≡ Menu

A couple of years ago, I asked how long it should have taken the Jaguars to move on from Blaine Gabbert. Today I want to revisit that general idea, but look at how long it takes the best quarterbacks to identify themselves as top-tier players. A couple of months ago, I looked at the greatest quarterbacks of all time. Using the top 75 quarterbacks from that list, I removed any player whose career began before the merger; that left me with 42 passers.

First, I looked at how each quarterback fared in relative Adjusted Net Yards per Attempt — i.e., ANY/A relative to league average — through their first 16 starts. Just over two-thirds of these passers were above average during their first 16 starts, with 1/3 of those quarterbacks being at least 1 ANY/A better than league average.  That group of fourteen quarterbacks — which Aaron Rodgers just falls shy of joining — can be categorized as above-average quarterbacks from the beginning. They are Kurt Warner, Dan Marino, Daunte Culpepper, Chad Pennington, Tony Romo, Mark Rypien, Jeff Garcia, Boomer Esiason, Ben Roethlisberger, Philip Rivers, Matt Ryan, Joe Montana, Steve McNair, and Ken Stabler. Obviously a number of those quarterbacks were not immediate starters in the NFL, but they did excel as soon as they became starters.

The graph below shows each of the 42 quarterbacks’ Relative ANY/A through their first 16 starts. The X-Axis represents the quarterback’s first year, and the Y-Axis shows their RANY/A value through 16 starts.

QB breakout 1

Now, let’s remove the 14 quarterbacks who had a RANY/A of at least +1.0 through their first sixteen starts. How did the other 28 quarterbacks fare in starts 17 through 32 in RANY/A? Eleven of them produced a RANY/A of at least +1.0 in their next sixteen starts: Bert Jones, Matt Schaub, Ken Anderson, Peyton Manning, Aaron Rodgers, Brad Johnson, Carson Palmer, Jim Everett, Steve Young, Dan Fouts, and Steve Grogan.

QB breakout 2

Just eight of these quarterbacks had below-average RANY/As in their second set of sixteen starts. Troy Aikman, Vinny Testaverde, and Terry Bradshaw were former number one overall picks who were drafted by very bad teams; they all started at least four games as rookies, and produced below-average numbers through both starts 1-16 and starts 17-32.   Michael Vick was a former number one overall pick who saw some immediate success — he had an above-average RANY/A through 16 starts (but was below +1.0 RANY/A), not to mention significant production with his legs — but he has had an up-and-down career.1

Warren Moon, Brett Favre, and Joe Theismann also produced slightly above RANY/A numbers through 16 starts, and then below average RANY/A averages in starts 17 through 32.  Other than Aikman/Testaverde/Bradshaw — and I think they deserve to be in their own group — that leaves just Drew Brees as the only quarterback in our sample to be below average through both his first and second set of sixteen starts.

If we eliminate the 11 quarterbacks who had RANY/A averages of at least +1.0 in starts 17 through 32, that leaves 17 quarterbacks who turned out to have very good careers but failed to be far above average through their first 32 starts. How did they fare in starts 33 through 48?

QB breakout 3

Here we see Drew Brees finally stand out from the pack. In addition, Mark Brunell, Trent Green, and Troy Aikman were at +1.0 RANY/A in starts 33 through 48.

That leaves thirteen quarterbacks who failed to be significantly above-average during any of their first three 16-game samples: Brett Favre, Tom Brady, Joe Theismann, Dave Krieg, Jim Kelly, John Elway, Donovan McNabb, Vinny Testaverde, Terry Bradshaw, Randall Cunningham, Warren Moon, Michael Vick, and Rich Gannon.

We’ve already discussed Bradshaw, Testaverde, and Vick: talent was never the issue with these quarterbacks, but a combination of factors prevented them from seeing immediate success. What about the other ten?

