Ron Rivera exceeded most reasonable expectations in his first year as head coach and face of the new-look Washington Football franchise, leading last year’s train wreck to a division championship and giving the eventual Super Bowl champions their hardest playoff contest. Even as WFT fans were celebrating the better than expected finish, Rivera began to make it clear that to take the team much further he will need to find a franchise quarterback.

Ryan Fitzpatrick was signed a few weeks ago to provide a veteran presence. At 38 and starting for his ninth team, Fitzpatrick is clearly not the quarterback of the future. His signing does set the team up to draft and develop a quarterback, by relieving the pressure to draft a day one starter, or start a rookie too early. That is the path that most of the league’s best teams have followed to find their franchise QBs.

In 2011, Rivera took over as head coach of the 2-14 Carolina Panthers. His first major move that season was drafting Cam Newton with the first overall pick. Within two years, Rivera and Newton led the Panthers back to the playoffs. In 2015, the Panthers were back to the Super Bowl.

In 2021, as Rivera looks for his next franchise QB, he doesn’t have the luxury of holding the first overall draft pick. Picking 19^th, it is highly unlikely that he will have a shot at any of the top three “elite” QB prospects: Trevor Lawrence, Zach Wilson and Justin Fields.

The fourth-rated prospect on most draft boards is Trey Lance. The dual threat sophomore QB has impressive physical tools, but he has one year of starting experience in the Division II Missouri Valley Conference, and is still very raw. Scouting reports cite a need to develop his field vision, ability to work through progressions and speed up the process. While many draft analysts give Lance a second-round grade, the supply vs. demand economics of the QB position could push teams to draft him well before the WFT’s 19^th pick. The next QB on most boards, Mac Jones, generally gets late first to second-round grades.

Should the WFT pick the best QB available at #19, use that pick to target an impact player at another position, or perhaps trade up to draft one of the top-rated prospects? My last article examined the historical outcomes of trades into the top 10 for quarterbacks, which is likely what it will take to land one of the top three QB prospects this draft, and possibly Lance as well. It has typically been the league’s worst teams that make those trades, and the outcomes have seldom met expectations.

Why is that? I think it is likely because bad teams fail to comprehend a basic feature of the draft, which better teams get. Teams trading up in the top 10 picks of the draft are willing to incur large opportunity costs by giving away additional high-value draft picks, which could otherwise be used to target more potentially high-impact players. It is likely that they are willing to take that risk because they are absolutely sure that the QB they are targeting is “the guy”. They shouldn’t be, because evaluation of NFL draft prospect does not lend itself to precise projections, at all.

Previously, I have discussed the draft in terms of two kinds of systems. The behavior of highly deterministic systems is governed by a few, readily measured variables. If you know what the key variables are, and do your measurements correctly, you can predict the outcomes of deterministic processes with a high degree of precision. If an outcome differs significantly from your prediction, then you probably got one of your measurements wrong. If scouting quarterbacks was highly deterministic, then we would not regularly see outcomes like third-round pick Russell Wilson becoming the best QB selected in the deep 2012 class, or the fifth QB selected in 2018, Lamar Jackson, vastly outperforming much more highly rated prospects in his draft class.

When teams trade away multiple first and second-round draft picks to chase their QB of the future, they are behaving as if the draft were a deterministic system. The opportunity risk of giving away multiple first and second-round picks could only be tolerable if they are highly confident in their evaluation of the targeted QB.

In fact, as we are reminded year after year, it is a highly probabilistic system. In probabilistic systems, it is difficult to predict outcomes with any kind of precision, because important variables are difficult to measure, like prospects’ character, and/or can’t be measured, such as how a college spread QB will adapt to NFL route progressions.

This is not to say that NFL scouts and some media analysts are not experts at what they do. Draft evaluations are accurate, because as a group QBs given a first-round grade will do better than QBs given a third-round grade. But they are highly imprecise, since some QBs with a third-round grade will outperform QBs with a first-round grade, and some QBs with a first-round grade will do worse than QBs with a third-round grade.

