even better those of the majority of his highly skilled, highly informed, highly compensated fund-managing peers. I've adopted, (and adapted), many of those key principles into a philosophy towards ranking rookies that I call "index investing", and in this series I will walk through the philosophical underpinnings of the strategy, the process of implementing the strategy, and the real-world impact of the strategy. Today, we will start with...
Index Investing 101: Is the NFL Draft an Efficient Market?
On the stock market, information is the most valuable commodity. Investors compile all of the information they can get on a company and synthesize it, deciding how much they believe a share of that company is worth. If another investor places a different value on that company, the two investors can buy and sell stocks between them until the price of the shares better reflects what the traders believe the company's value to be. Both parties have a strong financial incentive to price the company correctly. Because of that strong incentive, and because of the size and the scope of the market, it tends to be very efficient at pricing assets based on the information that is available at the time. This does not mean that the stock market prices assets perfectly; history will always show that, at any given time, many stocks were undervalued and many others were overvalued. Instead, it means that, given the available information, an individual trader will be unable to identify which assets are overvalued and which are undervalued at a rate that differs meaningfully from random chance. This is what's known as the Efficient Market Hypothesis, or EMH. As a consequence of this hypothesis, as many as 80% of actively-managed funds will underperform the market in general in a typical year. It's not that these funds are worse at pricing stocks than the market as a whole, it's that they're no better than the market as a whole, but they charge fees that reduce their average return below that of the market. In 1975, a man named John Bogle decided that there was no sense paying high management fees to an active manager who would fail to outperform the market benchmark, so he created the very first index fund available for public investment. The idea behind an index fund is that, instead of trying to beat the market average, he would merely try to track the market average. By buying a representative quantity of all of the stocks that make up a particular index, an index fund will in theory produce returns nearly identical to the market at large. Because the strategy is so simple, it can be accomplished with very little investment in terms of time, manpower, and trading fees, which allowed Bogle to reduce the fees he charged investors to levels previously unheard of in the investing world. At the time, the idea was wildly controversial- Bogle's fund was referred to as "Bogle's folly", and was even considered by some to be un-American. Experts predicted that the idea would never take off, because no investor would be satisfied with the prospect of merely average returns. By the end of the 20th century, though, index investing had become one of the most popular strategies, and Bogle's fund, the Vanguard 500, now manages over $100 billion in assets. In 2004, Time Magazine named Bogle one of their 100 most influential people.
By now you're wondering whether I'm going anywhere with this, or whether you've accidentally stumbled onto InvestmentHistoryGuys.com. You're still in the right place, I just felt that a little bit of background was necessary, because I'm going to present a radical concept that I suspect will be every bit as controversial to fantasy football players as index investing once was to investors. I'm going to propose a new form of the Efficient Market Hypothesis-- one that applies to the NFL draft. I believe that the NFL draft is a market, albeit one that trades in players instead of companies. I also believe this market shares many crucial characteristics with the stock market. For starters, NFL franchises, like investors, have a strong financial stake involved in ensuring that their valuations are as accurate as possible. Like investors, NFL franchises are gluttons for information, seeking out every single scrap of knowledge that might be at all relevant to a player's future. And like the stock market, while individual investors might not be perfect at pricing the assets involved, the draft involves an open competition between many different agents who are all looking to buy players when their price falls far enough, or sell players and draft picks if another team is willing to pay more than they are worth. These factors should combine to make the NFL draft a rather efficient market, a place where a player's listed value (or "draft position") accurately reflects the best understanding of all available information. This is not to say that the NFL draft is a perfect market- just like some stocks will be overvalued and undervalued, certain players will likewise be overvalued or undervalued. Sometimes players like Tom Brady or Marques Colston fall to the late rounds, while players like Jamarcus Russell or Charles Rodgers go in the top five. The EMH does not mean that every asset is valued perfectly, it means that individuals will not be able to identify which assets are overrated and which are underrated at a rate that differs to a statistically significant degree from pure random chance. That's the key takeaway. Yes, some assets are improperly valued. No, individual investors are not able to determine ahead of time which assets those might be. Some 6th round running back is going to be the next Alfred Morris, but the odds of picking which particular one it will be are no better whether you watch hours of tape or simply pull a name from a hat.
