SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.
Also, our existing PFR blog rss feed will be redirected to the new site's feed.
Pro-Football-Reference.com » Sports Reference
For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.
ShareThisGreatest QB of All-Time III: Career rankings
Yesterday, I explained the methodology behind my grading of every quarterback-season in NFL history. Today, I'm going to present the career results. As usual, I'll be using the 100/95/90 approach, where each QB gets 100% of his score in his best season, 95% of his score in his second best season, 90% of his score in his third best season, and so on. This is the key to rewarding guys who played really well for a long time, but without killing guys with really bad rookie years or seasons late in their career. It also helps to prevent the guys who were compilers from dominating the top of the list. The table below shows the top regular season QBs in NFL history, using three different metrics.
"VALUE" shows each quarterback's converted yards over average, as explained in yesterday's post. "REPL" shows each QB's converted yards over replacement, defined as 75% of league average. I like using the "Value" score as a HOF indicator and to answer the question of who were the best quarterbacks ever. However, after the top 30 or so QBs, I like using replacement value, which rewards guys who were good for a long time. Being average for 10 seasons means you were probably a better QB than someone who was good for two seasons. Using three-fourths of league average as the baseline is probably the best to judge a large group of quarterbacks, like when we want to separate the Eli Mannings from the Jeff Georges and Trent Dilfers from the Ryan Leafs. If we have to rank a random 100 QBs in NFL history (or if we're trying to judge how good an average draft pick was), the replacement category is best. Deciding who was the 42nd and who was the 43rd best QBs ever? I'd use the replacement value category, but note that this only works well for players in the same era. The replacement value formula, for various reasons, is biased towards modern QBs. For both the "VALUE" and "REPL" metrics, I pro-rated non-16 game seasons in the usual manner, splitting the difference between pro-rating and not pro-rating at all (i.e., a 9-game season is pro-rated to a 12.5 game season).
If you want something that is totally era-independent, you'll want to use the third column, "SEARK." That formula measures each quarterback's rank in each season. If you were the #1 QB in the league in any season, you got 10 points; if you were #2, you received 9 points; #3, 8 points, and so on. This actually helps the older QBs since they played in smaller leagues, and therefore it was easier to accumulate more "SEARK" points; however, since older QBs did not stick around as long as modern QBs, I think this metric is pretty era neutral. Note that I combined the AFL and NFL QBs in the '60s for the purposes of the "SEARK" column, although each player was only compared to the other QBs in his own league for the VALUE and REPL categories. Additionally, all AAFC stats have been excluded (sorry Otto Graham), as the NFL does not officially recognize them the way the league does with AFL stats (although the HOF does consider AAFC performances).
Finally, I showed the main team each QB played for, along with what percentage of his career value came with that team. It is possible (see Daunte Culpepper) to get over 100% of your career value with one team, if you are below average with your other teams.