To illustrate the probabilistic nature of drafting quarterbacks and explore what that means for Ron Rivera’s options with the 19^th pick, in this article I will take a historical stroll through the outcomes of the last 10 QB draft classes. Along the way, I will examine questions such as how often the top QBs in the draft have turned out to be best in class, how often the best QB has been picked later in the draft, and what chance does the WFT have of drafting a franchise QB with the 19^th overall pick.

Photo by Rich Graessle/Icon Sportswire/Corbis/Icon Sportswire via Getty Images

---

Methodology and Stats

My methodology this time is very simple. For each draft from 2011 to 2020, I listed the top two to five QBs selected, ranked according to overall production. In each draft I listed any long-term starters, short-term starters, and if there weren’t enough of them, quality backups in the top part of each table. In the bottom half of the table, I listed any of the first five QBs selected who failed to become long term starters. Long-term starter is defined as a QB who signed a starting contract following his rookie contract.

To rank QBs by overall production, I used Pro Football Reference’s statistic weighted Career Approximate Value (CarAV). CarAV is an attempt to place a single number on a player’s career value, equalized across positions. It is the sum of the player’s individual season AV values, weighted to emphasize peak seasons. AV is pretty complicated to explain. Interested readers should refer to this page.

CarAV has some limitations when comparing players at different positions and when comparing players from different draft years who have been in the league for less than around five years. However, it performs much better when ranking players at the same position from the same draft year, as I am doing here.

I also provide Adjusted Net Yards per attempt (ANY/A) statistics for each QB as a metric of passing efficiency. ANY/A is considered by many analytics experts to be an improvement over the more commonly used official NFL statistic Passer Rating. It is calculated as follows:

ANY/A = (pass yards + 20*(pass TD) - 45*(interceptions thrown) - sack yards)/(passing attempts + sacks)

Like other passing performance metrics, the general idea is to incorporate all the good (passing yards, TDs) and bad (interceptions, sacks) things that QBs contribute to the passing game into a single metric. One could fault ANY/A for failing to take into account a quarterback’s contribution to the rushing game. But that same criticism also applies to Passer Rating.

For readers who are unfamiliar with ANY/A, some general benchmarks that I’ve found from working with numbers from the last decade are: values below about 5.0 are below standard for the NFL, values between 5.0 and 6.0 represent the transition from back-up to starting-level performance, and values above 7.0 represent the transition from very good to elite passing performance.

In some cases the rank order determined by CarAV and ANY/A can differ substantially. Where it does, I have gone with the ranking by CarAV because it tends to provide better agreement with the consensus view of who the best QBs in each class were. ANY/A is just provided to give another perspective on these QBs’ performances.

Finally, the statistic Draft Order refers to the order of QBs taken in any particular draft. The first QB off the board has a Draft Order of 1 and the fifth QB selected has a Draft Order of 5, regardless of the round or pick number where they were selected.

---

The Drafts

2011

Heading into the 2011 draft, Cam Newton dominated draft discussions in the media as one of the more polarizing prospects in recent times. Draft analyst ratings ranged from best overall prospect to no chance of starting in the NFL. Amongst analysts that did not rate Newton as the best QB prospect, Blaine Gabbert and Jake Locker were considered possibilities as the first player selected.

Only two QBs selected in this draft went on to become long-term starters, Newton and Andy Dalton picked 35^th overall. Colin Kaepernick, picked right after Dalton, had an electric start to his career, leading the read-option craze of the early teens, and posted decent numbers in six years with the 49ers before his NFL career ended in controversy. After Kaepernick, Tyrod Taylor had a three year starting stint in Buffalo, but otherwise has been a quality backup. Locker and Gabbert, picked in the top 10, were major busts.

2012

This was the deepest draft of the decade for producing starting QBs. Andrew Luck was considered a generational prospect ahead of the draft and was a rare consensus number one overall prospect. After being selected first overall, he had a decent career starting for Indianapolis, which was unfortunately cut short due to injuries.

The cream of this QB crop, Russell Wilson, was the sixth QB off the board, selected with the 75^th pick in the third round, 73 picks and four QBs after consensus number two prospect, RG3. Third QB off the board, Ryan Tannehill, had a decent start to his career with Miami, and has turned into one of the NFL’s top QBs since moving to Tennessee in 2019. The eighth QB selected, Kirk Cousins, has posted good numbers as a starter for the Redskins and Vikings.

