โ‰ก Menu

Brad Oremland is a longtime commenter and a fellow football historian. Brad is also a senior NFL writer at Sports Central. There are few who have given as much thought to the history of quarterbacks and quarterback ranking systems as Brad has over the years. Today, he’s contributed this guest post, but also is asking for your feedback. So please, help Brad and help us, in the comments, with your thoughts.

In recent weeks, Football Perspective has hosted some lively discussions about the greatest quarterbacks of all time. I like to think my approach to these issues is balanced, but it begins with statistics. I am always looking for ways to improve my analysis, and Chase has graciously invited me to post the results of my statistical rating system for quarterbacks.

This is not my personal list of the best quarterbacks in history. My subjective list differs, at time significantly, and I’ll post that next month. The list below is purely statistical, with three notable limitations:

1. It measures regular-season statistics only.

2. It covers the years 1946-2014. The modern quarterback position didn’t really exist prior to the mid 1940s.

2b. QBs who played prior to 1946 are omitted, even if they continued to play after the end of World War II. I don’t want a ranking that shows Sammy Baugh 65th, since it’s missing the first decade of his career. Players like Baugh, Sid Luckman, and Bob Waterfield are deliberately excluded.

3. Only seasons in the NFL, AFL, and AAFC count toward these rankings.

My purpose in posting this list is to ask for help. There are a lot of smart readers and commenters at this site, and I want you to critique my results.

I’m not showing my work yet: I’m not looking for a critique of my process, but of my results. Who’s too high? Who’s too low? You can identify individual players, or patterns. Wherever you think I messed up, I want to hear about it. Please keep in mind, though, that this is purely a stat-based list. It doesn’t represent my opinion, and it’s not slanted toward or against individual players or teams.

But just because this system is unbiased, that doesn’t make it perfect. It is definitely not perfect. But I’m comfortable making subjective adjustments, and that may create blind spots that prevent me from improving the formula. I’m asking you to evaluate the list below and judge where you think it is counterintuitive or inaccurate.

Are players from the ’70s overrated? Are contemporary players underrated? What about players from good teams, and players from bad teams? Are running QBs overrated? Underrated? How about game managers vs. downfield bombers? Is the system fair to them? Are one-year wonders overrated? Are compilers overrated? Players who threw a lot of TDs, a lot of interceptions, players who got sacked a lot? Wherever you think the system is off, I’m eager for your feedback.

Hopefully you find this list interesting, and you can expect a fuller explanation of my rankings in the future, but in the meantime, I appreciate your input and assistance. I included each player’s numerical score, which I realize isn’t in context yet, but it can give you a more precise idea than a simple ranking. Troy Aikman, Donovan McNabb, and Joe Namath, for instance, are effectively tied. Below are the top 125 QBs of the modern era, as ranked by my stat-based system:

1.    Peyton Manning       51.83
2.    Dan Marino           45.73
3.    Fran Tarkenton       38.66
4.    Johnny Unitas        37.46
5.    Joe Montana          34.76
6.    Brett Favre          34.69
7.    Otto Graham          33.94
8.    Drew Brees           31.87
9.    Tom Brady            31.80
10.   Dan Fouts            30.52
11.   Steve Young          29.62
12.   John Elway           28.02
13.   Sonny Jurgensen      27.51
14.   Ken Anderson         27.45
15.   Norm Van Brocklin    26.47
16.   Y.A. Tittle          26.38
17.   Warren Moon          24.20
18.   John Brodie          23.35
19.   Roger Staubach       22.44
20.   Aaron Rodgers        20.41
21.   Boomer Esiason       19.95
22.   Roman Gabriel        19.11
23.   John Hadl            19.04
24.   Jim Kelly            18.99
25.   Bobby Layne          18.62
26.   Jim Hart             18.57
27.   Jim Everett          17.70
28.   Philip Rivers        17.65
29.   Kurt Warner          16.73
30.   Tony Romo            16.31
31.   Troy Aikman          15.96
32.   Donovan McNabb       15.94
33.   Joe Namath           15.93
34.   Bart Starr           15.74
35.   Ben Roethlisberger   15.61
36.   Randall Cunningham   15.42
37.   Len Dawson           15.35
38.   Terry Bradshaw       15.18
39.   Vinny Testaverde     15.10
40.   Rich Gannon          15.00
41.   Norm Snead           14.73
42.   Steve McNair         14.62
43.   Jeff Garcia          14.55
44.   Bert Jones           14.31
45.   Trent Green          14.30
46.   Charlie Conerly      14.23
47.   Daryle Lamonica      14.08
48.   Mark Brunell         13.73
49.   Joe Theismann        13.68
50.   Ken Stabler          13.60
51.   Drew Bledsoe         13.50
52.   Milt Plum            13.35
53.   Charley Johnson      13.34
54.   Phil Simms           13.07
55.   Billy Wade           12.90
56.   Bob Griese           12.41
57.   Dave Krieg           12.37
58.   Daunte Culpepper     12.359
59.   Craig Morton         12.357
60.   Steve DeBerg         12.23
61.   Brian Sipe           11.994
62.   Steve Grogan         11.987
63.   Bernie Kosar         11.98
64.   Billy Kilmer         11.97
65.   Neil Lomax           11.58
66.   Matt Ryan            11.08
67.   Joe Ferguson         11.02
68.   Ron Jaworski         10.983
69.   Earl Morrall         10.981
70.   Carson Palmer        10.97
71.   Eli Manning          10.80
72.   Tobin Rote           10.79
73.   Matt Hasselbeck      10.69
74.   Brad Johnson         10.50
75.   Jeff George          10.13
76.   Frank Ryan           10.04
77.   Don Meredith          9.84
78.   Tommy Kramer          9.81
79.   Mark Rypien           9.63
80.   Danny White           9.49
81.   Ken O'Brien           9.34
82.   Jim Zorn              9.28
83.   Greg Landry           9.201
84.   Steve Bartkowski      9.198
85.   Doug Williams         9.13
86.   Kerry Collins         8.95
87.   Jake Plummer          8.88
88.   George Blanda         8.83
89.   Bobby Hebert          8.78
90.   Matt Schaub           8.77
91.   Archie Manning        8.59
92.   Chris Chandler        8.57
93.   Johnny Lujack         8.47
94.   Jim Harbaugh          8.43
95.   Ed Brown              8.38
96.   Jim Plunkett          8.35
97.   Jim McMahon           8.16
98.   Neil O'Donnell        7.71
99.   Lynn Dickey           7.53
100.  Frankie Albert        7.51
101.  Chad Pennington       7.48
102.  Michael Vick          7.46
103.  Jeff Blake            7.40
104.  Marc Bulger           7.39
105.  Bill Kenney           7.37
106.  Jeff Hostetler        7.11
107.  Bill Nelsen           6.99
108.  Steve Beuerlein       6.97
109.  Jay Schroeder         6.87
110.  Jay Cutler            6.71
111.  Richard Todd          6.54
112.  Matthew Stafford      6.404
113.  Aaron Brooks          6.401
114.  Wade Wilson           6.31
115.  George Ratterman      6.30
116.  Elvis Grbac           6.29
117.  Chris Miller          6.25
118.  Scott Mitchell        6.19
119.  Cam Newton            6.02
120.  Stan Humphries        5.96
121.  Babe Parilli          5.90
122.  Doug Flutie           5.89
123.  Jake Delhomme         5.70
124.  Bobby Thomason        5.58
125.  Brian Griese          5.56

Russell Wilson and Andrew Luck are both over 5.0, and I expect them each to rank in or near the top 100 by this time next year. Both look to me like future Hall of Famers.

I know many of you are familiar with Chase’s stat-based ranking system. He and I use a lot of the same ideas. There are some players who are rated more accurately by Chase’s system than mine, and I believe some are rated better in my system than his. But I hope you’ll try to evaluate this list by what you know about the game, and not simply compared to a different methodology, even one as carefully refined as Chase’s.

I hope you find some value in the list, even though I haven’t yet explained its methodology, but in particular, I offer thanks in advance for any feedback you can offer.

  • Adam Steele

    Yay, my favorite topic! Thanks for posting this, Brad. For a purely statistical ranking system, I think this is a very accurate list. Going down the order, here are some things that stand out to me:

    Peyton and Marino are way ahead of everyone else…and I agree with that for this type of evaluation. IMO, they are clearly the two best statistical QB’s of all time.