This belief has rather profound implications for how we choose to structure our "portfolio" of rookies in dynasty leagues. Some people spend hundreds of hours researching players, or reading opinions from others who research players. This time spent is the equivalent of a fee, a cost an owner pays. An index investor, on the other hand, can achieve similar returns with a radically lower time investment. And when the active manager begins to reach on lower-quality prospects, (as indicated by draft position), at the expense of higher-quality prospects, an index investor can swoop in, snatch up the extra value, and even come out ahead in the long run. In other words, index investing does not doom you to average results. It is possible, and even likely, that you will outperform your more actively-managed peers despite investing a mere fraction of the time. I imagine many will still be skeptical, so let me address some of the more common concerns and see if I cannot yet persuade you.
How can we be certain the NFL draft is an efficient market?
In short, we can't. Even in the stock market, one of the most closely scrutinized institutions in human history, it remains the "Efficient Market Hypothesis" and not the "Efficient Market Law". The hypothesis draws plenty of critics. Even its proponents are divided on just how efficient the markets are, with various factions advancing a "strong" form, a "semi-strong" form, or a "weak" form of the hypothesis. Still, while EMH remains unproven (and, according to some, unprovable), its remaining popularity owes to the actual observed results. From 1966 to 1996, two thirds of professional portfolio managers were outperformed by the S&P 500 Index, and there was little correlation between those who outperformed the market from one year to the next. This observed data fits poorly with traditional beliefs about the market. If picking overvalued and undervalued stocks were an actual skill, we would expect to see a subset of managers who were very "skilled" who repeatedly outperformed the market average, and a second subset of managers who were very "unskilled" who repeatedly underperformed the market average. On the other hand, this data comports nicely with the Efficient Market Hypothesis, which posits that outperforming the market would owe more to random chance than to skill, and therefore predicts that managers who outperform the market in one year are no more likely to do so again in the next year than any other manager.
Similarly, there's no way to prove that the NFL draft is an efficient market, but applying the EMH to the NFL draft allows us to make certain predictions and then check those predictions against the historical record. If the NFL draft were an inefficient market, we would expect to see certain front offices or GMs display a "talent" for identifying undervalued players. Those front offices would then compile a track record where they consistently outperformed the market year after year, compiling many strong draft classes vs. few weak draft classes over a long timeline. If, however, the NFL draft were an efficient market, we would expect that today's "talented" front offices- the front offices who have been most successful at identifying undervalued players and avoid overvalued players in recent drafts- would perform no better than average going forward, despite their recent successes. And, indeed, that is exactly what we see- various studies using different methodologies have concluded that there is little correlation between the quality of a team's draft classes from year to year. Another study has found that there's little correlation between the quality of a team's outcomes even from pick to pick within the same draft- a team that nails its first rounder is no more likely to nail its second rounder than a team that whiffs completely in the first. According to the conventional wisdom that some teams are better or worse at drafting than their peers, these results should be shocking; we'd expect the "good" drafters to consistently do well, and the "bad" drafters to consistently do poorly, resulting in a correlation coefficient well north of zero. Instead, we see a correlation coefficient of just 0.07, which is barely distinguishable from zero (or no correlation at all). If it's hard to believe that some front offices simply are not better at identifying talent, then I invite you to look back at recent history. If you'd asked the average observer who the best drafters in the league were after the 2009 season, no doubt you'd get many nominating Bill Polian of Indanapolis or A.J. Smith of San Diego. After all, Polian's Colts were coming off of a record 7th straight 12+ win season and a second Super Bowl berth. Instead, Polian found himself out of a job within two years, after an injury to Peyton Manning revealed all of Polian's personnel blunders and left the Colts as the least talented team in the league. Smith's Chargers, on the other hand, had for years been acknowledged as one of the most talented teams in the league, and had finished the 2009 season with a 13-3 record. Several years prior, Smith had found himself in a power struggle with coach Marty Schottenheimer after a 14-2 season, and San Diego's owners fired the reigning Coach of the Year rather than risk losing Smith, who they believed was arguably the best personnel man in the league. Since 2009, San Diego's record has continuously declined, and Smith is now just an assistant in Washington after years of alienating his veteran talent and failing to replace them led to his dismissal from San Diego.
