Is “Draft and Develop” a Myth? Part 1: Offense

Feb 6, 2023 - 8:00 PM

<a href= — Photo by Justin Casterline/Getty Images

I recently had a look at the probabilities of addressing the Commanders’ major needs for the 2023 season in the draft. The results confirmed what people who follow the NFL draft should have expected. There is a very good chance of drafting a player in the first round who will be ready to play this coming season. The chance of drafting a first-year starter drops by about half with each successive round. What that means is that after the first, you are really picking players for next year or beyond and adding depth.

At least, that seems to be the obvious conclusion. While I was compiling the data on offensive tackles, I noticed something that didn’t quite fit the simple narrative. Quite a few players I was looking at who were drafted in the first round started as rookies, but then quickly faded into backup roles. Conversely, players picked in the middle rounds who became eventual starters seemed to stick once they cracked the starting lineup.

This raised a few questions. How much of an influence does draft position, rather than talent, have on players’ opportunities to start early in their careers? If my basic assumptions about players’ developmental timelines were wrong, what effect would that have on my conclusions about the probabilities of drafting starters in different rounds? To answer those questions and more, in this two-part series I will have a look at the timeline of drafted players entering and leaving starting lineups in the first five years of their NFL careers. In Part 1, I will have a look at the Commanders’ three biggest needs on offense: Quarterback, Offensive Tackle and Interior Offensive Line.

Estimating Probabilities of Drafting Starters

To estimate probabilities of drafting starters, I simply determined the proportion of players drafted at each position, over the decade from 2009 to 2018, who started 10 or more games in the first five years of their NFL careers. Data was sourced from Pro Football Reference. Starting 10 games, as opposed to 16, in a season was chosen as the criterion to avoid missing starters who lost time to injury or who took a few games to crack the lineup in their rookie seasons.

There is a fair bit of slop in the database with regard to specifying the offensive line positions, which is at least partially a reflection of the slop that exists in the NFL. For example, college offensive tackles frequently switch to guard in the NFL. Occasionally players drafted as guard have switched to tackle, and many players switch between guard and center. As a result I was unable to cleanly split players within position groups. Therefore, for purposes of this article, I have grouped right and left tackles together as Offensive Tackles (OT) and guards and centers together as Interior Offensive Line (IOL).

Also, to account for position switches from college to the NFL, offensive linemen were classified by their draft positions and counted as starters if they started more than 10 games in any season at any offensive line position.

Quarterbacks were much more simple. A player was counted as a starter if he was drafted at QB and started 10 games in a season at QB. Apologies to Logan Thomas.

Offensive Tackle

I’ll start with OT because that’s the position that first drew my notice, and also because the data is the cleanest, which makes it easiest to see the major trends. The estimated probabilities of drafted OTs becoming starters in the first five years of their NFL careers are shown in the following figure:

I initially came to this analysis with the simple assumption that development of drafted players in NFL programs over time would result in more players entering starting lineups in each year after they are drafted.

This “draft and develop” concept is borne out by the results for OTs drafted in the third and fourth rounds. Aside from a little jitter for player selected in the third round, the overall pattern for middle-round OTs is that more players become starters in each of the first four to five years of their careers. The effect of development time is most pronounced in the third round, where the proportion of starting OTs increases by 2.5 times from 16.7% in their rookie seasons to 41.7% in their third seasons in the league.

An apparent development effect is also seen between OTs’ rookie seasons and their second seasons at every part of the draft except the second round. However, for players drafted in the first round, the development effect is rather short-lived. By the fourth year of their careers, the proportion of first-round OTs that are still starting starts to come back down. This effect is most dramatic for OTs drafted at picks 1 through 16, where the proportion of starters drops from 91.7% in their third years to 66.7% in their fourth years.

A similar pattern is evident for OTs drafted in the 5^th through 7^th rounds, where the proportion of starters peaks at 17.9% in their second seasons and then falls for the next three seasons.

The pattern for players drafted in the second round is more complex, with the highest proportion of starters in the first year, followed by a peaked development timeline in years two through five. I’m not sure what that means. It might just be a flat distribution with a bit of random variation between years.

To sum up, in the third and fourth rounds, more OTs become starters each year after they are drafted, which is consistent with the idea that they are developing with time in NFL programs. In the first round, however, there is a little evidence that some players are developing from their first to second seasons. After that, the proportions of starters comes back down. I’d be tempted to speculate that this could mean that players drafted in the middle rounds have to earn their starts, while some players drafted in the first round are given early starting opportunities due to their draft status and lose them over time due to their play on the field. However, if this was all the data I had, I’d probably conclude that I’m just overanalyzing two little peaks on a noisy graph.

Another large scale trend is apparent in these data. In my previous article, I only looked at first-year starters, since the focus was on addressing immediate needs. That narrow focus gave the impression that the probability of drafting starting players drops dramatically from the first round to the third and fourth rounds. However, the present analysis reveals that, by the third through fifth years of players’ careers, that gap has closed somewhat due the development of players drafted in the middle rounds, as well as the apparent regression of some players drafted in the first round.