    Tarkenton at #3 is a bit too high, although I think he is generally underrated and should definitely be top 10. Perhaps his longevity and running are inflating his rating.

    Favre at #6 is way too high. Pretty clearly a function of his extreme durability since he was only dominant for a few years. However, I think your system does a great job overall of avoiding too much credit for compilers. The only other name that screams “compiler” is Bledsoe at #51, who had a handful of good seasons and a lot of bad ones.

    Brees ahead of Brady is curious, although at least it’s a reasonable debate. Actually, Brady being as low as #9 is also curious.

    Young at #12 is too low, as he was absolutely dominant for three years and excellent for all eight of his starting years in SF.

    Seeing Roman Gabriel at #22 makes me cringe. He was a good QB, but his only real standout skill was INT avoidance. Interestingly, the other INT avoiders seem to be ranked properly, so not sure what the deal is with Gabriel.

    Jim Hart at #26 really doesn’t make any sense. I’d offer a comment but I’m not sure how he ranks this high.

    Trent Green at #45 is too low. He had five seasons as one of the best QB’s in the league, which is more than you can say about several guys above him.

    The best thing about your system is that there doesn’t seem to be any noticeable systematic flaws. There isn’t one type of QB who is consistently overrated or underrated.

    • Thanks for your thorough approach, Adam, it’s really helpful for me to see so many players whose ratings seem off to you.

      I’m surprised that you think Tarkenton is too high. He was one of the most efficient passers of his generation, one of the greatest running QBs in history, and he played forever with basically no down seasons. You mentioned longevity and running “inflating” his rating, but those seem to me essential components of his career. If you want to elaborate on your feelings, I am persuadable on this issue and I’d be interested in your take. I rate Tarkenton a little higher than most people do, and that’s largely a function of how I interpret his stats, so it’s important for me to hear where I might have missed something.

      I tend to agree with you on Favre too high, but I’m not sure he was only dominant for a few years. He was pretty great from 1994-98, second-team all-pro three other years, and pretty strong in ’04 and ’09. That’s not Steve-Young-dominant, but it’s not just longevity, either. I’d love for you to expand on Bledsoe. I know I have him a lot higher than Chase (whom I believe has him outside the top 100), but 51st doesn’t seem too far off to me. Certainly he’s overrated by most people, but I don’t think he had a lot of bad seasons, I think he had a lot of seasons that were close to average. Because my system rates Bledsoe so much differently than Chase’s, I’m really interested in feedback on his ranking.

      Roman Gabriel is another one where I’d be interested in anything you have to add. I actually rate him about the same (22) as Chase (26th in the Modern Era). I’m not sure I agree about avoiding INTs being his only distinguishing asset. Gabriel produced a lot of points: he led the NFL passing TDs twice, and he scored 30 rushing touchdowns. He had a long career, made four Pro Bowls, and NFL MVP in ’69. Subjectively, I have him lower than this, but “cringe” is a strong word. Again, I don’t necessarily disagree, but if you want to expand on this, I’m interested in your analysis.

      re: Jim Hart, I simultaneously [1] agree with you, 26th is way too high; and [2] believe a stat-based analysis might have to go this direction. When you subtract early players like Baugh, Chase rates Hart even higher than I do. Hart was distinguished by three factors: his long, productive career; his high number of pass attempts; and his excellence avoiding sacks. I do think Hart is too high, but I’m not sure how to adjust his rating without making radical changes to my entire system.

      It’s nice of you to absolve me of any noticeable systematic flaws, but I wonder if your critiques don’t point to a slight trend of overrated compilers, or at least high-level compilers. Tark, Favre, and Hart all had long careers, while Young had a short one (as starter, anyway). Do you think my method might show a small bias in that direction?

      And thanks so much again for your feedback; I know this is something you’ve studied and you take seriously.

      • Kibbles

        I’m with you on Tarkenton. It should come as no surprise whatsoever that someone who hates entanglement issues as much as I do loves Tarkenton. I didn’t quite put him 3rd in my Wisdom of the Crowds ballot, but he checked in at 6th for me, behind Unitas, Elway, and Young, but ahead of Montana and Brady. And in terms of career value added to his franchise, I would probably prefer 239 starts from Tarkenton over 185 from Unitas or 143 from Young.

        • Richie

          Yeah, Tarkenton makes an interesting example of separating QB from his team. He averaged 12 AV/season in his first stint with the Vikings, who were an expansion team.

          Then he averaged 15 AV/season during his prime years, while playing for the Giants.

          He went back to the Vikings (who were presumably a much different roster 6 years later) and averaged over 12 AV/season in his twilight.

      • sacramento gold miners

        Appreciate your efforts on this table, but winning is such a key part of QB performance, and I could never put Drew Bledsoe ahead of Bob Griese. Bledsoe tailed off so badly after leaving New England, that his overall career is affected. Ken Anderson is way overrated, very efficient, but just didn’t take full advantage of his postseason opportunities.

        Sports is full of ironies. I could be mistaken, but Neil O’Donnell is near the top for lowest interception percentage career-wise, but is remembered mostly for the SB 30 mistakes.

        • Nitpicker

          The importance of factors other than stats is probably why Brad specifically noted that the list is purely statistical and can significantly differ from his personal, subjective list.

      • Adam Steele

        You’ve convinced me on Tarkenton. My initial comments were a knee-jerk reaction to my surprise at seeing Tark ranked #3. I’ve never seen a stats based system rank him so high before, so I thought “Oh there must be something wrong here.” After looking over his numbers again, the only thing wrong was my preconceptions. Now, that doesn’t mean I think Tark was definitely better than Montana or Unitas, but it’s a reasonable argument.

        On Favre, I think you and I just have a different definition of dominance. IMO 1995 was his only truly dominant season. Of course he had several other great seasons (96, 97, 01, 09), but his three year peak from 95-97 is not in line with the 6th best QB of all time.

        As far as Bledsoe, honestly I wouldn’t even rank him in the top 100. His career RANY/A is negative! I will give bad seasons a pass for rookies, old guys, and talent starved rosters. Other than his rookie year in ’93, Bledsoe has no excuses. While he never played for juggernaut teams, his supporting casts were hardly terrible. Between age 22 and 34, Bledsoe posted SIX seasons below -100 VALUE. This includes -224 at age 28, -356 at age 32, -553 at age 31, and -716 at age 23. In 14 seasons, he had only four good years. He had several awful seasons in the middle of his prime despite being healthy and playing on relatively strong teams. His career was smoke and mirrors.

        Check out Roman Gabriel’s career advanced passing stats: 94 Y/A+, 99 TD%+, 119 INT%+, 112 sack%+. I stand by my assertion that avoiding INT’s was his only prominent skill. His sack rate was pretty good although that excludes half of his career. But an average TD% and well below par Y/A are major red flags to me. I think INT avoidance is highly overrated, especially in the bombs away era Gabriel played in. I’d bet the majority of INT’s during that era came on downfield throws and weren’t that much worse than a punt. Thus I give Gabriel’s INT% very little credence. Any QB with a career 94 Y/A+ does not belong in the top 30 of all time for any reason.

        I guess compilers are slightly overrated by your system, but it’s not a glaring flaw. Personally I value peak and dominance more than longevity, but that’s just an opinion and I understand why people might feel differently.

        • Thanks, Adam, this is the type of feedback I was hoping for.

          You’re certainly right that with a high threshhold for what constitutes “dominance”, Favre doesn’t fare well statistically. He had a lot of good seasons — even very good seasons — but only one that was truly exceptional. Are there other players with whom you feel I may have given them too much credit for very good seasons (like Favre in 96, 97, 01, 09) that fall short of greatness?

          re: Bledsoe … this may simply be a philosophical difference. Something I believe is that a player doesn’t have to be above average to help his team. My method compares players to something approximating replacement level. There’s an adjustment to make sure the list isn’t dominated by compilers, and I think it’s largely successful: players like Steve Young and Aaron Rodgers rate much higher than the Vinny Testaverdes, Drew Bledsoes, and Kerry Collinses.