  • Brady finished as slightly above-average in RANY/A in each of his first three seasons as a starter. He also picked up a pair of Super Bowl rings. This was back in the “Brady is a winner with mediocre stats” days. It’s safe to say he had “broken out” by most definitions of the term by the end of the 2003 season, but it wasn’t until 2004 that he had his first great statistical season, and it wasn’t until 2007 that he had his first Tom Brady(TM) level year.
  • Theismann was drafted in 1971, but went to the CFL for three seasons. He barely played with Washington in ’74 and ’75 (although he did return punts), and then started 5 games in ’76 and 6 in ’77, before becoming the full-time starter beginning in 1978. His 1979 season was very good — he ranked 8th in ANY/A — but the cutoffs chosen prevent him from standing out in this analysis. Theismann had only one truly great season of his career, which came in 1983, although he had other solid seasons in his career. He’s a unique character in that he didn’t make his first Pro Bowl until age 33, and then was out of the NFL at the age of 36 (no link required).
  • Krieg, like Brady and Favre, was above-average (but below +1.0 RANY/A) in each of his first three sets of 16 games. After sitting as a rookie, Krieg started 45 games from 1981 to 1985, going 27-18 and posting strong (but not overwhelming) numbers. He was even better for the rest of the ’80s, and had a very good career, albeit one that puts him a few tiers below the Brady/Favre level. Could you see Dalton turning into this generation’s Krieg?
  • Kelly’s numbers hovered just north of league average during his first three years, but frankly, his career statistics are not commensurate with his reputation (albeit without adjusting for the fact that he played half of his games in Buffalo). His 4th and 7th years were very good, and his 5th and 6th years were great, but that was about it; of course, Kelly also lost three years to the USFL. He was a good-to-great quarterback almost immediately, though, so you wouldn’t cite Kelly as an example of how some quarterbacks take time to develop.
  • I’ve already written at length about Elway’s numbers.
  • McNabb’s numbers show a quarterback who steadily improved, but was still roughly league-average through his first five seasons. That’s a bit misleading, in my view, as the Eagles fielded some of the worst receivers in the league during that time, and McNabb added significant value with his legs. During his first 48 starts, the Eagles went 31-17; that lofty record was certainly a function of the team’s defense, but McNabb made three Pro Bowls during that stretch, too. Like Elway, I’d say early-career McNabb was better than his numbers.
  • Cunningham was another Eagles quarterback who was a bit undervalued by looking at just his passing stats. Cunningham was an even bigger threat on the ground than McNabb, and was stuck playing in Buddy Ryan’s offense. Cunningham finally produced strong passing seasons in 1990 and 1993 before his scorched-earth run off the bench with Minnesota in 1998. Cunningham was bad — really bad — during his first two years, but he quickly settled down into at least a decent passer and a dynamite runner by the time he was 24 years old.
  • Moon is in some ways a de facto number one pick, by virtue of landing with an absolutely horrible Oilers team. His numbers were not impressive during his first three years, but beginning in year four, he began building the groundwork for his Hall of Fame career. Moon’s career is unique, of course, but he does seem to have a bit in common with the Testaverdes, Youngs, Aikmans, and Bradshaws of the world when it comes to being brought in as a franchise savior for a bad team and struggling with such a Herculean task. P.S. Andrew Luck is pretty freakin’ good, isn’t he?
  • The last quarterback on the list is another unique player. Gannon sat his first three years, and then generally performed as a below-average quarterback during his next three seasons as the Viking starter. He then went to Washington and struggled in 1993, and missed the 1994 season with a shoulder injury. He sat on the bench in Kansas City in ’95 and ’96, but played well in five starts in 1997. He kept up his solid play in ’98 over ten starts, before going to Oakland and making the first of four straight Pro Bowls at the age of 34 in 1999. I’m not sure if we’ll ever see another career quite like that one.


