The Draft Research Project: Offensive Players by Position and Round

In the last article, I introduced the updated figures for the Draft Research Project, with eight years of drafts and just over two thousand players tracked across their first five years in the NFL. In that piece, I outlined the terms regular starter (a player who started at least 40 games across his first five years), impact player (a player who made a Pro Bowl in any way while also appearing in at least 40 games), and failure (a player who did not make it to at least a fourth season in the NFL. These three terms give a pretty effective snapshot of multiple dimensions of a player’s career. After all, even if he is not able to make a Pro Bowl or start games, if he is still in the NFL after four years it suggests that he has something to offer a team, at least, on some level.

Now it is time to introduce two more concepts. The first is Pick Cost. Put simply, the Pick Cost of a starter is equal to how many picks had to be spent to find a single starter. So, across the entire NFL in all rounds, players drafted at potential starter positions had a pick cost of 3.91 (because that’s how many picks it took to find a single regular starter in the NFL).

The second is even simpler, and that’s Likely Players. Simply put, these are the totals in the numbers of players found in each category divided by eight, to represent what a “typical” draft is likely to include; all totals are rounded to the nearest half player. The reason for this is to point out that while it is possible to find a starting tackle in the 7th round (i.e. Charles Leno), this is still not likely in any meaningful way. Betting on the possibility of finding such a player would be poor team-building strategy.

Quarterback

Overview: On average, 12 selections were made per draft, and based on historical trends there would be 2.5 likely starters per draft. Even though drafting a quarterback is traditionally seen as a “crapshoot”, it is actually fairly predictable. Of course, crapshoots are also predictable to people who just understand the math, and so the metaphor is probably more accurate than it seems. Rolling an ‘11’ is much less likely than rolling a ‘7’, and finding a starting-level quarterback in the fourth round is much less likely than finding one in the first.

Discussion: In the first round, it takes roughly three picks to find two starters, and the failure rate is remarkably low (8%); quarterbacks worth a first-round pick were often able to hang around in the NFL for at least four years. The profile is roughly what people might expect, in that if you do not get a quarterback early, you are likely not going to get one. There’s a weird eddy in the third round where quarterbacks last longer even though they don’t become starters, with the obvious conclusion that many of those players were as much planned backups (e.g. Mason Rudolph, Ryan Mallett, and Jacoby Brissett) as they were efforts to find a new starter.

Conclusions: If you want a starting quarterback–and especially an impact quarterback–you must take one in the first round. Otherwise you are just hoping to get lucky.

Running Back

Overview: By contrast with quarterback, running backs attracted 22 picks per draft with 3.5 likely starters per draft. The classic “devalued” position in the NFL, and that devaluing is upheld by virtually every metric I could think of to look at. Drafting a running back in the first round is the least efficient use of draft value I could find. Not only are running backs the ultimate boom-bust first round selections, there is hope for value later in the draft for teams that are patient.

Discussion: Obviously, the extraordinarily high failure rate (42%) jumps out, but it’s also worth pointing out that it took fewer first-round picks at quarterback to find a starter than it did picks at running back. Can’t miss running backs like Trent Richardson do, in fact, miss.

By contrast, later on in the draft things smooth out, and there is real efficiency to be found in the second and third rounds. The numbers suggest that there are likely two regular starters at running back available every draft by round three, and one will be taken in that round and the other will be discovered later. Meanwhile, even the non-starters found in the third round contribute fairly consistently, with a very low failure rate.

Conclusions: That failure rate for first-round running backs is not only the highest found, it is so much higher that it merits further analysis in the future. However, running backs seem like they are likely to contribute at least some value if drafted from Round 2 through Round 6, and there’s even one likely impact player left after round three every draft. This is the ultimate “should wait” position.

Tight Ends and Fullbacks

Overview: There were 14 picks and 3.5 likely starters per draft at tight end. Drafting a tight end in the first round happens only half as often as drafting a running back, and while it has a 0% failure rate, that’s another way of saying that a roughly equal number of first-round tight ends and first-round running backs succeed, but teams are smarter about not taking tight ends in the first round. Note that I included fullbacks in this grouping because they had to go somewhere, but they were strictly a late-round proposition (averaging being taken with pick #199 in the 6th round).

Discussion: There is probably only one impact player at tight end per draft, but there are probably two or three starter-level tight ends available after Round 1. Rounds 2-4 are where it’s possible to find the best efficiency, with returns all but disappearing after that point. Teams probably just shouldn’t draft fullbacks (13 selections made with 0 turning into regular starters or impact players and a failure rate of 54%; the only exception to the rule–Kyle Juszczyk–was actually drafted as a running back).