          Maybe it makes sense to use a higher standard like “average” when evaluating the very best of all time, but when we’re talking about #50, or #75, or #100, I think using average as the standard is a mistake. One of the most common misconceptions in sports is that it’s easy to find average players. It’s really hard! Finding an average QB is not something any team can take for granted. In the last couple years, average-ish QBs like Joe Flacco and Jay Cutler have gotten 9-digit contracts. Players like that are not easy for a team to obtain.

          Drew Bledsoe was a productive starter for 12 seasons, about half of which produce some type of meaningful value in my system. If I were running a team, I would not want four years of Chad Pennington, and eight years of whomever I can find, over the full career of Drew Bledsoe. You’re welcome to disagree, but I don’t think average is the appropriate standard against which to measure players. I compare QBs to replacement level and adjust to reward dominant peaks. In this part of the rankings, Bledsoe’s not competing against Steve Young and Brett Favre; he’s up against Mark Brunell and Jeff George and Brad Johnson. In that context, Bledsoe’s sheer production — he was such a huge part of his team’s offense — looks valuable by my methodology.

          re: Gabriel … I don’t know that looking at career averages is the appropriate way to judge a player’s career. While we don’t want to place a lot of value on hang-around years, I don’t think we want to punish them either. In 1976, Joe Namath threw 4 TDs with 16 INTs. That season really hurt his averages, but it doesn’t diminish what he accomplished early in his career. In his last four seasons, Gabriel averaged 4.4 net yards per attempt. That doesn’t really concern me. Almost all of Gabriel’s value (in my system) is from 1967-73; his other nine seasons are icing, but he’d rate in the top 40 without them.

          From 1965-74, Gabriel ranked 4th in passing yards, 3rd in TDs, 2nd in TD/INT +/-, and 1st in QB rushing TDs. His efficiency numbers are fine, but his value came more from production than efficiency, and I just don’t know you accurately rate a quarterback solely on his passing stats: Gabriel rushed for over 1,000 yards, and he is one of only 16 QBs with 30+ rushing TDs. Chase and I both rank Gabriel in the top 30 (statistically), because we both place tremendous value on avoiding INTs, and Gabriel was the best of his generation. I would love to hear an argument that INTs should be valued differently in different eras, because I think that’s probably true, but realize that if I even slightly de-value the importance of INTs, Hadl is in the top 20, Norm Snead is in the top 40, and Jim Hart flips with Gabriel. Creating rankings like this is nasty business.

          I’m a little disappointed that I convinced you on Tarkenton! You raised some great points about other players, and to improve my methodology, I need smart people to make vigorous attacks on it, like you did with Bledsoe and Gabriel. If you want to shoot holes in my defenses of their rankings, please do! Every critique helps me figure out what I need to improve.

          • Adam Steele

            Our Bledsoe discussion highlights an important issue. The various tiers of historical performance require different criteria to evaluate the players in them. If you’re trying to measure Bledsoe in the context of all time greatness, then his average and sub-par seasons should be counted against him. But if you’re measuring him in terms of value to a franchise, his average seasons should boost him since, as you said, it’s hard to find even average QB’s. However, in the latter scenario, I don’t think his high volume of attempts should benefit him in the rating. Unlike rushing, passing opportunities do not necessarily equal an endorsement of ability. Are 600 mediocre pass attempts “better” than 400? I don’t think so. In Bledsoe’s case, I believe his #1 draft pick status afforded him more chances than he otherwise would have had. If the 125th pick produced Bledsoe’s stat line from ’93-’95, would he have gotten the chance to chuck the ball every year for another decade?

            That’s a fair point about Gabriel and tossing out the bad seasons at the beginning and end of his career. I shouldn’t have been so lazy with the numbers I presented. I do strongly believe that INT’s were less damaging in the downfield passing era, but without play-by-play it’s hard to prove or quantify. And in general, I’m at the extreme end of not caring about INT’s, so most people would probably not have the same issue about Gabriel that I do. Now it does give me pause when you say Hadl would be in the top 20 with a smaller INT penalty, but as you said it’s a tradeoff no matter what.

            Haha, I don’t necessarily agree that Tarkenton should be 3rd, but I am conceding that the logic you used in your argument was sound. I could see him ranking anywhere from 3rd to 8th and it being reasonable depending on the parameters. What I like most about Tark is that he checks all the boxes and isn’t relying on one or two amazing seasons to prop up the rest of his career.

            Others have mentioned this, but I have a problem with Staubach at #19. He was great for enough years that his lack of playing time shouldn’t hurt him so much. Staubach is the slightly lesser version of Steve Young in that regard.

            • Apologies in advance for a long reply. It’s surprising how much room there is for debate among people who largely agree with each other. Don’t scroll down until you’re in the mood to read a whole lot about Drew Bledsoe!

              I’m actually having a hard time separating my subjective, personal feelings on Bledsoe from my objective, stat-based ranking. Let’s start with full disclosure on the former: I think Bledsoe was overrated for most of his career, and wildly misremembered by those who use career yardage as their preferred metric, but the stat-based community has gotten so used to talking about how overrated he was that many of us have forgotten the things he did well. I don’t know how well you remember Bledsoe’s reputation in the mid-to-late-90s, but he was very highly regarded. He took over a team that was 2-14 and had them in the playoffs two years later, the Super Bowl in just his fourth season. He was the youngest QB to reach 10,000 passing yards.

              Bledsoe was a four-time Pro Bowler. Other than a couple years with Curtis Martin, he never really had a running game, and opposing defenses knew he was going to pass. He faced aggressive blitzes and nickel defenses on first down, even in two-receiver sets. He was the engine that drove the Patriots’ offense, and he produced points on teams without a lot of weapons in the receiving corps. Bledsoe would be in my top 75 of all time, probably the top 60.

              The thing is, a stat-based formula doesn’t know all that. It doesn’t know his team wasn’t very good, and it doesn’t care that the team improved when Bledsoe started playing. What it does know is that Bledsoe passed a lot, with average-ish efficiency. I guess that’s the definition of a compiler, and here I am advocating it — but keep in mind, I’m not pushing Bledsoe in the top 20, or 30, or 40. My formula ranks him 51, which seems reasonable for a four-time Pro Bowler who ranks among the all-time top 10 in passing yards.

              You asked two specific questions, and I’ve taken a while getting to them.

              If the 125th pick produced Bledsoe’s stat line from ’93-’95, would he have gotten the chance to chuck the ball every year for another decade?

              Yes, absolutely. Bledsoe’s ’94 season in particular showed great promise. The Patriots’ leading receiver that season was tight end Ben Coates, a nice player, but hardly a game-breaker. Their top WRs were Michael Timpson and Vincent Brisby. The leading rusher was Marion Butts, who gained 703 yards with a 2.9 average. The defense produced zero Pro Bowlers. Only two of the starting offensive linemen played on New England’s 1996 Super Bowl team. The others were backups on other teams, or out of the league entirely.

              Was Bledsoe’s stat line impressive? Other than the yardage, not really. But he was doing heroic things on a talent-starved team, and people throughout the league knew it. No one suggested, 19 years ago, that the Patriots had made a mistake by drafting Bledsoe, or that they needed to consider finding another option at quarterback. They had the same backup in 1996 (Scott Zolak) that they did in ’94, and they used their top draft pick on Terry Glenn, recognizing that they needed to give Bledsoe some help. You want a mid-90s Patriot draft pick who was a disappointment, point at Terry Glenn, not Bledsoe.

              If you take the stat line out of context, maybe you would worry about Bledsoe, though I think that age-22 season would buy anyone a year’s grace period. In context, I think Bledsoe was viewed, entering 1996, as the most promising young QB in the league. Was that influenced by his college career and draft position? Probably a little bit, but it was not the defining factor. He led the league in passing yards with Michael Timpson and Vincent Brisby as his top WRs!

              Are 600 mediocre pass attempts “better” than 400?

              I guess it depends on what we mean by mediocre. An average passer with 600 attempts should absolutely rate higher than an average passer with 400 attempts. He was a bigger part of the offense, and he probably faced more pass-oriented defenses.