I looked at the best 42 quarterbacks to enter the league since 1970. Then I divided each quarterback’s career into sets of 16 starts. Just four of those quarterbacks produced below-average passing numbers in each of their first two sets of 16 starts: three former first overall picks (Bradshaw, Aikman, Testaverde), and Brees. If a quarterback is below-average through two years worth of starts — say, Ryan Tannehill — then it seems highly unlikely that such a player will turn into a franchise quarterback absent extenuating circumstances. In the case of Bradshaw/Aikman/Testaverde, the extenuating circumstances were landing with terrible teams; for Brees, well, he also landed with the worst team in the league: the Chargers went 1-15 the year before he arrived, and Brees was the first pick in the second round.

For Tannehill, the extenuating circumstances could be an offensive line that blocked and acted like high school girls. There are other “excuses” one could make for him, but in general, I don’t think his situation was analogous to what some of those other quarterbacks inherited. If Tannehill turns into a star quarterback, he’ll be a very unique case.

On the other hand, if we set our baseline north of league average, a number of star quarterbacks failed to average at least +1.0 RANY/A through two years worth of starts. Some of those quarterbacks were solid but not spectacular (Brady, Favre, Krieg, Kelly), others had unexpected bursts of late career success (Theismann, Gannon), while some were stuck in rough situations (McNabb, Cunningham, Elway, Moon). If a quarterback has been mediocre after two years worth of starts, the quarterbacks you want to point to are Aikman/Testaverde/Bradshaw/Brees/McNabb/Cunningham/Elway/Moon.2 But the odds are much more likely, I suspect, that they just turn into the next Mark Sanchez.

  1. To say the very least. []
  2. Expecting anyone to be the next Theismann or Gannon is to expect the improbable. []
  • Nwebster

    Kelly missed two NFL seasons, while playing two seasons in the USFL – no Canada for him.

    Seems likely there’s some survivor bias here, you just don’t stick around long if you don’t perform relatively early, Gannon a notable exception. Between Thurman Thomas, Andre Reed and Jim Kelly, do the Bills have the weakest set of skill position players in the Hall from any one team era?

    • Chase Stuart

      Fixed, thanks.

      Yes, there’s quite a bit of survivor bias here, since the sample is the 42 best quarterbacks to enter the NFL since 1970. I do think Reed/Kelly are bottom-tier HOFers; Thurman is a more interesting player. He led the league in YFS in four different seasons, but those were the only years where he averaged at least 4.0 YPC and gained at least 1,100 YFS (or 900 rushing yards).

      • Erick Delgado

        Hard to address survivor bias in this instance without determining how to define a QB with franchise expectations.. is it any QB drafted in the first three rounds? Any quarterback that was given an above average contract? With either, there will be still be selection bias. One solution would be to include any QB that started at least 16 games.

        Its easy to see that using only RANY/A to evaluate QB production does not capture the value added from mobile QBs like McNabb and Vick. RANY/A captures yards loss due to sacks, but does not account for rushing yards.

        I would suggest a logistic regression that includes additional independent variables that measure QB performance (adjusted rushing yardage, completion percentage). Other variables that should be considered are a team’s winning percentage, a variable quantifying the WR corp’s and offensive line’s relative value.

  • JD


    Thomas was legitimately great, but Kelly and Reed are both among the weakest in the hall at their respective positions. The thing is, it’s very rare for a team to put three skill position players who played together in the Hall. Here are the candidates, as I see them:

    50s-60s Colts: Unitas, Berry, Moore, Mackey
    90s Cowboys: Aikman, Irvin, Smith
    60s Packers: Starr, Hornung, Taylor
    80s Redskins: Riggins, Theismann, Monk
    70s Steelers: Bradshaw, Harris, Swann, Stallworth
    70s Dolphins: Griese, Csonka, Warfield
    80s Chargers: Fouts, Winslow, Joiner
    80s 49ers: Montana, Craig, Rice, (Young)

    The following is my opinion. The only possible contenders are the 80s Redskins and the 70s Steelers. For the Redskins, Riggins is one of the worst RBs in the HOF, Theismann was good but nothing special, and Monk was an ageless compiler who was never that great (he only made 3 Pro Bowls in 16 years!). For the Steelers, Bradshaw was an inconsistent QB who was often the weak link on some great teams, everyone knows the arguments against Swann, Stallworth didn’t peak until the Steeler dynasty was over, and Harris followed four great years with nine mediocre ones.

