The Fantasy Mythbusters: Does Draft Capital Matter?
The 2023 NFL Draft is in the past, which only means that rookie dynasty drafts are in our present and future. Being on the clock can be intimidating; after all, this player will (hopefully) be on your roster for years to come. Do they have the right skill set to succeed in the NFL? Will their landing spot help or hurt them? Did the NFL team that drafted this player invest enough on draft day to signal they are bought in?
You’ll hear this last point referenced often by fantasy pundits. Who could forget James Robinson, the waiver wire wonder — overall RB7 — of the 2020 season? Unfortunately, the Jacksonville Jaguars were not as grateful as championship-winning fantasy managers. They promptly drafted Travis Etienne Jr. in the first round as a replacement, and 24-year-old Robinson now shores up the backfield of the New England Patriots.
What happened? Why were the Jags so quick to move on? Many have answered with two simple words: draft capital. Robinson was an undrafted free agent, meaning that the team didn’t actually invest their draft picks/significant cap space with him. On the flip side, players with high draft capital — i.e., first and second-rounders — aren’t kicked to the curb as often, since the team spent dearly to sign that player (in addition to using high picks, rookie contracts have contingencies based on rounds drafted).
Or so the conventional thinking goes. Here, though, at the Fantasy Mythbusters, we are interested in testing convention. Are teams really less apt to move on from guys they drafted early? Or are they immune from some version of the sunk-cost fallacy? Just because a player was drafted high doesn’t mean he will be helpful going forward.
To answer, we can scrape draft data from Pro Football Reference since 2012 (and yes, thank you ChatGPT for helping me prepare the scraping script). Here is an example of the first seven players from the 2021 draft:
This is a very rich dataset, but we’ll only be focusing on a few columns. We want to know the impact of the round drafted (a proxy for draft capital) on how long a team sticks with a player. To measure this, we’ll use the “St” column, which tells how many seasons each player is the primary starter for a team.
It’s tempting to just immediately run a regression for the round drafted on the number of seasons started, but not so fast. There are a few variables we have to control for. First off, the year a player was drafted. Naturally, players from 2015 will generally have more years started than players from 2022; the latter can have a maximum of one year as a starter! So we include year variables in the regression to remove this impact (these are called “time-fixed effects”).
More notably, we want to control for the actual performance of a player. If we ignore player performance, we would get the result that the higher a player is drafted the longer they play, simply because higher-drafted players are generally better! We know this, and it’s not the question we are trying to answer. Instead, we want to know for two equally talented players, does being drafted in later rounds impact staying power? This isolates the effect of draft capital.
To measure value, we’ll use the wAV column in the PFR data, which is the “Weighted Career Approximate Value.” This is essentially PFR’s metric on how much value a player brings to their team over time; you can read more about it in their glossary.
I ran a regression for the round drafted on the number of seasons started, controlling for the year drafted and value added, for each of the major fantasy positions (QB/RB/WR/TE). Unsurprisingly, wAV has a large, positive impact on the number of years started across positions, which makes sense (better players play longer). But what about the round these players were drafted?
The major takeaway is that RBs drafted in the 1st round start about 0.8-1.0 seasons longer than RBs drafted in later rounds. This supports what we saw in the James Robinson vignette: an undrafted rookie being supplanted by a first-round pick! Interestingly, this is the only significant round effect; we don’t have evidence that RBs drafted in any other round have more or fewer seasons started. This bodes quite well for Bijan Robinson and Jahmyr Gibbs, the two backs taken in round one this year.
The other result of note is that WRs drafted in the first round start about 0.2-0.6 seasons longer than WRs drafted in later rounds. Once again, the first round is the only significant effect. It also looks like the effect is about half as big as the first-round impact for RBs
What about TEs and QBs, you ask? Interestingly, none of the draft rounds were significant, even the first. It’s kind of tough to draw many conclusions from this. For one thing, QBs and TEs have a much smaller sample size (especially quarterbacks). In addition, high-skilled QBs generally go early in the first round, and controlling for wAV accounts for all of the difference in longevity by round.
There seems to be clear evidence that a first-round RB will start for about a season longer than an equally talented RB drafted after the first. There’s also an effect for WRs, although only about half of the size. The takeaway: draft capital does matter for these skill positions, and adjust your rankings accordingly. On the flip side, don’t sweat it too much that QB Will Levis was drafted (just barely) in the second round.
Want to hear more? Let me know on Twitter.