              I’m not sure how this relates to anything, but I looked it up and don’t want that to go to waste… nine QBs of some consequence were born in the early 70s: Kurt Warner, Steve McNair, Jeff Garcia, Trent Green, Mark Brunell, Drew Bledsoe, Jake Plummer, Kerry Collins, and Jeff Blake. Everyone ahead of Bledsoe is top-50, and even in a purely statistical analysis, I wouldn’t want my system to rank Plummer or Collins or Blake ahead of Bledsoe. It’s hard for me to see how his ranking could be off by enough to push him out of the top 100. Granting that Pro Bowl selections are unreliable, how many four-time Pro Bowlers aren’t among the top 100 at their position? I also looked this up: http://pfref.com/tiny/iOqxy … the numbers are cherry-picked, but this suggests something other than mediocrity.

              It probably sounds like I like Drew Bledsoe a lot more than I do. Actually, I never really liked him, and it’s weird to post such a long defense of him. I guess I feel like Bledsoe did a lot of things that cause a player to be overrated, but he did many of them in a very tough playing environment, and the objective rank is fairly close to where I rank him subjectively (maybe 55-60-ish).

              Switching gears (finally!) I agree about Staubach, and my subjective ranking is quite a bit higher. I would have expected him to rank near Steve Young, a comparison you drew as well. Staubach is hurt, obviously, by the fact that he has no hang-around value. My system is designed so that low-impact seasons don’t have much effect on a player’s ranking. But most elite players have a few seasons like that near the beginning and/or end of their careers. Staubach doesn’t. I completely agree with you, but I think this is a case where Staubach is a special exception requiring subjective adjustment — I’m not sure what can be done within the formula that wouldn’t create more problems than it solves.

              • Adam Steele

                Sorry for the very long delay in response time. For someone who doesn’t like Bledsoe, you make a damn good case for him. Being that I was in elemtary school during Bledsoe’s early years, I didn’t realize how vapid his receivers were. I’d never heard of Timpson and only heard of Brisby from seeing his jersey on the screen during NFL films, and they did nothing without Bledsoe. When a TE (Coates) is the most famous receiver he threw to during those seasons, that’s probably a weak supporting cast. And as you mentioned, defenses were able to sell out to stop the pass, which certainly made his job harder. And of course playing in Foxboro winters adds to the degree of difficulty.

                However, I don’t agree that a higher volume of average efficiency necessarily indicates a better performance. In some cases it definitely does, and I’d include Bledsoe’s 1994 as one of those seasons. The Pats were competitve so most of his pass attempts came against teams playing real defense. The converse is a high volume of attempts on a bad team, which lends itself to plenty of garbage time stat padding. Carson Palmer’s 2012 immediately comes to mind. He had 565 attempts, and I would argue the final 165 were easier than the first 400 because they came against soft prevent defense. His 103 ANY/A+ is very deceiving for this reason. Bledsoe’s 2002 season with 600+ attempts involved a lot of garbage time despite being an 8-8 team, and again I think those extra passes aided his efficiency numbers. After thinking through this in more detail, my assertion that Bledsoe should be outside the top 100 was ridiculous. Now I would place him somewhere between 60 and 70.

                Just for reference, how would you rate a season like Matt Stafford’s 2012 (average efficiency on 700+ attempts)? Do you think he added significant value to the Lions? I could see it both ways.

                • Just to play Devil’s Advocate against myself, the stat-based formula doesn’t know how crappy Bledsoe’s receivers were. But I don’t think any knowledgeable fan really doubts that Bledsoe was the premier offensive player on the Patriots in the ’90s. It’s also worth mentioning that Coates never did anything without Bledsoe: he was an all-pro with Bledsoe, and a nobody without him.

                  For the 2012 season, Stafford ranked as the #12 QB in my formula, basically a touch above average. As a point of reference, that season scored almost exactly half as much career value as Matt Ryan the same year. Stafford is probably a touch too high, but I don’t think the error is significant. I do believe he added value for the Lions, but not as much as a Tony Romo or Russell Wilson.

                  I love that you brought up Carson Palmer’s 2012, a great example of garbage-time stat-padding. I didn’t think of Bledsoe’s 2002 that way, however. But you raise an important point: stats are the beginning of this type of analysis, not the end, because you have to consider context. Is there any real value in garbage time production? Should we credit a 6-yard gain on 3rd-and-8? What about massive H/R splits?

                  I’ll give examples of each: Ken Anderson, Derek Carr, and Drew Brees. Anderson is underrated outside the analytic community, but wildly overrated within it. Here’s a quote most people haven’t seen before, from Bill Walsh to Dr. Z: “A team can pad the stats for its passer by letting him throw a lot of dink passes at the end of the game … I know. I did it for Ken Anderson in his first few years, when I was the Bengals’ offensive coach.” I wish there were an easy way for statisticians to throw out garbage time data. Analyzing QBs on bad teams is so hard. Should they get a break because their supporting cast was weak? Or do they deserve demerits for padding their stats in meaningless moments?

                  Speaking of meaningless, in Week 10 last season, Carr completed 30 passes and only picked up 9 first downs. At least half your completions should produce first downs. Games like that, players like that, aren’t as good as the stats suggest.

                  In 2008, Drew Brees was a lion at home (20 TD, 5 INT, 114.0 rating) and a lamb on the road (11 TD, 12 INT, 80.3 rating). He only played well in half his games. New Orleans went 2-6 on the road, and finished 8-8. This goes back to that saying about statisticians, if you have one foot in a fire and one in a block of ice, on average, you’re comfortable. The Saints had three home games they won by at least two TDs, and three road games they lost by a field goal or less. Brees could have been less spectacular at home, and it probably wouldn’t have made a big difference, but if he had been above-average on the road, the Saints might have been a playoff team. His overall stats don’t reflect his value to the team. Drew Brees is a great player, but his stats are misleading, and my system consistently overvalues him.

                  I apologize for my over-long replies, but you keep raising points that interest me.

                  • Bryan Frye

                    Regarding this: At least half your completions should produce first downs.

                    Since 2002, the highest first down percentage any starting QB has produced in a season was 45.04% by Peyton Manning in 2004. If a full time QB gained first downs on half of his passes, he’d probably go down as the greatest of all time.

                    • Bryan, I’m afraid you misread my comment. There’s a difference between attempts and completions. Last year, every quarterback in the league was over 45.04% first downs per completion.

                    • Bryan Frye

                      Ha, I did read it backwards. But thank God you explained the difference between attempts and completions. I don’t think I ever would have figured it out without your help.

                    • I’m not sure you’re in a position to make sarcastic comments about reading comprehension just now ๐Ÿ™‚

                    • Bryan Frye

                      I’m seated in front of a computer. I’d say that’s the perfect position for making comments on the internet.

                      I get it. You made a comment, and I misread it and sent some irrelevant facts your way. You realized I misread your comment and, instead of simply stating such, you replied with a bit of snark (perhaps because you thought “who is this prick to correct me when he can’t even read; I don’t know). In turn, I replied with more snark. We both took the easy way out, and we could keep doing so if we really wanted to. Or we could realize that we are grown men who share a common love of football and put the silliness aside. I respect you as a human and as a football thinker, and I suspect the feeling is mutual. I think we can agree that getting off the sarcasm train is the way to go.

                    • Kibbles

                      Honestly, I didn’t think Brad O.’s initial response was particularly snarky. He diagnosed the problem, (you’d misread his comment), he offered clarification, (he was talking about completions, you were talking about attempts), and then he provided some background data to support his initial statement, (50% first down rate on completions is a quite common achievement).

                      But then again, that’s just my own attempt at divining a stranger’s tone and intent solely from text with few contextual aids- a task that has often humbled more gifted communicators than myself.

                    • Bryan Frye

                      As a frequent victim of Nathan Poe and his laws, I can definitely see where you’re coming from. I’m thinking of just reading everything at face value and never taking anything in print as sarcasm again. I think it would make reading the Onion a lot more fascinating.

                    • You’re right, I wasn’t trying to embarrass him. I just wrote a very literal response to address his error. When I’ve been misunderstood, I try to reply as clearly as possible.

                  • Adam Steele

                    I love long replies! More information to debate.