Dan Snyder’s pick, RG3 became the most costly bust of the salary cap era. The other QB selected in the first round, Brandon Weeden, also failed to live up to expectations. It is not clear whether any team other than the Cleveland Browns expected much of him, though.

2013

No long term starting QBs were selected in this draft. The best QB in class was second-round pick Geno Smith, with a CarAV of 13. Nothing to see here.

2014

This draft has only produced a single long-term starter to date, Derek Carr. He was the fourth QB off the board, selected near the top of the second round. After a three year stint as Tom Brady’s backup, Jimmy Garoppolo has posted good numbers in limited starts in San Francisco, but the 49ers’ trade up to the third overall pick of the 2021 draft would seem to indicate he has reached the end of the line there. The tragically over-drafted first QB off the board, Blake Bortles, lost his starting job after five years in Jacksonville. Following a preseason injury in 2016, Bridgewater has taken on a role as a backup and injury-replacement starter.

The consensus most-likely-bust prospect, Johnny Manziel, did just that after being selected 22^nd overall by Cleveland. Once again, only Cleveland was surprised.

2015

Jameis Winston was considered by many analysts to be a rare QB talent ahead of this draft, but there were a fair few who thought that Mariota might even be better.

It turned out to be the second draft in the decade to not produce any long term starting QBs. The first and second QBs off the board, picked first and second overall, were both released by the teams that drafted them following their rookie contracts and signed with other teams as backups. The most notable QBs selected after these two were Trevor Siemian, picked 250^th, and Brett Hundley, picked 147^th. Neither of those two are really that notable.

2016

The majority of draft analysts rated Goff as the top QB prospect in this draft, but Division II prospect Wentz generated his fair share of dark-horse buzz, as well. Picked first and second overall, they were both outshone by the eighth QB off the board, Dak Prescott, selected in the fourth round. Goff has had a good starting career in LA, helping to turn a losing team into a regular playoff contender, highlighted by a Super Bowl appearance in his third year. But apparently not good enough for the Rams, who decided to trade him for Matt Stafford a few weeks ago. Wentz had a promising start to his career before suffering a sharp decline in play in 2020, prompting Philadelphia to trade him to Indianapolis for a third-round pick. The fourth best QB in this class, Jacoby Brissett has had a career as a backup and spot starter. The third and fourth QBs selected in this draft, Paxton Lynch and Christian Hackenberg could be considered minor busts.

2017

In my pick for the most interesting QB draft of the decade, the Chicago Bears traded up one spot to make sure they got their guy, one-year starter Mitch Trubisky, at second overall. This move seems comical in hindsight, given who they left on the board. But it was not seen as crazy at the time at all. Trubisky was widely considered to be the favorite to be the first QB taken. NFL.com’s Bucky Brooks even called the pick number correctly. Trubisky has just begun the second phase of his career, as a backup to Josh Allen in Buffalo. After Mahomes and Watson were selected second and third, no more notable QBs were selected in this draft.

2018

Lamar Jackson is a good example of how the fifth QB off the board can outperform the first QB selected. Of course, that statement might be a little misleading because Baker Mayfield was a surprise pick at first overall by the Browns. The top-rated QB on most analysts’ boards heading into this draft was Sam Darnold. Cleveland seems to have got that one correct, since Baker Mayfield has proven to be a worthy starter, whereas Sam Darnold has yet to demonstrate the same. Based on the reported trade interest in Darnold , it appears that many front offices might think the issue is the team, not the QB.

Arizona also traded up to acquire the fourth QB in this draft, Josh Rosen, who has turned into a spectacular bust.

2019

We are starting to get into “too early to tell” territory. After his second season in the league, it appears that Kyler Murray was drafted appropriately, despite concerns of some analysts that he might be too small to play in the NFL. The second QB selected, Daniel Jones, is in a virtual dead heat with the tenth QB taken, Gardner Minshew, in terms of overall production (CarAV) at this point in their respective careers. But Minshew has a clear advantage in terms of passing efficiency (ANY/A). The fourth QB selected, Drew Lock, struggled early on in Denver, as do many rookie QBs, but he still has plenty of time to develop into a starter.