Therefore, teams with the patience to wait for players to develop give themselves a decent chance of drafting starting OTs as late as the third round.

Now, let’s have a look at quarterbacks.

Quarterback

It should come as little surprise that teams looking to draft starting quarterbacks should take their shots in the first round. Not just the first round, but the first half of the first round. Based on these data, there is slightly better than 50% chance of drafting a QB who is still starting in his fifth year in the top 16 picks. That number drops to 30% in the second round.

The data on QBs gets pretty sparse after around pick #16, because only ten out of 94 players drafted after that were starting by their fourth seasons. That’s OK, because the main point I want to make relates to the players drafted in the first round.

The main thing that stands out is that the effect we got hints of with OTs drafted early in the first round really stands out with QBs. The proportion of QBs drafted in the first 16 picks who get starts increases from 83% to 87% from Year 1 to Year 2 and then drops to 52% by their fifth year in the league. This is consistent with the idea that QBs drafted early in the first round get starting opportunities early due to their draft status, rather than by earning their starts with their play on the field.

Unlike OT, there is little evidence at the macro level that QBs earning starting time with development is a major trend at any point in the draft. Of course, data from QBs like Patrick Mahomes and Lamar Jackson is in there. But it appears to be drowned out by the larger number of QBs who became starters early and then lost their starting jobs.

There is an apparent spike in the proportion of starters from years one to two at picks 17 to 32. But that is only going from two (Brandon Weeden, Teddy Bridgewater) to four starters (Bridgewater, Josh Freeman, Lamar Jackson, Tim Tebow), so I wouldn’t place too much stock in it.

Once again, there is an odd phenomenon in the second round, where the proportion of starters peaks in Year 1. This is based on five starters out of ten drafted players, so it might be real and not just a glitch due to small sample size. The most likely explanation would seem to be that teams are quicker to give up on the QBs they drafted in the second round than the ones they drafted in the top 16.

Interior Offensive Line

That brings us to the last position group, IOL. One major difference from the two positions we have looked at is that guard and center are not considered premium positions in the draft. That is clearly reflected in the numbers of IOL selected early in the first round. From 2009 to 2018, 23 QBs and 24 OTs were selected in the first 16 picks, but only five IOL were picked in that range. Let’s see if being a non-premium position has any impact on the development timelines.

What really stands out here is how flat the graphs are through the first and second rounds. Overall, the probability of drafting a starting IOL is not that much different in the top 16 picks and the second round.

As I mentioned above, the sample size of IOL picked in the top 16 is really too small to draw any meaningful conclusions from. However, the samples of IOL drafted 17 to 32 and in the second round are 16 and 22 players, respectively. So I would say the lack of major difference between the second half of the first round and the second round is probably real.

In the early two rounds, the proportion of starters peaks in Year 1 and then falls in Years 2 and 3. It comes back up in Years 4 and 5 in the second round.

The proportion of drafted starters then falls in the third and later rounds. From the third round onward the proportion of starters increases from Year 1 to Year 2, but falls again (rounds 3 and 4) or remains about the same (rounds 5 through 7) after that.

Overall, as far as development timelines go, the two major trends with IOL are: 1. Players taken early in the draft tend to start as rookies and then the proportion of starters dips in Years 2 and 3; From round 3 onward, the proportion of starters increases from Years 1 to 2 and then declines or stays the same. Beyond that, there is not much evidence that players earning starting time through development is a major trend.

I have two major take-home points from the IOL data. First, the best value for IOL is in the second round. Using an early first-round pick on a “generational” guard would seem to be a waste of draft capital, since you are just as likely to find a plug-and-play, five-year starter in the second round. Second, after the second year, you should generally have a good idea about whether an IOL is a starter or not.

Conclusions

The main takeaway from this analysis was that, at the aggregate level, development of players in the first five years of their careers is not a general trend across the board. It does appear to be the norm for players taken in some position groups in some parts of the draft, but not others.

Of course, this type of coarse analysis does tend to obscure individual players’ stories. Even where development isn’t evident at the group level, there will be individual players who buck the trend but greater numbers of players are likely to be starting early and losing their starting positions.

There was also considerable variation between parts of the draft and position groups. There was clear evidence of player development over time being a major trend for OTs drafted in the middle rounds. Similarly, IOL drafted in the third round and later showed overall development trends from Year 1 to Year 5, with some dips and troughs along the way. Conversely, QBs drafted in the top 16 picks, tended to start in their first two seasons and then lose their starting jobs from years three to five. A similar, but less pronounced effect is evident in OTs drafted in the same range.

In conclusion, the idea that drafted players tend to develop and become starters in the first few years after they are drafted is not a myth. It’s just overstated. At two of the three positions I examined, it tends to occur more often for players drafted in the middle and later rounds. While players drafted in the first and second rounds tend to start early and then some of them lose their starting jobs over the next few years.

Acknowledgement: Edited by James Dorsett

No one has shouted yet.
Be the first!

Are You Watching This?!