                    I don’t think garbage time is the issue as much as facing prevent defenses. Of course they often go hand in hand, but not always. For example, a two minute drill at the end of the first half will often pad a QB’s stats with easy completions against a soft defense, but it’s definitely not garbage time. If I had the data and ability to mine said data, I would discard all passes against prevent defense, because I don’t think those situations provide an accurate measure of QB ability. In reality we just have to live with the distortion, and make mental adjustments in extreme cases like Palmer in 2012. Both Carr brothers have been guilty of prevent inflated seasons, Jason Campbell in `08 and `09, Tim Rattay’s entire career in SF, and many others that I can’t think of off the top of my head. Check out Craig Nall’s small sample in 2004; he looks like a world beater until you realize ALL of his passes came during meaningless end game situations. I have never heard that quote from Walsh about Anderson, but I agree that he’s actually overrated by the analytics crowd. I do remember Scott Kascmar mentioning that Anderson’s playoff numbers are padded by lots of garbage time, but without PBP it’s hard to know exactly how much.

                    Funny you mention Brees’ 2008. That’s like my poster boy season for overrated high volume numbers. Yes his efficiency looks good on paper, but as you showed, that’s quite deceiving. Without looking at the actual numbers, I would guess Brees has the largest home/road and indoor/outdoor split of any great QB. Kurt Warner has a major decline outside the comfort of a dome, as does Matt Ryan. In 2014, Rodgers and Peyton both had serious home/road splits, although they were still great overall. I have a couple semi-plausible ideas to combat this phenomenon: Divide each QB season into separate home and road seasons, and measure the home seasons against a higher baseline. Or calculate a separate rating for home and road, then take the harmonic mean of the two ratings as the overall rating.

                    When are you going to publish your formula? I’m very curious!

                    • I agree about prevent defense, though I also think it’s unusual for a team to maintain great defensive play when they’re up by 30. Even if they’re not in a prevent, per se, you frequently see backups at that point. But yeah, I think it’s safe to say we’re on the same page here.

                      That’s an interesting idea regarding H/R splits. The harmonic mean idea is particularly intriguing. Have you done any poking to see how it comes out?

                      As far as my timeline, the plan right now is to begin my subjective series next week. It’s a seven-part series (over seven weeks) on the top QBs of all time — and it is my actual, personal list, so I can’t blame any critiques on a statistical formula! — and then to publish my stat-based methodology.

                      The stats help shape my opinion, but I won’t refer to the mysterious formula in the subjective series (and everyone here has seen the results anyway), so I don’t think that will create any problems for readers.

                    • Adam Steele

                      I have not yet experimented with the harmonic mean idea because it has a couple issues I have to figure out how to deal with. One, you can’t use negative numbers, so any value based metric with a zero baseline would have to be recalibrated. Two, I’m not sure what to do if the splits have a significant gap in the number of attempts. A weighted average doesn’t work with a harmonic mean like it does with the standard mean. Any ideas?

                      I’ll be looking forward to your series. Hopefully I can learn about the QBs I’m less familiar with because there are some major holes in my historical knowledge.

                    • re: H/R splits and the harmonic mean idea — I have lots of ideas, but none of them really practical. As you noted, there are some tricky issues to work around, and the nature of the project suggests more work to be incorporated (do you control for road environments, etc.).

                      I hope the QB series lives up to your expectations. This isn’t primary research, so it’s not ground-breaking, but hopefully I’ve collected a lot of information in one place, made it easy to access, and performed some worthwhile analysis. But I also hope I haven’t built this project into something it’s not — it’s just my own, personal list, with some stories I find interesting and explanations of why players rate where they do.

                    • Adam Steele

                      That would be a whole ‘nother project. Maybe it’s one I’ll take on…

                      I’ve never seen a detailed subjective ranking with commentary from an analytics guy. I’m more curious to hear the reasoning behind your rankings than the numerical list itself. I respect your ideas and want to see how your approach compares with mine. If nothing else I’m sure we’ll have more stuff to debate!

                    • The first article will post this Tuesday. Please make comments and ask questions! The series is written for readability more than detail — many of these players have had entire books written about them, so I’m mostly going for overviews — but I’m happy to explain (and debate!) in the comments.

                      Also, a heads-up: some of these bits you will have read in the comments on this post. Specifically, with regard to Roman Gabriel and Drew Bledsoe.

  • Michael Carlson

    Obviously, it’s not a quarterback rating, per se, but a passer rating. The thing that jumps out is Bart Starr and Len Dawson (34,37) behind Roman Gabriel, John Hadl, Boomer, Jims Hart and Everett. Hadl and Hart are personal faves but I wouldn’t rank any of those guys higher than Starr and Dawson. But they’re also behind Bobby Layne, who you might think of as a similar type. Yet Terry Bradshaw is behind Starr and Dawson. It suggests those guys lose a team context benefit in your system? Makes me very curious about the methodology…

    • It’s a quarterback rating (rather than pure passer rating) in the sense that it accounts for things like rushing and fumbles, but yeah, this rating is exclusively based on regular-season stats. Great postseason players like Starr and Bradshaw are underrated. Dawson seems low to me, too.

  • John

    John- The thing with marino for me is that he peaked very early in is career(84-88) and had a big decline of mediocre and terrible seasons in the second half of his career. He was also a bad decision maker and careless with the football.

    • Richie

      “He was also a bad decision maker and careless with the football.”

      Where are you getting that? Marino had a 3% interception rate for his career. Joe Montana was playing in the same era and threw 2.6% interceptions. That works out to about 1 or 2 interceptions per season difference. And Montana generally had better teams around him.

      Marino’s final season was a disaster, but I don’t think he was outlier-bad at decision-making and ball-protection for his era.

  • Bryan Frye

    It’s interesting to see the differences and similarities between this list and some of Chase’s lists. The biggest difference is Norm Snead, who is 41 here but I don’t recall ranking in any of Chase’s lists.
    Matt Schaub ranks about 40 spots lower here, which I think probably jives with public perception.
    Chad Pennington rates pretty low, despite having incredible efficiency numbers, which leads me to believe there is more emphasis on volume here (it would explain Favre’s ranking).
    You have Michael Vick rated way lower than Chase does, which I agree with.
    You have Dawson rated way lower than most people do, and you have Brady rated lower than almost everyone does.
    Trent Green’s ranking seems low to me, as does Steve Young’s.
    Your rankings for Aikman, Warner, Bradshaw, Layne, and Kelly seem low at first glance; until you recall that the RANY-based list has them ranked much lower. You can throw in Elway and Moon, too. Your rankings for them are closer to what I imagine the general public would think.
    I see very little variance from guys like Manning, Marino, Tarkenton, Unitas, Graham, Fouts, Brees, Starr, Anderson, Rivers, Romo, and Montana. Statistically speaking, you both have John Brodie way higher than the average football fan would (considering the average fan has probably never heard of him).

    Two guys I find interesting here are Cam Newton and Jim McMahon. You have them at 119 and 97, respectively. In some of Chase’s most recent rankings, Newton ranked 95,133, and 99. McMahon ranked 86, 113, and 90. This is especially neat in Newton’s case, as your rankings differ by 24, -14, and 20 (McMahon is less pronounced, at 11, -16, and 7).

    • Thanks Bryan, a lot of good stuff here. Also, congratulations on being the first person to notice Norm Snead. That’s one of my least favorite rankings. I would love to flip him with Bob Griese. What the system likes about Snead is his yearly volume — he threw a lot of passes, almost every season — and his average yardage, which is pretty good for his era. He was a productive goal-line runner, and his huge number of INTs is partly offset by a low fumble rate.

      I notice two themes in your comments, and please correct me if I’m mistaken:

      1. Volume passers too high (Snead, Favre) and high-efficiency passers with short careers too low (Young, Pennington, Green). Or maybe it’s mostly the latter, since you seem okay with the rankings for high-volume players like Marino, Tarkenton, Unitas, Fouts, and Brees.

      2. More closely matched to public perception than RANY.

      That second point could be good or bad, or (perhaps most likely) good on some players and bad on others. If you have any further thoughts, I’m interested in them.

      • Bryan Frye

        I’d say you’re mostly right about both points. I think peaks are more important when judging overall greatness (although I certainly understand why you would prefer a really long career of good over a short career of great). By no means would I ever argue that Pennington should be ahead of Favre. I don’t know if anyone can honestly say that he or she believes that to be true….even Chad’s parents.

