Pro Football Reference Blog

What are the chances that Tony Romo . . . or Matt Schaub . . . or Brodie Croyle . . . or Jim Sorgi . . . or Chris Simms . . . or Troy Smith . . . or [insert name of quarterback not drafted in the first 50 picks of the draft] becomes the next later round QB success in the NFL?

That's what I am going to try to evaluate. There are lots of problems in trying to come up with an answer, though. First, what constitutes "success?" To define Brady, Bulger, or Trent Green as a success is easy. But is it a success if a later round pick even starts in the NFL for a season? After all, the vast majority never see any substantial playing time in the regular season. What about guys like Scott Mitchell and Elvis Grbac? They did accomplish more than most later round picks. Successes or not?

To attempt an answer, I divided the later round quarterbacks into performance tiers, so that there were varying degrees of "success". I sorted all quarterbacks who entered the league between 1983-2002, and were not selected within the top 50 selections in the draft. For now, I am going to set aside those drafted since 2003, as only a handful have played a sizeable number of games in one season, and only Charlie Frye has seen significant playing time in more than one season. I then found all quarterbacks from that group who threw at least 150 pass attempts in at least one NFL season. 88 quarterbacks qualified. These 88 QB's were divided into two groups, those who threw 150+ pass attempts in more than one season (64 total), and those who threw 150+ pass attempts in only one season (24).

Next, to evaluate the quarterbacks within these groups, I used Adjusted Yards Per Attempt (AYPA), and compared the player's AYPA to a baseline number of "the Nth ranked quarterback in terms of AYPA among all who threw 150+ passes", where N = the number of teams in the league for that season. The total number of quarterbacks throwing at least 150 passes in a season fluctuates, but the average is approximately 1.25 x N (40 for a 32-team league, for example). So, this performance baseline would put about 20% of passers below the baseline (and thus having negative value) for the average season.

I took the player's AYPA, subtracted the baseline AYPA, and then multiplied that difference (Let's call it AYPA over replacement level) by the raw number of pass attempts to calculate "value added" for each season.

I was not as concerned with ranking the quarterbacks in exact order, as much as in creating tiers of relative success levels to then evaluate to see what those players had in common. I used the career passing "value added" and "value added" per season as the main tool, along with considering some other factors: the best years, the consistency of performance from year to year, and rushing ability and health, to make some decisions at the margins between tiers.

The 64 quarterbacks who threw 150+ attempts in more than one season were divided into five groups:

THE TOP GUNS make up the first tier, and consist of the following nine quarterbacks, sorted in order of career passing value added (for all 150+ attempt seasons) to date: Warren Moon, Trent Green, Mark Brunell, Kurt Warner, Rich Gannon, Tom Brady, Marc Bulger, Jeff Garcia, and Matt Hasselbeck. All of the Top Guns have posted at least two Top 5 Value Added Seasons, except Brady, who has just missed a few times and has five Top 15's. Only one QB, Mark Rypien, had multiple Top 5's and is not in this group, because his third best season was not even in the Top 15.

The next two tiers are what you might call Groups 2A and 2B. The next 21 QB's all started (from now on, when I say started, I mean threw at least 150 attempts) for multiple seasons. I split these two groups up, not strictly by raw value added, but by their career paths. The first, the STEADY PERFORMERS, all were Top 15 in at least 40% of their starting seasons, Top 25 in at least 75%, and started at least 4 seasons. The second, the UNEVEN PERFORMERS, had more up and down years, a greater deviation in year to year performance, and a higher incidence of bad years. On balance, I would take most of the STEADY PERFORMER group ahead of the UNEVEN PERFORMERS, though a few of the Uneven guys do rank high in "value added" due to their best season.

THE STEADY PERFORMERS are Chris Chandler, Brad Johnson, Neil O'Donnell, Steve Beuerlein, Aaron Brooks, Jeff Hostetler, Bobby Hebert, Jake Delhomme, and Charlie Batch.

THE UNEVEN PERFORMERS are Mark Rypien, Jay Schroeder, Brian Griese, Gus Frerotte, Jeff Blake, Elvis Grbac, Stan Humphries, Scott Mitchell, Doug Flutie, Erik Kramer, Jon Kitna, and Kordell Stewart. Kordell Stewart was moved up to this group, even though his passing profile fits more in line with the next group, based upon his extraordinary rushing contributions and seven starting seasons.