Conclusions: One of the disadvantages of a tight end taken early in the draft is that he is just as likely to underproduce as any position on offense, if not more so. One of the advantages is that even an under-producing tight end still provides some value, and the true “failure” rate seems very low. It can be a safe use of resources, but there is probably only a single impact player at the position available in every draft.

Wide Receivers

Overview: This position attracts more draft selections than any other on offense. In fact there were 32 selections and 6.5 likely starters per draft. Likewise, no other position on offense receives more first-round investment despite having the lowest rate of return on picks per impact player found. With the second-highest first-round failure rate (after running backs), the issue is not that there are not impact players available at wide receiver–there are. Teams are just very poor at finding them.

Discussion: There are likely three impact-level wide receivers per draft, and they can be found more or less equally across the first two rounds and then afterward. Round 4 jumps out, because wide receiver investment peaks there, even though the chances of finding an impact player or a even a starter there are almost non-existent. The failure rate peaks before settling back down. Very obviously, that’s where teams are starting to reach, or else gamble on prospects who fell for good reasons.

Conclusions: Teams should really take wide receivers in the second and third round. Especially compared to other positions, the position has a high rate of return. Waiting in this way is also safer than trying to find “the best guy” too early.

Offensive Tackle

Overview: Fewer tackles are taken in drafts than wide receivers and running backs, but more starters are found. With 19.5 selections made per draft (and 7 likely starters per draft), tackles are probably a sound overall investment of draft capital. For these tables, I used drafted position (and, when drafted position was general, Over The Cap’s contract information) when it came to offensive linemen, so it is worth pointing out that some of these players likely because interior linemen before all was said and done.

Discussion: There seems to be only one impact-level tackle per draft, and that player is usually gone in the first round. After that, finding an impact tackle is unlikely. However, it is reasonable to prospect for starters across the entire second day of the draft, in that the odds of finding a starter are still decent and especially in the second round the failure rate is low. However, after that point, it seems obvious that big men can stay on rosters if they stay healthy, but they are doing so only in a rotational capacity.

Conclusions: With fairly high failure rates in the last rounds and with a clear and steep drop in the likelihood of even finding starters, the tackle position is one of the most urgent positions to draft in the NFL. Almost every other position on offense still has about a 1:4 (or better) chance of finding an impact player in the second round–including quarterback. Tackle? 1:10.

Offensive Guard

Overview: At 15 selections made per draft with 5 likely starters per draft, Guards are one of the few positions where the decline from Round 1 to Round 2 is less noticeable, and there’s even a slightly greater trend of starters being found in the second round. Teams invest in guards about as often as they invest in tight ends, but guards produce starters at a slightly greater rate.

Discussion: What is interesting about the guard position is that there is a likely at least one starter per draft per round across the first three rounds, and while the first two rounds have a very low failure rate, the fact that they have such a good ratio of pick per performance is not a sign of there being more quality guards so much as it is a sign of how many fewer guards are drafted than tackles.

Conclusions: One of the big factors at work here has to be the rate at which tackles are converted to guards, but the fact remains that guards are solid investments for picks in the first three rounds, often outperforming other positions drafted on the second day, especially.

Offensive Centers

Overview: Centers receive little to no attention, with fewer than 7 picks spent per draft on the position in order to find the likely 3.5 starters available per draft. That means that there are roughly as many starters found every year at center as there are at tight end and running back, but that those selections are found with far fewer picks. Impact players at center are functionally nonexistent outside of the first round, with one important caveat.

Discussion: It is highly unlikely that NFL teams are two or three times better at finding centers than they are at finding tight ends and running backs, respectively. Instead, what is more likely is that teams are simply willing to settle for a lesser player at center than at those two positions and they are more willing to invest picks at the other spots trying to get an edge. However, a first-round center is likely to “pay off” in terms of staying on the roster and becoming an impact player, and the alternative is likely settling for just another guy–or converting a drafted guard.

Conclusions: The weird non-failure rate of 7th-round centers makes for a very interesting bookend to the high success level of 1st-round centers. This again suggests something about team building strategies on the parts of GMs as much as it does anything about the players available for the position.

Summary

One thing that might not be obvious from all of these tables is the urgency of each position relative to the other. Again, presuming that the goal is to have players of impact across all positions (instead of simply guys who can start), this table lists how many players are found in each draft, on average, who go on to both appear in 40 games and to earn at least one Pro Bowl.

Teams that put off investing in the offensive line in favor of other positions are highly unlikely to find impact players at those positions later in the draft. The players simply aren’t there. By contrast, teams can afford to wait on tight ends and receivers. There is no reasonable reason to rush on finding running backs, because even if a team believes it must have a Pro Bowl-level talent there, they are the most common and replaceable offensive players at every level of the draft.

In the next piece, defensive positions will be evaluated.