        I don’t take issue with high volume passers, so long as the high volume consists of good passes. I wouldn’t want people 40 years from now looking back and thinking Matthew Stafford was one of premier quarterbacks. The guys you mentioned, however (Marino, Tark, the Original Johnny Football, Fouts, Brees), were efficient over a high volume of passes. I think each of them had at least one year when he was the best QB in the league (I argued for Brees as 2009 MVP).

        The second point is neither good nor bad. I just notice that your ordinal rankings match up more closely with what the general public tends to think (except for Brodie, of course).

        As I said to Richie, I don’t feel super comfortable judging the rankings without knowing how the sausage was made. Every process based solely on stats is going to produce some rankings some people disagree with, and that’s okay. I’ve often felt that if everyone is in agreement on something, you might want to think about why that is. I am excited to see the actual process behind your rankings, either here or on SC.

        • Thanks, Bryan. Don’t worry, the methodology is coming … eventually. You can expect my subjective list first, shortly after the NFL Draft, and then I’ll get into the process that produced these numbers.

          In the meantime, just wanted to note that I agree on Brees in ’09. He does score as the top statistical QB that season, as well.

          Unrelated note, I remember that you liked my comments on Shane Lechler in the 05-14 All-Decade piece … if you read the punter section in the link above, there’s a stunning stat about Lechler.

          • Bryan Frye

            I have a bunch of different homegrown stats for measuring QB play, but there’s always a weird ranking in there somewhere. For those who consider QB wins a real stat, I have one stat where championships are worth a certain number of points. It’s sort of like a much better version of the legacy tracker that the new Madden games feature.

            I have another one that just weights games differently, with each round of the playoffs worth more points. The issue with that, of course, is that it completely screws over older guys who didn’t have anything like the modern playoff format. There is a guy named Christian Bolton who came up with something called the single number back indicator (SNBI) that gives more credit for stats accrued in successive playoff rounds. I don’t agree with his methodology, but it is worth a read just for the sake of understanding another perspective. https://sites.google.com/site/snbifootball/home

            I have other ones where I weight years and leagues differently (e.g., winning a title in 1942 is much less impressive than winning one in 2014; winning the 1961 AFL title is less impressive than winning SB1). There are copious ways to come up with these lists, and they’re probably all wrong. It’s still fun to make them.

          • Bryan Frye

            Oh, and if another punter ever makes the HOF, it will probably be Lechler, whether he deserves it or not.

    • Richie

      I would have expected Favre to come out higher if compilers were emphasized too heavily.

      • Bryan Frye

        Yea, I’m still trying to figure it out. Montana ranks over Favre, but Young doesn’t. Favre ranks over Brady. I’m not sure what kind of formula Brad is using, so I don’t really feel super confident judging the output without knowing the input. Compiling years aside, Favre’s prime was about as good as most QBs ever get.

        • “Favre’s prime was about as good as most QBs ever get”

          We have got to put you and Adam Steele in touch with each other.

          • Bryan Frye

            Considering most QBs never even make a practice squad, I stand by my comment. Now, if we’re only talking HOF QBs, then that’s another story.

  • There’s pretty clearly a bias towards compilers, specifically passing yardage compilers. Vinny Testaverde is tied with Terry Bradshaw. They both went 1,1 in the draft and were both QBs but that’s pretty much the end of their similarities.

    Roman Gabriel’s ahead of Kurt Warner, Bart Starr and Big Ben. Steve DeBerg and Dave Krieg ahead of guys like Earl Morrall and Babe Parilli. John Brodie! ahead of Aaron Rodgers.

    Being really durable is a skill, but it’s probably not the most important quarterback trait.

    • Thanks, Chris. I am concerned about a possible bias toward compilers. You compared Bradshaw to Testaverde, and there’s one more similarity you didn’t mention: they both began their careers with bad teams, posting horrible efficiency numbers, before playing well later on.

    • Kibbles

      The skew towards longevity was the first thing that I noticed, too, though I can’t say whether that’s a bug or a feature. One could argue with a perfectly straight face that just because a guy played longer doesn’t mean he wasn’t actually a little bit worse. One could also argue with a perfectly straight face that the guy who played longer was more valuable to his franchise, even if he was a little bit worse. It’s just a question of what you want to calibrate your list to reward and/or measure. I think Johnny Unitas was probably slightly “greater” than Fran Tarkenton, but at the same time, I wouldn’t disagree with anyone who said that Tarkenton added more value over the course of his career.

  • JoeS

    Having grown up with football in the 70s, it’s strange to see Staubach rated so relatively low at #19. He retired with the #1 QB rating of all time in the modern era. And, yet, he’s behind Unitas, Jurgenson, Tittle, Anderson and Brodie on this chart.

    Unitas, you can make an argument about. The others are highly debatable. And, considering that Brodie and Staubach were direct rivals, I don’t think even most fair-minded 49ers wouldn’t have switched QBs at the drop of a dime.

    • Staubach seems low to me, too, and it’s weird, because my system loves him on a year-to-year basis: he rates as the top QB four times. The only others who do are Otto Graham, Steve Young, and Peyton Manning. I don’t think there’s really any question that Staubach was the best QB of the 1970s.

      I agree about Staubach and Brodie, but the rating system says that 13 years of Brodie is just as good as 8 years of Staubach and 5 years of a replacement-level QB. Like you, I’m not sure I believe that. As you noted, Staubach doesn’t get credit here for his postseason play. You’re the first person to mention Staubach; I’m curious to know how others feel about his ranking.

      • James

        Tangotiger, one of the if not THE best baseball sabermatrician, uses a much higher replacement level when calculating GOAT/HOF rankings. His preferred method is to use Wins above *Average* not Replacement, and to not count negative seasons.

        Essentially, peak greatness matters more than long term good but not great play, and it doesn’t penalize players that stick around past their prime. It better reflects the fact that a couple 13-3 seasons are more likely to result in a Super Bowl than a handful of 10-6 seasons, and fans cherish great seasons disproportionately more than good ones.

        • JoeS

          Where does this Tangotiger place Staubach, then?

          And, Brad O., thank you for your promp reply. Staubach, like Ted Williams in baseball, seems to be one of those tricky players who will never get their full due because each of them lost approximately 5 seasons of their prime to being in the military. One of the effects is that both of them are ‘penalized’ for their last couple of seasons where there ARE numbers, compared to numbers for those ‘missing’ seasons which we will never know.

          It would be an interesting experiment for someone to conduct to try and predict what Staubach’s numbers would have been if he had played even one or two more full seasons (the Cowboys did have Meredith and Morton, so it’s doubtful it would have been five).

          • James

            “Where does this Tangotiger place Staubach, then?”

            Far, far away from the baseball HOF, that’s for sure.

  • Thanks so much for the thoughtful feedback, everyone. Please keep it coming!

  • Cotcat

    It appears your formula places too much emphasis on volume and not enough on efficiency. Even though Wilson/Luck haven’t been in the league as long as Newton, they should both still be well above him. Other big volume guys/guys with longevity (i.e. Favre) seem too high as well

    • Richie

      Newton over Luck is pretty crazy when you figure Newton only has 1,500 more yards despite one more full season. Luck has more TD’s and a better interception rate.

      But, I can’t get worked up about guys who only have 3 or 4 seasons. There will definitely be unusual results for guys without much data.

      • Picking nits, but it’s not one more full season — Newton was injured last year and only threw 82% of the team’s passes. And you can’t compare Newton to other players if you don’t include his rushing. Newton actually has 2,600 more yards than Luck, and 17 more touchdowns. Newton has nine more turnovers (-11 INT, +2 FMBL).

        Luck’s rookie season does relatively little for his score in my system. Luck’s ’14 rates better than any of Newton’s years, but Cam has three good seasons, and Luck has two. Luck’s going to pass him, but I think it’s easy to forget how well Newton played his first two seasons.

  • sacramento gold miners

    Neil Lomax is a QB who had terrible luck of playing with a struggling franchise, then a hip injury ended his career early at 29. He could be among the best under the radar QBs in league history. Excelled at the Run & Shoot at Portland State, and starred with the Cardinals. Those teams did have some weapons in Roy Green and J.T. Smith as the wideouts, and Ottis Anderson and Stump Mitchell in the backfield, but the defense was porous.