Here are the top 100 QBs in NFL history according to my formula, sorted by converted yards of value over average. An * means the player is in the HOF, while a + means the QB was active in 2008:
Name ATT VALUE RK REPL RK SEARK Team % Peyton Manning+ 5960 10492 1 16093 2 83 clt 1.00 Dan Marino* 8358 10410 2 16770 1 98 mia 1.00 Steve Young* 4149 8783 3 12735 6 77 sfo 1.02 Joe Montana* 5391 8451 4 13284 3 80 sfo 0.95 Fran Tarkenton* 6467 8286 5 12838 5 89 min 0.67 Dan Fouts* 5604 7192 6 11828 7 61 sdg 1.00 Ken Anderson 4475 6752 7 10337 9 53 cin 1.00 Johnny Unitas* 5186 6696 8 11233 8 78 clt 1.01 Sid Luckman* 1744 6529 9 7844 27 78 chi 1.00 Sammy Baugh* 2995 6351 10 8432 17 81 was 1.00 Brett Favre+ 9280 6157 11 13149 4 63 gnb 1.03 Roger Staubach* 2958 5987 12 8548 15 58 dal 1.00 Len Dawson* 3741 5733 13 8711 14 54 kan 1.01 Sonny Jurgensen* 4262 5524 14 9632 11 55 was 0.72 Norm Van Brocklin* 2895 5517 15 8234 20 82 ram 0.76 Otto Graham* 2626 5029 16 6600 42 50 cle 1.00 Y.A. Tittle* 4395 4762 17 8299 19 59 nyg 0.52 Kurt Warner+ 3557 4479 18 8300 18 33 ram 0.75 Tom Brady+ 3653 4442 19 8468 16 21 nwe 1.00 Daryle Lamonica 2601 4363 20 6839 40 43 rai 0.99 Jeff Garcia+ 3676 4328 21 8214 21 34 sfo 0.82 John Elway* 7250 4314 22 10256 10 36 den 1.00 Boomer Esiason 5205 4242 23 8777 13 47 cin 0.98 Joe Namath* 3762 4023 24 7258 37 36 nyj 1.01 John Hadl 4687 3975 25 7553 33 38 sdg 0.88 Jim Hart 5076 3965 26 7393 36 34 crd 1.00 Trent Green+ 3740 3962 27 7960 25 30 kan 0.91 Bart Starr* 3149 3948 28 6998 38 41 gnb 1.00 Drew Brees+ 3650 3872 29 7913 26 27 nor 0.76 Warren Moon* 6823 3848 30 9584 12 34 oti 0.84 John Brodie 4491 3821 31 7735 29 32 sfo 1.00 Bert Jones 2551 3818 32 6191 45 30 clt 1.02 Roman Gabriel 4498 3755 33 7474 34 29 ram 0.70 Steve McNair 4544 3717 34 8095 23 23 oti 1.01 Jim Kelly* 4779 3621 35 8138 22 29 buf 1.00 Rich Gannon 4206 3506 36 7572 31 27 rai 0.95 Donovan McNabb+ 4303 3451 37 7961 24 18 phi 1.00 Cecil Isbell 818 3440 38 4056 93 41 gnb 1.00 Troy Aikman* 4715 3439 39 7774 28 36 dal 1.00 Bob Griese* 3429 3384 40 6008 48 23 mia 1.00 Mark Brunell+ 4594 3018 41 7566 32 18 jax 0.98 Arnie Herber* 1175 2995 42 3414 115 72 gnb 0.90 Jim Everett 4923 2944 43 7696 30 25 ram 0.84 Charlie Conerly 2833 2891 44 5383 59 51 nyg 1.00 Terry Bradshaw* 3901 2885 45 6090 47 31 pit 1.00 Craig Morton 3786 2814 46 5593 54 19 dal 0.67 Daunte Culpepper+ 3042 2806 47 6254 44 24 min 1.13 Vinny Testaverde 6701 2695 48 7416 35 17 nyj 0.46 Randall Cunningham 4289 2692 49 6858 39 18 min 0.59 Joe Theismann 3602 2679 50 6144 46 27 was 1.00 Tommy Thompson 1424 2671 51 3959 97 33 phi 1.02 Steve Grogan 3593 2669 52 5509 56 18 nwe 1.00 Earl Morrall 2689 2643 53 5014 70 23 clt 0.41 Billy Kilmer 2937 2640 54 5020 69 25 was 1.07 Ken Stabler 3793 2624 55 5441 57 22 rai 1.26 Bobby Layne* 3700 2546 56 5798 51 53 det 0.66 Mark Rypien 2613 2389 57 5039 67 18 was 1.00 Bernie Kosar 3365 2365 58 5714 53 24 cle 1.03 Matt Hasselbeck+ 3347 2330 59 5999 50 15 sea 0.99 Brad Johnson+ 4326 2264 60 6393 43 16 tam 0.36 Chad Pennington+ 2395 2251 61 5033 68 16 nyj 0.62 Johnny Lujack 808 2173 62 3048 124 25 chi 1.00 Milt Plum 2419 2156 63 4759 74 30 cle 1.43 George Blanda* 4007 2131 64 4697 75 14 oti 0.82 Bob Waterfield* 1617 2061 65 3566 111 36 ram 1.00 Carson Palmer+ 2165 1961 66 4557 79 14 cin 1.00 Tony Romo+ 1307 1959 67 3618 107 13 dal 1.00 Frank Ryan 2133 1954 68 4473 82 16 cle 1.04 Brian Sipe 3439 1937 69 5224 63 16 cle 1.00 Greg Landry 2300 1932 70 3907 99 21 