Not many draft guides will tell you this, but the way their projections and rankings are presented to the consumer are faulty and a fragile source to begin your draft planning process.
As today’s fantasy football has become more statistically saturated than ever, it has also become a game that is beginning to develop a crutch for rankings and projections, which is a scary endeavor to me, because we lose some big picture focus on what is relevant to playing successfully. What your other sources of projections won’t honestly tell you up front is that these numbers attached to players are simply a mark of probability. A wave of probability that is undoubtedly the highest it can possibly be in a sea of variance.
Player projections begin with opportunity first and foremost. The projection model starts off with historical and league environment trends coupled, which is then integrated with the same trends on a team and coaching staff level. To get a grasp on what kind of opportunity will be available for each player, we use that information to generate overall play volume and a split of running and passing plays for each team and then layer in individual usage and historical player trends to the process.
From there, things run into a mountain of variance because from a top-down view, efficiency in the NFL is unbelievable volatile and carries a staggering low rate of correlation coeffeciency (the measure of which two variables are related) year over year. Football is a sport where even the strongest correlation coefficients leave a lot of room for variation. For example, the highest year-on-year correlation for a statistic on a team output level is pass completions, which carry a R^2 of .43. In other words, 43 percent of team’s passing completions can be predicted by the previous season. That’s it. Not even a coin flip of predictability. And that’s the strongest measure we have on a team level. On an individual player level, the strongest measure we have is receiving targets, which check in with a correlation of .53. When it comes to NFL football, the data we have available is still largely more descriptive than it is predictive.
That’s why it is important to incorporate a range of outcomes into your approach for a specific player over linearly following just one set of projections. Projections are the most probable full season scenarios, but again, I don’t have the ego to blindly follow them alone, and neither should you. If we hold ourselves accountable based on what the industry has defined as accuracy, even the very best in that regard have a hit rate that closely resembles Shaquille O’Neal’s career free throw percentage. Relying on the confidence placed in vacuum projections is a large part of the reason(s) why you should also never rely on value-based drafting. Tying my entire draft plan to a tent of statistics over a four-month span leaves a lot of variance open on the table and can lead to a lot of hollow weeks. The most important part of any question is accounting for where you’re wrong, and that is something amplified in any field of prognostication. The goal is to win weeks like everyone else, but emphasizing overall points scored first isn’t as pressing as chipping away on the supply and demand, opportunity cost, replaceability and team structure elements of the game.
That is why you’ll find that our rankings here aren’t just a linear list of what the projections say. They are created with a wider lens over overall scoring, because even if you are supremely accurate in nailing those end game numbers, they still mean wildly different things depending upon how a weekly game like fantasy football is played. For example, Travis Kelce and Greg Olsen had nearly identical overall statistical seasons a year ago using the counting stats that are scored for fantasy football and ended the season as the first and second-highest scoring players at their position. Kelce caught 85 passes for 1,125 yards and four touchdowns while Olsen caught 80 passes for 1,073 yards and three scores. On a per-game level, they were just .3 receptions, 3.2 yards and .6 touchdowns apart in production. But depending upon who you had rostered in fantasy has drastically different results at different moments in the season. Through nine games, Kelce had 42 receptions for 466 yards and was a top-12 scorer in just four of those games. So, in less than half of Kelce’s opening nine games played – which was roughly 75 percent of your fantasy regular season – he didn’t even turn in a starting-caliber performance based on his position. On the other hand, Olsen started off the season with 50 catches for 712 yards and three touchdowns over that span with seven top-12 scoring weeks in nine games and in six of those seven he was a top-six performer at the tight end position. For most of your fantasy regular season, Olsen was significantly more important to your team.
Of course, we know how the rest plays out. Kelce went nuclear over the final stretch of the season while Olsen faded with the Panthers’ season. Olsen had just 27 catches for 339 yards and zero touchdowns the rest of the way through Week 16 while Kelce had 42 catches for 651 yards, which led the entire NFL. The end of the season will show both as being equal through a broad scope, but the way fantasy football is scored and played shifted the values of each player significantly over the course of the season. Similar projections can still mean significantly different results at the end of the day.
That’s one anecdotal example used to make a point, but there is a litany of others that are similar. Football – and most importantly for context, the fantasy side – is a week-to-week game in which what goes into the results of that week will vary greatly. There are only a handful of David Johnson-types of week-to-week players available per season, and there’s a reason many of them are selected in the front half of the first round.
That is why the ranks and tiers won’t always align linearly with the projections you’ll find here, because those projections are just one outcome and a lone outcome that is tallying overall performance for a week-to-week game. It is also why when you reach the tiers section, the tiers are structured in a way that places players in buckets based on how similarly they score points. If you are following a linear list of rankings and projections, you may end up in a raft where your guide is encouraging you to select Amari Cooper, Brandin Cooks, Donte Moncrief and DeSean Jackson as your wide receiving unit, all players that score their fantasy points the same way. Sure, you’re going to run into some monster weeks when the stars align for that group, but you’re also going to heighten the variance of your team performance and run into many low-scoring weeks because each of these players scores their stats in a similar fashion. This is where team structure comes into play in unison with those rankings and projections that often are overlooked by the average player because it’s not explained on a list, and rankings are handled and approached differently by nearly all of those who create them. The tiers are set up to closely resemble the rankings, but also offer a path to proper roster construction when selecting a balanced roster based on the way players accrue their scoring.
In closing, the most important takeaway is that you shouldn’t approach fantasy football with the old school mindset of “I have to win the draft” because you and your opponents are going to make a lot of mistakes during that process. Preparing yourself for a multiple range of outcomes heading in will lead you to not only selecting a better team during the draft, but will also prepare you for swings that occur during the campaign.