  • Ryan

    Excellent list and thanks for sharing…Taking your list in chunks:
    Top 10: Tom Brady has an argument for Top 3 regular season value…not sure how you have him at 9th?
    Brett Favre is a case of replacement level, a Top 13 regular season performer at worst…6th might just be on the high end valuation?
    Tarkenton was awesome, glad to see him at 3, just wish he could have won at least 1 Super Bowl!

    11-20: Like many, I see Steve Young as a Top 5, at worst Top 9 guy…a peak argument for sure.
    Elway places ~20 group on pure regular season stats, do you have some factor for this bad teammates?
    Moon is a tough one, seems like low peak, fine career…would be cool to see you put in a value for CFL years.
    Love the placement of criminally underrated John Brodie (Chase had a post discussing strength of schedule and I remember Brodie having a brutal if not all-time toughest schedule…maybe this causes us to generally overlook him?) and riser but not that yet all-time great Aaron Rodgers.

    21-30: Boomer and Gabriel seem a tad high, but defensible.
    I like the modern guys a tad more in Rivers and Romo, but again defensible.
    Interested in how you came to Kelly, Layne, and Everett…they seem more like 40s guys.

    31-40: Namath seems a tad low, but Dawson has an argument for ~20…not sure why he is this low.
    Cunningham is this good due to great legs?
    Testaverde regular season is comparable and maybe better than Bradshaw.
    Maybe Vinny is too high, but Terry would be also and maybe more so…great post-season career, so-so regular.

    41-50: Like others mentioned, I could see Trent Green ~10-15 spots above.
    The Mad Bomber has an even stronger peak argument then Green, the 25-30 range might not be out of line.
    Stabler feels a tad high, while I agree that I am not seeing Norm Snead.

    The rest: Schaub seems overlooked, maybe Landry, Rypien, Pennington, and Rypien deserves bumps.

    • Thanks, Ryan, I like your groups-of-10 approach.

      I agree about Jim Everett, but I would love for you to expand on why Kelly and Layne seem more like 40s guys. Both were first-ballot HOFers, and they ranked 19th and 31st, respectively, in Adam’s crowd-sourcing project.

      Keep in mind for Namath, he’s basically tied for 31st. Cunningham, like everyone else, is rated as a complete player. Both passing and rushing contribute to his position. I don’t see how we can evaluate Cunningham without considering his rushing.

      • Ryan

        For Kelly, relatively short NFL career of 160 starts, offset with only two excellent peak season in 1990/1991…adding in USFL and post-season gives him a Top 30 argument though.

        For Layne, it’s a disservice not too include his two NFL championships…maybe I am underrating his fairly lengthy/medium peak career?

  • Perhaps it was inevitable that several commenters would compare my list to Chase’s work. To make things easy for you, here are some of the most dramatic differences between my stat-based ranking and the one Chase published last year. I subtracted pre-Modern Era players (Luckman, Baugh, etc.) from Chase’s list, so we’re comparing apples to apples.

    Brad Higher (my rank first)

    John Elway: 12, 23
    Warren Moon: 17, 32
    Jim Kelly: 24, 51
    Bobby Layne: 25, 40
    Jim Everett: 27, 45
    Randall Cunningham: 36, 52
    Terry Bradshaw: 38, 53
    Norm Snead: 41, >89
    Drew Bledsoe: 51, >89

    also: Charley Johnson, Phil Simms, Billy Wade, Neil Lomax, Joe Ferguson, Tobin Rote, Jeff George, Don Meredith

    Chase Higher (Chase’s rank first)

    Len Dawson: 18, 37
    Daryle Lamonica: 27, 47
    Trent Green: 30, 46
    Bob Griese: 42, 56
    Matt Schaub: 44, 90
    Greg Landry: 46, 83
    Earl Morrall: 48, 69
    Carson Palmer: 49, 70
    Johnny Lujack: 55, 93
    also: Mark Rypien, George Blanda, Chad Pennington, Michael Vick, Doug Flutie, Bill Munson, Wade Wilson, Russell Wilson, Cam Newton

  • Looks pretty good to me as someone relatively familiar with all of the different QB posts Chase has done throughout the years. I’d probably only make very minor tweaks here and there, depending on what variables you’re using.

  • Tim Truemper

    Two basic statistical questions/observations come to mind. Once you get below 35 to 40, the range of the rankings seems rather restricted. That is, 120 QB’s are within 30 points of each other. Above that and it shoots up about 20. Also, are the differences between the numbers you use on the scale equal intervals of difference? If its a ranking system only then I can see that it would not be equal intervals per se. Have not looked at the other comments but want to say that I appreciate any effort at objective rankings of player careers and performance.

    • The numbers reflect that there’s a bigger difference between #1 and #20 than between 21 and 40, a bigger difference between 21 and 40 than between 41 and 60, and so on. So for instance, my system comfortably rates #7 Otto Graham (33.9) over #12 John Elway (28.0), but it doesn’t provide a lot of confidence that #77 Don Meredith (9.84) was better than #85 Doug Williams (9.13) … it’s not easy to distinguish between the 80th- and 100th-best players: the distinctions are minor, and it depends heavily on what you value.

  • Joseph

    I found Brad’s other interesting piece, on ranking the best QBs by decade, consistently more compelling. But then it probably is much easier to pick out the very best over shorter time frames than a larger ranking like this one. As others have commented, I think the most questionable ranking has to be Staubach at 19 (and below the erratic Brodie? below Moon?). If we tried a thought experiment, and imagined we were GMs drafting for QB, I suspect Staubach would not be drafted lower than 5. Sonny Jurgensen is my all-time favorite performer, for flair and the sheer magic of his passing, but Roger would have to rank higher, objectively speaking. Fran at 3 is too high in my opinion (first team all-pro only once). And with the GM draft experiment, I doubt he would be drafted in the top ten. Steve Young for me is a much better QB than Brees (something of a stat-padder who, like Manning, loves to add to this TD total with dink passes under 5 yards from goal — back in the day, that glory went to running backs).
    Ideally, it would be great if we could combine quantitative analysis with expert qualitative judgments across the eras. For example, in the 70s there was a magazine called Pro Quarterback that rated QBs each season, with๏ปฟ input from team scouts. Used a 5
    pt scale for Leadership, Reaction to Pressure,
    Set-up, Throwing, Reading Defense. 1972 issue: Lamonica
    15.5 total. Gabriel 16. Brodie 16.5 – [only 2.5 for
    Pressure]. Fran 17 [only 3 for Throwing]. Staubach,
    Dawson, Griese each 18. Sonny, even at age 38, scored
    20.5, the second highest total. Top-rated was Namath,
    with 23.5 [who I would score lower, due to interception
    Thanks for all the hard work!

    • Thanks, Joseph, I really appreciate your input, both here and at SC. And I’m always pleased to see people advocate for Sonny Jurgensen. As I wrote last year, Sonny is “without question, the most underrated quarterback in the Pro Football Hall of Fame.”

      I agree with most of your comments here, and I think you’ll enjoy my series next month on the greatest QBs of all time.

      • Joseph

        Very much look forward to that series. For me, the biggest challenge in all-time rankings of QBs is factoring in the immense changes in the game over the last 60 years or so. I know there are sophisticated techniques to adjust statistically for things like “Dead Ball” era vs. “Live Ball,” etc. But there always remains a significant degree of ambiguity because the nature of the passing game has evolved so dramatically. To take one example, I recall a Grantland article by Kirk Goldsberry, “Quarterbacks Illustrated” (2014), which contained the following stat: “Overall, 40 percent of NFL passes targeted receivers less than five yards away from the line of scrimmage. League-wide, quarterbacks completed 74 percent of them.” No surprise there! Before rules changes, DBs could of course hammer WRs at the line and all over the field before the pass was in the air. Also consider all the “protect the QB” changes (remember when Unitas heroically jammed mud up his broken nose, and then let fly a bomb to win the game!), or letting offensive linemen use hand-grip blocking, the use of special gloves for receivers and QBs …

        And then there’s the elusive “team factor”. My favorite comparison here is Archie Manning and Terry Bradshaw. Is there any doubt that if Archie had been drafted by the Steelers, and Bradshaw had gone to the Saints, it would be Archie with the rings and the HOF honors?

