The Draft Research Project: Starters and Impact Players

Last year, I rolled out the Draft Research Project, in which I looked at the first seven years of drafts under the new rookie salary scale, which substantially altered the economics of selecting and retaining football players in the NFL. With another season in the books, it is possible to add the data from 2018, as well. What this means is that there are now 2,035 players who were tracked across the first five years of their careers in the NFL. The good news was that most of the averages and ranges stayed appreciably the same, so that bodes well for year-to-year consistency (at least once a large enough pool of players was established). With eight years of drafts seeing second contracts at this point, I thought it would be interesting to look at the data in different ways.

Overview

First, just some general trends. On average, if a player was drafted, he appeared in 42 games but only start 22 (with a median of 46 games and 11 starts). Only 11% of the drafted players earned at least one Pro Bowl in his first five years (and only 4% earned a 1st-Team All Pro selection in that time).

I am defining a failure as a player with a career of three years or less, in the sense that those players failed to play out their first contracts. By this standard, 647 draftees (32%) failed. On the other hand, I am deeming as a success the players who earned a fifth year in the NFL (with their team or with another, showing that they succeeded at getting a team to reinvest in them); thus 1,178 players were success stories (58%). This obviously leaves about 10% of all drafted players in a middle category.

How Many Starters?

It should be noted that simply starting a game does not necessarily make a player good, it simply means that he was the person chosen to start. Sometimes it’s because he’s good, but sometimes it’s because the alternative is worse. Sometimes, likely, there is also pressure to start a player who was highly drafted regardless of whether or not he is the best available option from a football perspective.

Conventional wisdom says that teams should find three starters per draft, but is that really true? Probably, at least in the sense that teams need players to start games and the draft is the primary source of those players. More than a thousand players (1024) earned a “starter” designation from Pro Football Reference for at least a single season. That allows for exactly four starters per team per draft. However, just because a team designated a player as a starter does not mean that he was able to keep that position, and many of the players designated as starters did not hold that position once their teams had enough of a chance to replace them.

The definition of starter I used for the draft research project was much stricter, asking that a player started at least half of the games available to him in his first five years in order to be deemed a starter (with 40 starts as a consistent cutoff despite the growing season). It’s difficult to ask for starts from kickers, punters, and long snappers, but if we narrow the pool down to the remaining 2003 draftees, only 520 of them started at least 40 games in their first five years. That means that each draft contained an average of 65 regular starters, or basically enough for two starters per team per draft.

Pro Bowlers and Impact Players

If you want a Pro Bowler, you probably need to invest a high pick to get one. Of the 214 players drafted in these eight years to earn a Pro Bowl in their first five years, only 37 of them were taken with a pick outside of the first three rounds. However, the urgency in finding stars early is even greater than that. While the “average” Pro Bowler was taken in the second round and with pick #53, the median pick spent to select a Pro Bowler was actually #34. If you want this Pro Bowler to be more than a 1-year wonder–say, you want him to appear in 40 games–you don’t change the median pick or average round very much, but you do rule out four players (including two quarterbacks).

However, it is actually far more nuanced than that, because results vary substantially by position. Because even “position” can be debated, I defaulted to the drafted position listed on Pro Football Reference and then (if there was ambiguity) the player’s contract data as reported on Over The Cap. Under those definitions, here are the positions where the majority of the impact players (Pro Bowlers who appeared in at least 40 games) were drafted in the top 34 picks, compared to how many players at those positions were drafted afterward.

In each of these cases, there’s a noted imbalance in terms of players picked before and after the inflection point. Meanwhile, two positions are by contrast at essentially a dead heat, in that roughly as many impact players were drafted in the first 34 spots as were drafted with all of the remaining selections.

If someone wanted to combine guards and centers into interior linemen (as I have in the past) or corners and safeties into defensive backs (as many organizations do on their official rosters), then those more general groupings are also relatively balanced at the tipping point.

Finally, here are the positions where more of the impact players were actually found after the first 34 selections than before.

Now, it is obviously more than fair to point out that more picks are spent prospecting for these impact players in the later rounds, but the fact remains that those players are being reliably found. In fact, if the goal is to have impact players at as many positions as possible, teams would be better off prioritizing offensive tackle, center, defensive line, edge rusher, and cornerback somewhat urgently. Meanwhile, it’s a coin toss whether or not it’s worth spending a prime pick on guards or safeties. Finally, trends over these eight years show that impact-level off-ball linebackers, running backs, tight ends, and wide receivers are more readily available later in the draft, and therefore should be deprioritized (again, barring other information or goals).

As a simple illustration of this concept, consider the 24 players found at wide receiver, and edge rusher. Two-thirds of the impact edge rushers were taken before the 35th selection in each of their drafts, making the chances of finding one in the later rounds minimal, at best. By contrast, fewer than a third of the impact receivers were taken by then. This can also be broken down more specifically, and later articles in this series will do exactly that by position and by round. For now, however, here is another illustration of the same concept of prioritization spelled out in three tiers (based on days of the draft).

Which position or positions should be prioritized and which positions can wait? More impact-level running backs were found on the last day of these drafts than edge rushers, tackles, and wide receivers put together.

Note that this does not mean that taking a tackle early means he’s going to be a star. Rather, it suggests that by the time later rounds happen the star-level talent at the premium positions is already taken, while players with the potential to become stars are still there to be found later on. The reasons for this being true are likely varied, but for teams building through the draft, it’s clear that some positions need to be prioritized over others.

Next up, I will try to dive into position groups by round...