Dwayne Haskins, selected third in this QB class, was an egregious bust. That might have something to do with reports that he was selected by the owner, against the better judgement of his draft experts.

2020

After just one season in the NFL, it is clearly too early to form a meaningful opinion of many of the QB’s selected in 2020. The first QB selected, Joe Burrow, looked like a rookie on his way to becoming a legitimate starter before going down with an injury in game #10 against the WFT. But he and second drafted Tua Tagovailoa were upstaged by the third-selected QB, 2020 AP Offensive Rookie of the Year, Justin Herbert. Jalen Hurts selected fifth did okay for a rookie in his first season. While fourth-selected QB, Jordan Love, held the clipboard for Aaron Rodgers and is a complete unknown at this stage.

---

Photo by Jonathan Newton/The Washington Post via Getty Images

Some Answers

To recap, the seven draft classes from 2011 to 2017 produced 12 long-term starting QBs, defined as a QB who signed a starting contract following his rookie contract, for an average of 1.7 per draft. It is still too early to call some of the QBs drafted from 2018 to 2020, but it appears that the 2018 draft produced three, 2019 could is looking like one to two, and early returns from 2020 are looking like two to four. That gives a total estimated yield of somewhere between 18 to 20 long-term starters, for an average of 1.8 to 2 per draft. The deepest QB class in that period, 2012, produced four long-term starters, while 2013 and 2015 failed to produce any.

That decade average is small compared to the number of QBs projected to go in the first round this April (four to five), suggesting to me that some of the top-rated prospects in 2021 draft won’t make it in the NFL. Predicting which ones those are is the hard part.

How often is the first QB selected the best in class?

Out of the last 10 drafts, only the 2013 class failed to produce any QBs that started for at least the duration of their rookie contracts. In the nine draft classes that did produce starters, the best QB was the first QB selected three times (2011 Cam Newton, 2015 Jameis Winston, 2019 Kyler Murray), for a hit rate of 33%. I would hazard a guess that’s probably lower than teams selecting QBs in the early part of the first round might have hoped for. It is also a little inflated, because, while he was the best in a poor class, Jameis Winston did not become a long-term starter.

In each of these cases, the best QB in class was the first overall pick in the draft. In 2017, the best drafted QB of the decade, Patrick Mahomes, was the second QB off the board. In the remaining five of nine draft classes (56% of the time), the best QB was selected third or later in the QB cohort.

In summary, in the last decade it has been more common for QBs selected third or later in their cohort to be the best in class than the first or second QB off the board.

Which part of the draft produced the best QBs over the last decade?

One way to answer this question is to list the pick numbers where the best QB in each draft class was taken. In ascending order, that looks as follows: 1, 1, 1, 6, 10, 32, 36, 75, 135. By that account, five of nine (56%) of the best QBs drafted in the past decade were taken in the top 10, and four of nine (44%) were drafted with the last pick of the first round or later.

But I don’t think that gives quite the right answer, because that approach equates elite best in class QBs like Patrick Mahomes and Russell Wilson with the QBs in weak draft classes, like Jameis Winston, who was borderline starter quality.

Another way to tackle that question is to ask where the elite QBs were selected. In my purely subjective estimation, the established elite QBs drafted in the past decade are: Russell Wilson, Patrick Mahomes, Deshaun Watson and Lamar Jackson. Based on early returns, it is possible that some more recently drafted QBs might eventually join their ranks, such as Josh Allen, Kyler Murray, Justin Herbert, and Joe Burrow. However, for now I will stick with the established elite QB list. Those QBs were selected at pick numbers: 75, 10, 12 and 32. None of the elite QBs was selected in the top nine picks, but that could change if any of the more recently drafted QBs claim elite status.

Where were quality starting QBs taken in recent drafts?

Of course, focusing on just the elite QBs results in small sample sizes, by definition. That could be an issue, since drawing conclusions from small samples can be misleading. To broaden the sample, it might be better to look at long-term starters, and recently drafted QBs who are showing signs of becoming long-term starters.