      • Joseph

        Dear Brad,

        Just revisiting this and a couple of other anomalies jumped out. Starr at 34, below the likes of Everett and Hart at 27 and 26? Is Bart being penalized in your calculus for low volume, as in passes, tds and yardage per season? I agree Starr could put up higher efficiency passing stats because of the greatness of the Packer teams and its formidable line and ground game, but he still was consistently clutch. I would put Len Dawson, who looks surprisingly low to me at 37, in the same category with Starr. Not a great arm, but he knew his limits and played with great poise, racking up consistently strong performances relative to his peers. I also think Norm Snead is way too high at 41, and I’m 100% confident that no real or imaginary GM would ever pick him to play QB over Snake Stabler ๐Ÿ™‚

        Looking forward to your full report in due course! Thanks again, Joseph

        • Starr is definitely a player I wish my formula rated more favorably, and you can expect him higher than that in the upcoming series. Starr’s biggest problem, statistically, is the absence of an obvious peak. He was NFL MVP in 1966, but my stat-based system doesn’t see it that way. Volume, as you guessed, is part of the issue. By way of example, let’s compare Starr in ’66 to our friend Sonny Jurgensen.

          That season, Sonny threw for 1,000 more yards and twice as many TDs. He also had a lot more attempts, and a lot more interceptions. But it’s not like Sonny was an inefficient passer: based on his production, my formula believes that every pass Jurgensen threw helped the team. Starr threw so infrequently that if you didn’t know GB’s style of offense, you would guess he was a part-time player, probably injured for a third of the season.

          Put another way, would you rather have 462 pass plays by Jurgensen, or 277 by Starr? I don’t blame the computer for saying Jurgensen.

          I agree about Dawson and Snead, too. Dawson is hindered by a relatively brief starting career (he threw 45 passes his first five seasons) and by the way I treat AFL stats. In the AFL, Dawson posted an insane 7.7 TD%, by far the best of that era, and a figure that would rank number one all-time. Over the rest of his career, that figure was 4.1%. It’s not an entirely fair comparison, because his prime years were in the AFL, so of course he was better then. But I don’t think Dawson’s excellent efficiency stats accurately reflect his play relative to contemporaries in the NFL, and I don’t believe Dawson was ever the best QB in football.

          I really struggle with what to think of Snead. There’s a little bit of an Archie Manning comparison, because he played mostly for bad teams. He had a long, productive career, and he made three or four Pro Bowls (I’ve seen both figures, and it drives me a little crazy that I haven’t been able to ascertain the correct one) — he was a pretty good player. I struggle to rate guys with weird careers. But you’re right that 41 is way too high, and no one would actually prefer Snead to Lamonica and Stabler.

          • Joseph

            Needless to say, I will like any computer that has the good sense to recognize Sonny’s greatness!

            I’m also very keen to see what your formula does do with the AFL stats. I personally would qualify everything for the first 5 years or so of the new league. One could also make a stronger case for discounting the AAFC stats. Just look at Otto Graham, an undoubted all-time great, but when the Browns joined the NFL in 1950, his passing stats took quite a hit: from ratings of 112, 109, 86, 97 down to 65, 79, 67, 100, 73, 94. The competition was just so much more consistent and deeper. Same with the early AFL, filled with players cut from the NFL or who were frustrated being buried so deep on the pine.

            Incidentally, there’s an interesting new book out called Football Morsels: Quarterbacks, by Julio Castaneda. Haven’t digested it fully, but he offers a “Best Combined QB Skills Adjusted” rating that places Sonny third highest, behind Montana at two and Manning at one. That said, there are plenty of jarring scores as well (Brodie at seven, behind Marino at six?? Staubach at thirteen, just behind Romo, and Johnny U at fifteen, seven places behind Brees???).

  • Thanks so much to everyone who has provided feedback! Any further thoughts are certainly welcome, but I think I’m getting a good sense for which parts of the list seem off to an informed community of readers.

    The most notable names listed as overrated are Roman Gabriel and John Brodie. That’s interesting, because they’re basically contemporaries, who had their best seasons in the late ’60s and early ’70s, and because that’s an era where my rankings tend to differ from Chase’s system — and he has many players from that era rated higher than I do.

    Underrated, I’ve heard a lot about Steve Young and Roger Staubach in particular, as well as Len Dawson, Tom Brady, and Trent Green. I tend to agree on the first four, but Green surprises me. Watching Green play, I never felt like I was seeing a great quarterback. Even subjectively, I wouldn’t put him ahead of contemporaries like Gannon and Garcia. But Green’s inclusion is consistent with the theme several of you noticed: a sense that brilliant players with short careers are underrated.

    I would really appreciate if anyone can offer input on whether you think it’s more likely my system is overrating volume or longevity. Am I giving too much credit to guys who pass a lot, or who play for a long time? I know that’s hard to assess without seeing my methodology, and I really will reveal that in the future, but in the meantime, thanks for any additional insight you can offer, and for the feedback you’ve given already!

    • Lord Dre X

      I have to say I think you favor longevity over volume.

      To make my point, I use the comparison of Matt Ryan vs Dante Culpepper.

      The two have stats that are “similar” as are their scores. Culpepper has a better Y/A and TD%, but Ryan has better int%, yards/game, completion %, and better sack %.

      Culpepper has 11 seasons in the league, vs Ryan’s 8, with Ryan having roughly 700 more attempts and roughly 500 more completions in his career.

      I am doing this on a phone so I can’t dig too hard, as maybe Culpepper’s rushing numbers inflate his rating over Ryan’s, thought I would think this would be negated by the advantage Ryan has in overall passing stats.

      I also suspect that your rankings adjust for the inflation of passing stats over the last few years, so that may also be to blame.

      I hate the Falcons, and if they lost every game to Carolina by 50, I’d be mad they didnt lose by more, but Matt Ryan is a very good quarterback that everyone forgets about.

      A 1.3 point difference in your system is the same difference between Steve McNair and the Donovan McNaab, Troy Aikman, and Tony Romo tier. Or in the difference between Joe Theisman and Dante Culpepper. I can’t see Culpepper and Ryan being so far apart.

    • Adam Steele

      Here’s my argument for Trent Green……From 2002-2005, his Chiefs scored the most points in the NFL, with the Colts being the only team within 250 points (!) of the Chiefs’ 1,837. Green’s efficiency stats were among the league leaders in all four seasons. How many QB’s in history have the led the league’s best offense over a four year span? But here’s the kicker – KC only made the playoffs in ONE of those four seasons. A combination of bad defenses and a stacked AFC left Green on the sidelines come playoff time. In his lone playoff game in 2003, Green played well and led his team to 31 points but lost anyway. That is horrendously bad luck from a QB legacy perspective. Given the firepower of his offenses, Green probably should have made the playoffs in all four seasons, and from there it wouldn’t have been a stretch to see him reaching or winning a Super Bowl.

      There’s also the 1999 what-if that probably influences some peoples’ opinions. After seeing Green post 8.6 Y/A, 16 TD, 5 INT in relief for the 2000 Rams, it’s natural to conclude that the 1999 offense would have been nearly as good with Green as it was with Warner. Of course legacies are built on what DID happen, not what might have happened, but it’s hard not to give Green a few sympathy points for his terrible luck.

      The knock on Green is that he played with great offensive teammates, which likely inflated his passing numbers. That’s a fair criticism, but many great QB’s also played with a strong supporting cast. Except in rare cases, I believe the QB is the main driver of his offense’s success/failure, so I give Green a lot of credit for his 4.5 fantastic seasons. I think Green was slightly better than Gannon or Garcia because he could throw intermediate and deep better than either of them (Gannon and Garcia were classic WCO short passers). He was also an underrated scrambler who could pick up key first downs with his legs. I’m not trying to argue that Trent Green is a legend, but I don’t think it’s unreasonable to rank him in the 30-35 range all time.

      • Green really was amazing. It’s just too bad he wasn’t better at throwing blocks.

        • Adam Steele

          He didn’t have the swagger to inspire his defense like Trent Dilfer.

      • Green is profiled in the upcoming series, so I’ll save most of my comments in anticipation that this discussion will continue then. But I will say that the great offensive teammates you mention influence my thinking a great deal, and Green did not pass the eye test for me. Even in his best seasons, he looked like a good QB, not a great one.

  • Tim Truemper

    Per your response to my comment on restricted range, thanks for pointing out your reasoning and I see the logic in your scaling.