What Matters More for Rookies: Skill or Landing Spot? (Fantasy Football)
The NFL draft draws ever-nearer and football fans wait with bated breath. Each team has the eternal hope of finding a diamond in the rough (in case you somehow forgot, Tom Brady was taken 199th overall) and turning a heretofore-hapless franchise around. More importantly for our purposes, though, are the fantasy implications. For the avid fantasy manager, the NFL offseason is a time to scout incoming rookies. The possibilities are endless, and excitement can turn to joy – or sorrow – during the NFL draft.
I’m talking about a common concept for the fantasy landscape: a player’s landing spot, or which NFL team they actually end up on. Conventional wisdom holds that the landing spot plays a crucial role in determining a player’s output in their first year. Najee Harris and Elijah Moore are great examples from 2021. Both proved to be talented NFL players with the promise of a long career ahead of them, but Najee ended up on the Pittsburgh Steelers (a team committed to running the ball where Najee would be the featured back) and Elijah Moore on the New York Jets (no explanation needed). Najee ended as the RB4, and Elijah Moore as the WR48 (although an injury did cut his season short in Week 13).
Fast forward to this year’s draft. Jason has professed his faith in top RB prospect Breece Hall, but notes that dreams can turn to dust if Breece is drafted into the wrong situation. For example, the Buffalo Bills would be a dream spot (great offense, lack of established RB) while the Tennessee Titans would be a fantasy disaster (it’s not likely that Breece would ‘goblin up’ a lot of carries from King Derrick Henry, at least not yet).
In this article, we’re going to try to measure the effect of team landing spot compared with a player’s skill. What is more important, how good a player actually is, or how good a situation they find themselves in? All data is from nflfastR (half-PPR scoring since 2002).
I’ll start by acknowledging that these effects – landing spot and individual skill – are very tricky things to measure. We’ll be considering purely redraft fantasy leagues (performance in just this upcoming year) since looking later into a career makes the estimation problem significantly harder. I will also be ignoring one major variable: draft capital, or how high a team drafts a rookie. This can indicate how much a front office expects a player to be involved in the team system, but I simply don’t have reliable data for it.
Anyways, to proxy ‘landing spot’, I’ll be using the ‘points returning’ and ‘points vacated’ across all positions for a team. For example, the Los Angeles Rams had 203.9 running back points vacated in 2020. This is all coming from Todd Gurley, who scored those 203.9 points in 2019 and then departed for the Atlanta Falcons in 2020. If Gurley had stayed on with the team, those points would have been considered ‘returned’ (in addition to 78 other points from back-up Rams RBs that stayed). Note that this doesn’t include how the upcoming season actually progresses, it’s just a measure of which players stayed or departed during the offseason.
To me, this seems like a reasonable measure for landing spot because it gives us a sense of team prowess (total points scored at each position) and opportunity. For example, the Kansas City Chiefs have a ton of points vacated at the WR position with Tyreek Hill taking his talents to South Beach. This ‘target vacuum’, as its often referred to, represents a choice landing spot for incoming wideouts.
Measuring skill is a bit more difficult. Ideally, I would include all of the NFL combine metrics, as well as the prospect rankings of various analysts. Unfortunately, I don’t have that data and, what’s more, those predictions are often pretty noisy. What I settled on is using future performance; as Jason likes to say, the best predictor of future fantasy performance is past performance. By symmetry, the best predictor of rookie performance is the future performance of that player!
Specifically, to measure skill, we will be looking at the maximum season-long point total scored by a player after their rookie year. This might not seem like the best metric at first glance, but it does give us a sense of what players are capable of once we’ve really seen them shine. For example, Christian McCaffrey scored an absurd 413 points in his third season (2019). This is a good measure for his innate, ultimate ‘skill’ or ‘potential’ in the fantasy sense, and we can compare how the rookie season stacks up. I’m only including players that scored 100+ fantasy points post-rookie season at some point in their career to (mostly) filter out less relevant players; it also means that one-year players like Najee Harris will be excluded (since we don’t have a big enough sample size to estimate his post-rookie ability that we can then compare the rookie campaign to).