That definition gives me the following, list with draft pick numbers in parentheses: Cam Newton (1), Andy Dalton (35), Russell Wilson (75), Ryan Tannehill (8), Kirk Cousins (102), Andrew Luck (1), Derek Carr (36), Dak Prescott (135), Jared Goff (1), Carson Wentz (2), Patrick Mahomes (10), Deshaun Watson (12), Lamar Jackson (32), Josh Allen (7), and Baker Mayfield (1). I think it’s also safe to add Kyler Murray (1), Justin Herbert (6) and Joe Burrow (1) from the last two drafts to the list. Some others could emerge, but I’d prefer not to speculate more than I already have.

The list of long-term starter pick numbers is as follows: 1, 1, 1, 1, 1, 1, 2, 6, 7, 8, 10, 12, 32, 35, 36, 75, 102, 135. Six of the last ten drafts have featured a QB picked first overall who went on to become a long-term starter, or is showing signs of doing so.

Eleven of eighteen (61%) long-term starting QBs drafted in the last decade were selected in the top ten picks of the draft. Only 2/18 (11%) were selected after pick #10 in the first round. The remaining 5/18 (28%) were selected in the second through fourth rounds.

In summary, 61% of long term starting QBs were drafted in the top 10 picks of the draft and 39% were selected later. It is worth mentioning here that no undrafted free agent has become a long term starting QB in the past decade.

What is the chance of the WFT drafting a starting quality QB at #19 or later?

I might be in the minority, but I think that NFL front offices are skeptical of one year starters, and that could cause Trey Lance to be available at the WFT’s 19^th overall pick as the fourth QB in the class. Of course, it only takes one GM to fall in love, in which case there is a very good chance that Mac Jones will be available in the fifth position.

Since 2010, six quality starting QBs have been drafted as the fourth QB in their class or later: Andy Dalton (5^th in class), Russell Wilson (6^th), Kirk Cousins (8^th), Derek Carr (4^th), Dak Prescott (8^th), Lamar Jackson (5^th).

While it is still too early to tell, Gardner Minshew, drafted 10^th in his class, might eventually join their ranks. I suppose it’s possible that Jalen Hurts (5th) could emerge as a quality starter, and who knows what to make of Jordan Love (6^th)? But to be safe, I’ll call that six starting quality starting QBs drafted fourth or later in their class in eight drafts from 2011 to 2018. If that history is any guide, there is a six in eight chance (75%) that a starting quality QB will be selected after the third QB comes off the board in 2021, and a five in eight chance (62.5%) that one will be selected after the fourth.

What is much harder to estimate is the chance that the WFT picks that guy at #19 and not someone else. Things get a bit gloomier when we consider the bust rate of the fourth and fifth QBs selected in the nine drafts from 2011 to 2019 (excluding 2020, which is too recent). Eight of the nine (89%) fourth-in-class QBs did not cement starting roles, although it’s still possible that Drew Lock (2019 class) might, dropping that down to seven of nine (78%). Things are about the same or slightly better at the fifth-in-class position, where seven of nine (78%) drafted QBs failed to become long-term starters.

---

What It All Means

In conclusion, it is not impossible for the WFT to draft a starting QB with the 19^th pick in 2021. However, given the relatively low hit rate with the fourth and fifth QBs off the board in recent drafts, which is most likely what they will be looking at, they might be better advised to use that pick to target a high impact player at a less highly valued position, such as linebacker, defensive back, receiver or offensive tackle.

Other good options are to trade back to add picks and possibly target a QB with a late first- to second-round grade, or using a less valuable later-round pick to take a chance on a developmental project, like Jamie Newman or Davis Mills.

But what about trading up into the top ten to target one of the top QB prospects in 2021? I have previously shown that trading up into the top ten to target a more highly ranked QB prospect has nearly always backfired, and has a high opportunity cost through the loss of additional high value draft picks.

The examination of recent QB drafts in this article shows one reason why that is. While a (slight) majority of the quality starting quarterbacks have been selected in the top 10 across all the drafts, in any particular draft class it can be difficult to predict whether the best QB will be selected first, third, fifth or later. And the flip side of that is that it is equally hard to predict which of the first two to four QBs in a class will be a total bust.

Acknowledgements:

As usual, this article was greatly improved by editorial and substantive input from James Dorsett. Football statistics CarAV, ANY/A and player draft order was sourced from Pro Football Reference.

---