Smith and Polian are hardly isolated examples. After the Patriots had won three Super Bowls in four years, observers believed they could do no wrong in the draft and gave them the benefit of the doubt for every bizarre-seeming decision. Since the last Super Bowl title, however, New England's record in the NFL draft ranks anywhere from below-average to outright poor. Kansas City hired Scott Pioli, the general manager responsible for those New England drafts, thinking he'd quickly raise the talent level on the team. Pioli also currently finds himself among the ranks of the unemployed and Kansas City has produced four or fewer wins in four of the last six seasons. Meanwhile, some of the most hapless or snakebitten franchises have produced shockingly good drafts in recent years. The Cincinnati Bengals, famous for their owner's thriftiness and unwillingness to spend on a decent scouting department, suddenly find themselves with some of the best young talent in the league after drafting players like A.J. Green, Geno Atkins, Andy Dalton, Andre Smith, Carlos Dunlap, and Leon Hall. The Detroit Lions, the saddest of the sad-sack franchises, became a punchline after spending high-1st round draft picks on receiver busts Roy Williams, Charles Rogers, and Mike Williams. They find themselves the current home of the best receiver in the league after spending a fourth high first rounder on Calvin Johnson, and have acquired several other core franchise players such as Matt Stafford, Ndamakong Suh, and Nick Fairley through the draft. San Francisco has gone from one of the most talentless franchises in the league to arguably the most talented with a stunning run of draft success that netted them players like Patrick Willis, Navarro Bowman, Aldon Smith, Colin Kaepernick, and the best offensive line in the league.
Perhaps the most dramatic example is the Denver Broncos, who have been a veritable roller coaster in the draft over the last two decades. In the early part of the Mike Shanahan era, Denver drafted a talented core that would contribute to two Super Bowl championships and contend for division titles through the early part of the 2000s. Then, Shanahan seemingly lost his touch, producing some of the worst results in the league from 2000 to 2005, before seemingly regaining it and producing some of the best results in the league from 2006 to 2008. Denver then fired Shanahan and turned their draft decisions over to Josh McDaniels, who became an instant punchline by wasting multiple high 2009 draft picks on busts like Knowshon Moreno, Robert Ayers, Richard Quinn, Alphonso Smith, and Darcel McBath. McDaniels seemed to be destined for punchline status again in 2010 after spending a 1st rounder on Tim Tebow, but instead the 2010 class quietly turned out pretty well, headlined by McDaniels' decision to select the raw Demaryius Thomas over the more highly-regarded Dez Bryant, and bolstered by the later selection of receiver Eric Decker. Then, in 2011 and 2012, under the guidance of the inexperienced John Elway, Denver continued its success, assembling the rest of the core that would win the AFC West in 2011 under Tim Tebow and earn the #1 seed in the AFC in 2012 under Peyton Manning. After so much early success, it'd be tempting to say that John Elway and company simply have the draft figured out. If history is any guide, however, we should expect them to perform no better than average going forward.
So that must settle it, right?