THE JOURNEYMEN are next. These guys got a few chances to start, maybe even put together a solid season among some not so good ones, and some of them hung around to play for several franchises. They are Rodney Peete, Bubby Brister, Jay Fiedler, Don Majkowski, Mike Tomczak, Damon Huard, Craig Erickson, Steve Bono, John Friesz, Rob Johnson, David Garrard, Steve Pelleur, Ty Detmer, Shaun Salisbury, Jim Miller, Josh McCown, and Ray Lucas.

Finally, among those getting 2 or more opportunities to throw 150 or more attempts, are the STOPGAPS. These guys were failed starters or career platoons/backups, maybe even guys who looked good in limited action one year, and were overexposed the next. They are Kent Graham, Hugh Millen, Quincy Carter, Billy Joe Tolliver, Eric Zeier, Chad Hutchinson, David Archer, Jamie Martin, Kelly Holcomb, Anthony Wright, Shane Matthews, Danny Kanell, Bobby Hoying, A.J. Feeley, Randy Wright, Doug Pederson, and Craig Whelihan. Quick aside--can someone explain numerous football commentator's fascination with mentioning Kelly Holcomb every time they think a team needs a backup that could challenge for a starting spot? I just do not get it.

The final two groups are simply the guys who threw 150 passes or more in only one season. I just took these 24 guys, and divided them into half based on their value added (or in most cases subtracted), into the SERVICEABLE BACKUPS, and the DISASTERS. I use the term "backup" here even though some were failed experiments as starters.

THE SERVICEABLE BACKUPS are Billy Volek, Jason Garrett, Tim Rattay, TJ Rubley, Cody Carlson, Alex Van Pelt, Bucky Richardson, Anthony Dilweg, Chris Redman, Steve Stentstrom, Frank Reich, Tommy Hodson. Bonus points if you can tell me where Dilweg went to school without looking.

THE DISASTERS were Doug Johnson, Billy Joe Hobert, Koy Detmer, Scott Campbell, Bruce Mathison, Donald Hollas, Mike McMahon, Stoney Case, Joe Dufek, Stan Gelbaugh, Rusty Hilger, and Chris Weinke.

When I look at the players in these tiers, and their characteristics, a few things stand out.

DRAFT POSITION DOES NOT MATTER IN PREDICTING SUCCESS

It only matters in determining the likelihood a quarterback even gets an opportunity, and how soon that opportunity occurs. When trying to evaluate the chances of an undrafted free agent, there is a problem. What is the relevant population? I settled on a rough estimate of 6 undrafted free agents per season, relying on the number of quarterbacks drafted in later rounds back when the draft went twelve rounds, and my guesstimate as to how many new undrafted free agents are make a roster each year, or at least make a practice squad. The other problem when trying to estimate success percentages of quarterbacks by draft position is that not all "quarterbacks" were drafted with a realistic thought of playing QB in the NFL. It's easy to spot a guy like Ronald Curry, but not as easy when we are talking about the quarterbacks drafted as athletes who never made a position switch successfully. My numbers below may not match up entirely with the draft results, because I removed some other guys who, in my opinion, were not drafted to play quarterback in the pros even though they played it in college and were listed as such on draft summaries. So, take my numbers below as educated estimates.

Here is the breakdown by draft pick selection number, sorted by tier groups:

GROUP        51-80  81-110  111-140  141-170  171-200  201-260  261-UFA
================================================================
TOP GUNS        0      1        1        1        2         1        3
STEADY          4      1        1        0        0         1        2
UNEVEN          1      3        0        3        1         1        3
JOURNEYMEN      1      3        3        2        1         2        5
STOP GAPS       3      2        1        1        1         2        7
SERVICEABLE     5      0        1        0        0         4        2
DISASTERS       2      2        0        3        1         1        3
ALL 150+ ATT   16     12        7       10        6        12       25
ALL DRAFTED    22     29       19       26       27        38      120
================================================================

GO WITH GUYS FROM QUALITY BCS CONFERENCE SCHOOLS

We know about the small college guys that have made a big splash in the NFL. Everyone wants to find the next Kurt Warner. The problem is, teams show no ability to find "Kurt Warner" in the actual draft. Out of 49 selected, here is the list of all quarterbacks drafted between 1983-2002 who attended a non-division IA program, and threw at least 150 passes in one season:

Rich Gannon (Delaware)
John Friesz (Idaho)
Josh McCown (Sam Houston State)
Craig Whelihan (Pacific)

That's only 8% of all non-DIA QB's drafted. Compare that to approximately 42% for "BCS" school products (49/118) and approximately 33% for other D-IA products (12/36). Warner, Hebert, Kitna and Romo went undrafted, while Giovanni Carmazzi, Jeff Carlson, Chris Hakel, Perry Klein, Doug Nussmeier, and Jeff Lewis all went by the end of round four.

Of the Top Guns, three were free agents, one each from a BCS (Moon-Washington), other D-IA (Garcia-San Jose St), and non-DIA school (Warner-Northern Iowa). Of the remaining drafted players, all except Gannon went to schools that are BCS-caliber, and their teams were average to above average during their college careers. You cannot say that scouts had limited opportunity to see these guys. In fact, they had multiple players selected from their school in their same draft class, and most played with multiple eventual first day draft picks during their college careers. Hasselbeck (Boston College), Green (Indiana), Brady (Michigan), Bulger (West Virginia), and Brunell (Washington) all played in pro style, balanced to run-oriented offenses, but with the exception of Brunell, were not frequent runners themselves. If you expand it out to look at the Steady Performers and Uneven Performers as well, you bring in twelve more drafted BCS-level prospects, five other drafted D-IA products, and no non-DIA guys. That's approximately 14% of BCS caliber and other D-IA caliber drafted quarterbacks reaching at least "Uneven Performer" level, compared to only Rich Gannon for the others (2%).

It appears that teams could increase their chances to get a surprise top starter later in the draft by focusing on players from BCS conferences who play in pro-style offenses, and may not put up huge raw numbers like other system quarterbacks (run-n-shoot, spread offense). If teams are trying to guess at who will be the next Warner or Ken Anderson, they may miss the next Brady, Brunell, Bulger, Green, or Montana, and the odds of finding those guys are a little better.

HEIGHT MATTERS, BUT BULK IS OVER-RATED

I should say that height matters to an extent. It is not going to make a player a really good quarterback. It is, however, going to allow a player that otherwise shows himself competent at the position to have more upside and longevity. Here are the numbers sorted by height (in inches). "72" includes all players 72 inches or shorter, and "77" includes all players 77 inches or taller.

GROUP          72     73      74      75      76       77  
=========================================================
TOP GUNS        0      2       1       4       2        0
STEADY          0      0       2       3       3        1        
UNEVEN          2      2       2       2       2        2
JOURNEYMEN      2      3       3       4       4        1
STOP GAPS       0      3       5       5       0        4 
SERVICEABLE     0      3       3       5       1        0
DISASTERS       2      0       2       5       3        0
ALL 150+ ATT    6     13      18      28      15        8
=========================================================

The key height seems to be 6'3" or taller. 16 of the 21 (76.2%)quarterbacks who achieved at least one top five passing value added season were 6'3" or taller. Compare that to 51 of 88 (58.0%) for all late round passers who ever attempted 150 passes in a season. I did not go through every drafted QB's height who never played in the NFL, but I would guess that the number 6'3" or taller is somewhere in the 55% range, for all late round quarterbacks ever making a roster.

The guys under 6'3" who were successful primarily relied on their mobility and ability to throw on the run to make up for the inch or two in height. Brunell and Garcia are the most successful examples of this. Warner and Delhomme have been the most successful pocket passer types listed at 6'2".

Weight, however, does not matter, and might be overvalued. I am not sure how to quantify this, and it is just an observation, but alot of the most successful passers had a similar wiry body type coming out of college. Brady, Trent Green, Bulger, Chandler, Beuerlein, Warren Moon--none of these guys are built like Peyton Manning, Culpepper, McNabb or Roethlisberger.

So, let's turn to the current guys, but first take a quick look at how long it takes the top performers to become good or great, to get a sense of what to expect and when out of the current group of later round quarterbacks. Here is a chart showing how many seasons, after the first season in which the passer had 150 attempts, before that passer first reached certain milestones in terms of value added. It consists of all passers who had at least 2,500 in career passing value added--21 total passers. All members of the Top Guns are included, as well as everyone from the Steady Performers except Charlie Batch, plus Rypien, Griese, Schroeder, and Frerotte from the Uneven Performers group. If the milestone was achieved in year "0", that means that milestone was first accomplished the same season the player first began starting.