    I’d probably say the 80s Redskins HOF skill players are worse than the 90s Bills. Kelly/Theismann and Reed/Monk are more or less a wash, but there was a time when Thurman Thomas was one of the most feared RBs in the NFL, something that I don’t think anyone ever said about Riggo.

    • JeremyDe

      Theismann is not in the Hall, nor will he ever be. If he ever had a chance, he had too many years wasted in Canada or on the bench behind Billy Kilmer.

    • I would go ahead and add a few more teams to the list (*I will include non HOF guys who are borderline, as you did):

      Colts: Manning, James, Harrison, Wayne
      Browns: Graham, Motley, Lavelli, Speedie*
      Broncos: Elway, Davis*, Smith*, Sharpe
      Raiders: Stabler*, Biletnikoff, Branch*, Casper
      Raiders again: Plunkett+ (worst player I will name), Allen, Branch*, Christensen
      Cowboys: Staubach, Dorsett, Pearson*
      Rams: Van Brocklin/Waterfield, Fears, Hirsch
      Rams again: Warner, Faulk, Bruce, Holt
      Eagles: Thompson* (no worse than Theismann), Van Buren, Pihos
      Eagles again: Van Brocklin, McDonald, Retzlaff*
      Jets: Namath, Maynard, Riggins – not exactly in their primes, but whatever
      Bears (yes, on offense): Nagurski, Grange, Hewitt (I can’t speak to their actual merit, but they are all in Canton)
      Packers: Blood, Herber, Hinkle, Hutson
      Giants: Strong, Flaherty, Badgro (Danowski just missed the list, but he was probably better than Theismann too)
      Giants again: Tittle, Gifford, Shofner*
      49ers: Albert*, Perry, Wilson*, McElhenny (not a strong group, but pretty darn good)
      49ers again: Tittle, Perry, JH Johnson, McElhenney, Wilson*
      Washington: Jurgensen, Taylor, Mitchell, Smith*
      Texans/Chiefs: Dawson, Haynes*, Arbanas*

      So…it might not last for a long time, but many teams have had the privilege of fielding great skill units. I don’t know which unit is worst, but there are plenty to choose from.

      • Chase Stuart

        Good stuff.

      • If we’re willing to have them just play together for one season, there’s the ’08 Cardinals: Warner, James, Fitzgerald. James was their leading rusher, with just 514 yds. That offense was only 15th in DVOA that year. Playoffs were a different story, though.

        • I would feel weird leaving Boldin out. The guy is and was a beast. I know stay guys aren’t supposed to care too much for narrative, but the guy played with a broken face! I can barely get on my car with a hip flexor tweak.

      • Dan

        Prime Warner and Faulk in a discussion about worst skill-position (future) players in the Hall? Am I missing something?

  • Kibbles

    Good article, Chase. I’ve looked at this from a fantasy angle in the past and typically found that if a quarterback didn’t have a top 12 finish in his first two years as a starter, it’s unlikely he ever would. Even guys like John Kitna and Tom Brady managed top-12 finishes. The big exceptions in the last 20 years were Ben Roethlisberger, Drew Brees, Matt Ryan, and Steve Young. Ryan/Roethlisberger were both over a full standard deviation above the league average in ANY/A as rookies, while Steve Young went from by far the worst team in the league to by far the best team in the league. Drew Brees is pretty much the only guy you can point to as someone who started his career slow and still became fantasy relevant, and Brees really started turning it on in his 3rd year as a starter.