Not really. All of this evidence so far is merely suggestive, and a true proof of the EMH would be, for all practical purposes, impossible. And as with their financial counterparts, even supporters of the idea that the NFL draft is an efficient marketplace disagree among themselves to some extent. Just like there is a strong form, semi-strong form, and weak form of the EMH in finance, there is a strong form and weak form of the EMH as it applies to the NFL draft. Supporters of the strong-form EMH believe that all players' draft positions should be assumed to be the result of an efficient marketplace. Supporters of the weak-form EMH argue that the NFL draft differs from the stock market in that it lacks a mechanism to push prices down by any substantial degree. They will point out that it only takes one team to "reach" on a player and draft him higher than his talent warrants, while it takes 32 teams working in agreement for a player to fall further than anticipated. They will suggest that this means players who are drafted below expectations are necessarily the result of a consensus reached by an efficient market, while players drafted well above expectations are not. Instead, they'll argue that players drafted well above expectations might simply represent the flawed evaluations of a single franchise who decided to "reach" on that player. For instance, they might suggest that the draft position of Dez Bryant, who was selected many pick after he was expected to be, is meaningful. In essence, the 32 franchises all agreed that Bryant was not worth as high of a selection as the pundit class believed, because none of those franchises traded up to select him with a higher pick. On the other hand, proponents of the "weak-form" EMH might suggest that the draft position of A.J. Jenkins is not particularly valuable information, as perhaps the other 31 franchises agreed with the pundit class that Jenkins was a 2nd round talent, but the San Francisco 49ers reached on him and took him in the 1st round, anyway. Both positions certainly have strong arguments in their favor. The "weak-form" proponents are certainly correct to note that the NFL draft does not have a mechanism to push down player valuations more than a single pick at a time (compared to the stock market, which can quickly react to overvalued stocks to force the price down dramatically and immediately). I tend to favor the "strong-form" version of the hypothesis, though, out of sheer practicality; it is impossible to know what the actual NFL consensus really was. Perhaps the San Francisco 49ers overreacted and drafted Jenkins well before the consensus. Then again, perhaps the consensus was much higher than the pundits realized, and San Francisco was right to select Jenkins when they did. Since we have no way of knowing which scenario is the truth, I err on the side of trusting the front offices, who have greater resources and stronger incentives, over the pundits who publish their opinions for public consumption.
So if the draft is an efficient market, why do teams spend so much on scouting?
It is worth noting that the EMH doesn't imply that NFL front offices are not good at their jobs, or that they could be replaced by a dart-throwing chimpanzee with an ESPN Insider subscription. On the contrary, it means that NFL front offices are the best in the world at their jobs, and the constant competition forces them all to the highest level. Any edge to be gained has long been discovered and copied by the rest of the league, leaving a level playing field… until the next time some enterprising team discovers a market inefficiency or a new store of information and gains another temporary advantage, which will last only as long as it takes for the rest of the league to catch up again. Likewise, any team that sports an inherent disadvantage, whether in the form of outdated practices or incompetent leadership, is strong incentivized to fire their leadership and modernize their practices until they again can compete on equal footing with the other franchises. I'm sure if a team were to hire a random person off the street to run their draft room, that team would have the worst draft in the league. A random fan or follower of the league would be totally incapable of competing against the best and brightest minds the NFL had to offer. Instead, all 32 teams feature some of those best and brightest minds, and no single franchise is able to gain a lasting advantage over any of the others.
Does this mean rookie rankings should be a perfect reflection of draft order?
Certainly not. The NFL is quite good at distributing assets based on what is valuable to an NFL franchise. What the NFL values, however, diverges from what fantasy owners value- sometimes quite radically. In the NFL, players sign long-term contracts worth millions of dollars, and the salary cap ensures that mistakes can harm a team for years to come. As a result, the NFL has a strong and rational incentive to be risk-adverse, avoiding players with uncertain injury or character backgrounds. Fantasy owners in redraft leagues care far less about a player's long-term future, and can greatly discount those risks in their evaluations. Even in standard dynasty leagues, the worst cost of a failed pick is just the opportunity cost of acquiring and holding the player- unless your league features a salary cap, the long-term ramifications of cutting a player are nonexistent. As a result, while the NFL might prefer a less talented but less risky player, it is perfectly reasonable for a fantasy owner, due to the different risk tolerances, to prefer the higher-upside player with more red flags.
The NFL values many skills and abilities which are irrelevant in most dynasty leagues. Tight ends in the NFL are typically valued based on a combination of their blocking abilities and their receiving abilities. Few, if any, fantasy leagues award points for blocking, so while the NFL might prefer a tight end who is above average as a blocker and a receiver, fantasy owners would most likely rather have one who is an exceptional receiver and a liability as a blocker. Similarly, the NFL might rate a player higher based on potential contributions to special teams, or to scheme flexibility. The NFL might envision a player as an exceptional role player or complementary piece. In most leagues, these skills do not result in fantasy points, and players whose draft stock relies heavily on them should be discounted.