With all of these variables collected, I’ll be running a simple regression to estimate the points scored during the rookie year. I’ll also be controlling for the actual calendar year, since there are strong effects over time (rookie WRs perform way better now than they are used to!). Let’s go through the positions, starting with the most important in the rookie fantasy landscape.
Here are the significant* variables for predicting the rookie year performance of 195 RBs:
|Returning RB Pts||-0.37|
I’ve normalized all of the variables so we don’t have to worry about comparing units or different amounts. For example, the top number means that “for every 1-standard deviation (STD) increase in returning RB points, the model predicts a 0.37 lower standard deviation score for a rookie RB.” Here, the exact opposite is true for ‘Maximum Points’: a 1-STD increase in maximum points predicts a 0.37-STD higher score for a rookie RB.
Remember that ‘Returning RB Pts’ proxies how ‘good’ a landing spot is. This result is intuitive, since if a team is returning a lot of RB production (i.e., stud RBs didn’t depart in the offseason) the prospects aren’t great for an incoming rookie. It also takes into account teams where RBs didn’t leave, but also didn’t actually score much, like the Buffalo Bills. Devin Singletary and Zack Moss should be back in Buffalo, but they didn’t score much in 2021, so the team won’t be ‘returning’ many points at the position.
The ‘Maximum Points’ metric proxies skill, since this is the highest total the player will ever go on to achieve after their first year. It makes sense, then, that it has a positive relationship with rookie performance. Finally, it’s important to note that none of the other variables were significant; these were the only two that really mattered.
What’s the takeaway then? Well, interestingly, it appears that skill and landing spot matter the exact same amount! Both have the predicted effect: more opportunity and more skill mean a better expected rookie season. However, they have the same strength: the bump you get from being a +2 standard deviation prospect is exactly offset (in this model) by the drag you get from landing in a -2 standard deviation destination. We sort of saw this with Javonte Williams this year: he was a very talented back who ended up in a pretty bad landing spot with Melvin Gordon returning to the team. These mostly offset and Javonte was, on the whole, decent for fantasy (finished RB17).
Here are the significant variables for predicting the rookie year performance of 241 WRs:
|Returning WR Pts||-0.21|
Again, we only have two important variables: the WR points a team is returning, and the maximum points a prospect will go on to score. The direction is the same as for running backs, but here the skill proxy is twice as strong. It is interesting to note that none of the other variables – like QB points returning – are significant in the model. The opportunity for volume truly appears to be king, a phenomenon we saw Brandin Cooks exhibit last season.
This is good news for young wideouts, because it means they are more robust to landing spot: skill has a way of shining through in a stronger fashion. This is intuitive: while running backs need to get their foot in the door before they are actually given carries, wide receivers have a better chance to at least get on the field. Once there, player talent can help to demand targets, thus leading to higher fantasy production. A great example last year was Amon-Ra St. Brown, a talented receiver who made it work in one of the worst offenses in the NFL.
Here are the significant variables for predicting the rookie year performance of 97 QBs:
|Returning QB Pts||-0.48|
Another intuitive result: the harmful effect of a bad landing spot is the strongest for QBs across all positions that we’ve seen. This makes sense, since usually only one QB gets significant playing time in a game, which means that a rookie drafted behind an established signal-caller will have a tough time eking out production. The San Francisco 49ers are a great example: Jimmy Garoppolo stayed on with the team, which made Trey Lance‘s 2021 fantasy output a moot point.
Here are the significant variables for predicting the rookie year performance of 83 TEs:
|Returning TE Pts||-0.28|
Rookie tight ends are sort of a long shot and, outside of generational talents like Kyle Pitts, are less relevant for fantasy. Still, it’s interesting to note that, like with RBs, the effect of landing spot and skill offset.
There are certainly flaws with this approach, but I think the results are useful and intuitive. If I were to summarize what I found for the two most important rookie positions, it would be that the skill and landing spot effects are equal and opposite for RBs, whereas skill is much more important for WRs. It’s also important to note that no variables other than points returning at the position and the proxy for skill were significant. The moral of the story is that it’s less important how strong the rest of the team is; what we want to focus on is the volume opportunity at a player’s specific position.
Keep this in mind for the upcoming draft. I hope your favorite prospect ends up in a good situation.
Unless otherwise stated, we’re considering a 10% level of significance.
Questions? Let me hear it on Twitter.