YR/FINISH  TOP5    TOP10  TOP15  TOP20    TOP25    TOP30
=========================================================
0          1       4       9      16        19       21
1          4       8       8       4         1        0
2          3       3       1       0         0        0
3          1       2       0       0         0        0
4          1       2       0       0         0        0
5+         2       4       3       1         1        0
Never      5       0       0       0         0        0
=========================================================

Only Chandler (28th) and Gannon (26th) were outside of the Top 25 in value added their first starting season. Chandler, notably, was the only one to start straight out of college as a rookie among this group. Slightly less than half finished in the Top 15, and Kurt Warner was the one Top 5 season in the first season starting. Within one season of their first opportunity, all except one had already posted at least one Top 20 Value Added Passing season, and that one was Chandler, who did not get another extended opportunity until his 5th year in the league. Slightly over half (12/21) had already had a Top 10 season by their second year as a starter. The difference between the Top Guns, the Steady Performers, and the Uneven Performers is not so much how they started their careers, but how they improved or sustained as their careers went on.

In conclusion, I'm going to turn back to some of the notable guys drafted since 2003, to provide an assessment of their chances.

Matt Schaub: I'll give him about a 40% chance of of joining the group with at least one Top 5 type passing season in his career. I think he has a chance to join the "Steady Starter" group, and we should know fairly early if that is the case, given his age and years of experience in the NFL.

Tony Romo: Big first year as a starter means he probably is going to be somewhere within the Top 30 list (top gun, steady performer, uneven performer) when his career is over. His height (6'2") means his upside and consistency may be impacted in the long haul, but he is a more mobile guy who could emulate the career of a Mark Brunell.

Tarvaris Jackson: If he is a success, he would break the mold. It would take one above average passing season in his career to make him the second most successful passer drafted from a non-Division IA school since 1983. We know the Vikings staff likes him, and he was the 62nd overall pick. There were 16 other teams that liked their guy picked between 51-80, and gave their guy an opportunity, and only 6 reached Journeyman status or better. He is however going to at least join the Top 5 most successful draft non-DIA guys, because he is all but guaranteed to start a significant number of games this year, which should mean we have an idea of the answer by season's end.

Brodie Croyle: See some of the comments on Jackson, as he is a 3rd round guy who Herm Edwards likes and thinks can be the starter. He looked bad in preseason, looked bad on 7 pass attempts last year. The Chiefs still need to give him substantial starting time this year if they have a belief he is the quarterback of the future, so they know the answer. He does have the tall skinny kid thing going, but then again, TJ Rubley and Stan Gelbaugh were skinny too.

Andrew Walter, Kyle Orton, and Bruce Gradkowski: Yeah, they get lumped together. If they never start again, they would rank as 3 of the 7 worst Disasters on the above list. Here's guessing that two of them actually do never start again, while one may get another chance to add negative value for someone else. I could make excuses for each of them (Orton and Gradkowski were rookies, Walter had the Raiders offensive line and a bed and breakfast operator calling plays), but I am sure everyone on the list above had some excuse.

Charlie Frye: I think we have seen enough to say it is highly unlikely he rises above "stopgap". The eventual successes did not put together 2 seasons like Frye at the start of their careers.

Chris Simms: Intriguing. We don't know if he will ever be able to make a comeback from the spleen injury. He was 21 in value added in his only season with 150+ attempts, his 3rd year in the league. He has the height and big school background shared by many successes, and put up numbers similar to many of those guys in his first opportunity. If I were a team, I would take a low-risk chance on Simms. He could have a career like Chandler or Beuerlein if he can come back from that injury.

Troy Smith: Height is an issue. His absolute upside is probably along the lines of Jeff Blake or Erik Kramer, but I am not saying that is likely. He could be a serviceable backup, or we may never see him play a significant role in an NFL game.

Jim Sorgi: Another intriguing guy. He is second string, has looked solid in limited action, and fits the profile of the successful guys--- BCS school, run-oriented pro style offense in college, good height, skinny. At this point, we can surmise he is not as good as Peyton Manning.