The NFL values positions very differently than the fantasy community. Everyone is aware that the most important player on an NFL team is always the quarterback, so it's no surprise that NFL franchises are willing to take bigger gambles in an effort to secure a good one. Quarterbacks make up just 4.5% of a team's starting lineup, and account for roughly 5.5% of the players on a 53-man roster at any given point, yet over the last decade quarterbacks have accounted for 8.8% of all players selected in the first round of the NFL draft. For an NFL franchise, even securing an average quarterback counts as a success; just imagine how much the 2012 Arizona Cardinals would have paid to have a merely mediocre signal caller. In fantasy, on the other hand, an "average" starting quarterback is not the 16th or 17th best player in the league, but the 6th or 7th best. If an NFL team spends a 1st round pick on a quarterback who becomes the 14th best player in the league, he was a success by NFL standards, but a disaster by fantasy standards. Running backs, while the lifeblood of most fantasy football leagues, are widely viewed as fungible pieces in NFL offenses, and their draft stock is typically deflated as a result. The running back drafted in the first round is going to be a lot more valuable for fantasy purposes than the typical quarterback drafted in the first round. Fantasy owners must adjust for these differences in positional weights when creating their rookie draft boards.
Finally, NFL teams are drafting players based on how talented they are, while fantasy owners are drafting players based on how productive they are. This distinction is important, as a more talented player in a bad offensive system will often be less productive than a less talented player in a good offensive system. While the NFL draft gives us a good objective measure of a player's talent level, the old fantasy cliche is that "Production = Talent + Opportunity". Even if the NFL draft served as a perfect indicator of player talent, that still leaves us with only halfway to our goal, and it falls on us to do the required legwork to bridge the gap.
Another problem with the Efficient Market Hypothesis should be readily apparent simply from the description. The EMH states that the market is "informationally efficient"- that assets are priced appropriately based on all available information. This immediately raises questions of unavailable information. Teams will do their best to acquire as much information as is possible prior to the draft, but it's impossible to know everything about a single person, let alone each of the hundreds of people whose services are being bought and sold. New information will come out after the draft- a player might get a DUI, or make inappropriate comments in public, or show an unexpected competitive drive or an unusual degree of resilience or mental toughness. Players will impress more than expected during training camps, or hold out for an extended period. Players will get injured, or recover from injuries more quickly than anticipated. Any of these developments, and thousands more like them, represent new information. This information was not available at the time a player was drafted, and is therefore not reflected in the player's draft position. The Efficient Market Hypothesis suggests that if we know nothing more about a player than that he was drafted in the 5th round, we should be skeptical about how good of a prospect he is. It does not say that we should remain skeptical as we learn more; if that same 5th round RB rockets up the depth chart during the offseason, lights the world on fire during preseason, and earns the starting job by opening week, our skepticism should diminish with each new development. We should assimilate each new piece of information, weigh it against the old information, and continually update our beliefs and expectations to reflect the total sum of our knowledge rather than anchoring our expectations to a single data point.
So where do we go from here?
Hopefully by now I have established the basis of my belief that the NFL draft is an efficient market. Hopefully I have adequately laid out the case that no individual, no matter how smart or talented, is capable of judging player talent at a rate that consistently outperforms the market consensus. If so, it's time to transition from the realm of theory into the realm of practice. In the last section of this article, I briefly touched on the process of adapting the information we receive from the NFL draft into something that is useful for our specific purposes in fantasy football. In Part 2 of this series, I will expand on that concept by providing a step-by-step walkthrough of how to use the Efficient Market Hypothesis to generate a set of rookie rankings. To illustrate the process, I will take you through the steps I use to create my own rookie rankings which you see posted on this site. Finally, in Part 3, I will close by laying out the case for implementing an Index Investing strategy in your own dynasty league. Stay tuned!
More from Adam Harstad:
Dynasty, in Practice: Valuing in the Face of Uncertainty - September 19
Dynasty, in Theory: A Paean to Uncommon Sense - September 16
Dynasty, in Practice: Early-Season Overperformers - September 12
Dynasty, in Theory: Thinking Like a Bayesian - September 9
Dynasty, in Practice: Keeping a Fantasy Journal - September 5
Dynasty, in Theory: Musings on Confirmation Bias - September 2
A Narrative History of Fantasy Football - August 28
Diversification 101 - August 19
Dynasty, in Theory: The Components of Player Value - August 14
The Opportunity Cost of Top Tight Ends - August 11