    One thing I will say, though, is that I found that analysis based on games started is a little bit unfair to quarterbacks who start week 1 of their rookie year. I think it’s pretty clear that a guy who sat for three years will be more prepared in his first 16 starts than a guy who played immediately (see: Rodgers, Rivers, Palmer, Romo, and a whole host of other guys who had amazing “first” seasons.) That would be the one argument Tannehill fans could make, and I think they’d have a legitimate point- if Aaron Rodgers had started his first two seasons, there’s a very good chance he would have looked every bit as bad as Tannehill (remember the reports coming out about how awful Rodgers looked in camps and preseason early in his career).

    Because of that bias towards players who sat, I also looked at it in terms of years removed from entering the league. There I found that fantasy-usable quarterbacks typically were fantasy-usable within three years of entering the league, whether they sat or not. By that measure, we should be willing to give Tannehill one more year, and we should be very skeptical of Ryan Mallett.

  • The Ancient Mariner

    Could I see Dalton turning into this generation’s Krieg? I have a hard time seeing that, because Dalton’s key weakness was Krieg’s key strength. Dalton’s really a much more talented passer than Mudbone was, but as Bill Barnwell put it, “He can’t reliably handle pass pressure. . . . Dalton just crumbles when attacked. ” He’s more like a less athletic version of Krieg’s predecessor:

    “Jim Zorn learned to live life as an NFL QB on the run, literally. By 1983, he developed some very bad habits and seemed to panic at even the slightest hint of pressure. . . .

    “Contrast that with Krieg, who despite being sacked on average of about five times per game and hit every which way, stood tall and commanded a chaotic pocket. . . .

    “He was surprisingly weak-armed and had zero natural skill to put touch on anything; receivers had to be really aware on shorter passes because EVERYTHING floated on him in that range even in his prime. However, it was the simple fact that Dave let a play have a chance before screaming and running around like it’s 3rd and forever and it was the playoffs that allowed the offense to find its legs. It was that simple.”

  • Jonathan

    I’m an Eagles fan, and I can hardly contain my excitement when I think about the great year Nick Foles had last year, and the potential to have a strong QB behind center for the next decade. My question is basically the inverse of the study you did here. How often has a quarterback had an incredible, breakout first season or two, and then NOT become the franchise leader/perennial superstar that I hope Foles will be?

    Foles was +3 RANY/A last year and has only played 20 games, so my instinct is that he’s bound to regress…a little bit. Are there any quarterbacks that had numbers like this in their first 15-25 games and then didn’t become great (due to injury, or otherwise)?

    (P.S. Please tell me no, and that Foles is bound to be a Hall of Famer and lead the Eagles to 3+ Super Bowls. I need something to look forward to.)

  • Jordan

    Always watched the McNabb/Reid Eagles from afar. Those Eagle WRs were pretty putrid and seeing that Reid was a damn wizard with getting the most out of his RBs, Can we call his Eagle career underrated?

  • James

    “[Theisman] barely played with Washington in ’74 and ’75 (although he did return punts)”

    Let me do my best Al Michaels impression: He did WHAT?!?

    • JeremyDe

      Theismann was drafted by the Dolphins in 71. Negotiations broke down, and he went to Canada. In 1974, Washington obtained his rights because their top 2 QBs were a 35 year old Billy Kilmer and a 40 year old Sonny Jurgensen. In his first season, Theismann was grateful to be back in the NFL, but he knew he wouldn’t get any playing time at QB, so he volunteered to return punts.

      6 different players returned punts that year for the team. Theismann led the team with 15 returns for a 10.5 avg. He also returned 4 of the 5 punts in their playoff loss to the Rams.

  • James

    Great article Chase.
    I’m currently arguing with a friend about how Manziel will turn out as a pro. We’re going to try and compare his stats for his first 16 starts (whenever that happens) to the statistics of good to great qbs in their first 16 starts. What stat categories do you think we should use and do you have the average stats for the first 16 starts across the board for the qbs you compared above?