det 1.06 Doug Williams 2507 1882 71 4508 80 12 tam 0.74 Philip Rivers+ 1428 1849 72 3655 105 15 sdg 1.00 Tom Flores 1715 1830 73 3582 109 18 rai 1.15 Don Meredith 2308 1827 74 4581 78 16 dal 1.00 Joe Ferguson 4519 1827 75 5175 64 24 buf 1.08 Jim McMahon 2573 1821 76 4351 87 9 chi 0.90 Phil Simms 4647 1754 77 6001 49 10 nyg 1.00 Ron Jaworski 4117 1710 78 5510 55 9 phi 1.05 Charley Johnson 3392 1707 79 4983 71 11 den 0.83 Bill Nelsen 1905 1681 80 3615 108 16 cle 0.81 Doug Flutie 2151 1666 81 3948 98 8 buf 0.88 Ed Danowski 605 1663 82 1953 144 30 nyg 1.00 Bernie Masterson 409 1653 83 1913 146 37 chi 1.00 Neil Lomax 3153 1614 84 5126 65 18 crd 1.00 Jake Delhomme+ 2434 1598 85 4431 84 5 car 0.99 Steve DeBerg 5024 1528 86 5720 52 10 kan 1.13 Danny White 2950 1502 87 4485 81 10 dal 1.00 Jeff Hostetler 2338 1484 88 4030 94 11 rai 0.61 Dave Krieg 5311 1479 89 5414 58 11 sea 0.46 Bill Kenney 2430 1469 90 4061 92 10 kan 1.00 Marc Bulger+ 2924 1440 91 4830 73 14 ram 1.00 Neil O'Donnell 3229 1429 92 4633 76 7 pit 0.80 Ben Roethlisberger+ 1905 1416 93 3778 102 9 pit 1.00 Wade Wilson 2428 1380 94 3554 112 14 min 0.71 Chris Chandler 4005 1363 95 4922 72 12 atl 1.19 Elvis Grbac 2445 1334 96 3993 96 7 kan 0.88 Jeff George 3967 1332 97 5325 60 18 atl 0.65 Norm Snead 4353 1309 98 5121 66 14 phi 0.71 Bobby Thomason 1346 1238 99 2637 129 24 phi 0.93 Babe Parilli 3330 1226 100 4067 91 19 nwe 1.46
I know what you're thinking: but what about the post-season? You're right, we need to include post-season stats in any debate about NFL QBs. I used the exact same methodology to grade the QBs in the post-season as I did in the regular season, although I weighted championship games by three times as much as a regular playoff game, and conference championship games were assigned double the value of other playoff games. Listed below are each QB's number of pass attempts, his career rating (his value score combined with his post-season score in each season), his career rank, his regular season value (same as above), his regular season rank (same as above), and his post-season score. Note that simply adding a player's post-season career grade to his regular season career grade will not match his new career score, because of the year-by-year approach. As many might expect, we have a new best QB ever once we include the playoffs, but I don't think he's going to stay #1 for very long. Note: This is not a ranking of the best post-season QBs of all-time, but rather a ranking of the best QBs when combining post-season and regular season performance.
QB Att Career Rk Reg Rk Post Joe Montana* 5391 11444 1 8451 4 3397 Peyton Manning+ 5960 10934 2 10492 1 474 Dan Marino* 8358 10787 3 10410 2 45 Steve Young* 4149 9950 4 8783 3 1182 Sid Luckman* 1744 7949 5 6529 9 1368 Fran Tarkenton* 6467 7862 6 8286 5 - 605 Ken Anderson 4475 7412 7 6752 7 667 Dan Fouts* 5604 7311 8 7192 6 97 Johnny Unitas* 5186 7175 9 6696 8 264 Brett Favre+ 9280 7124 10 6157 11 1017 Sammy Baugh* 2995 6797 11 6351 10 590 Roger Staubach* 2958 6746 12 5987 12 848 Len Dawson* 3741 6523 13 5733 13 909 Kurt Warner+ 3557 5915 14 4479 18 1535 Bart Starr* 3149 5613 15 3948 28 1949 Norm Van Brocklin* 2895 5608 16 5517 15 - 115 Otto Graham* 2626 5595 17 5029 16 776 Sonny Jurgensen* 4262 5447 18 5524 14 - 112 John Elway* 7250 5409 19 4314 22 1246 Troy Aikman* 4715 5303 20 3439 39 1830 Tom Brady+ 3653 5212 21 4442 19 873 Daryle Lamonica 2601 5031 22 4363 20 657 Terry Bradshaw* 3901 4310 23 2885 45 1617 Jeff Garcia+ 3676 4241 24 4328 21 - 84 Joe Namath* 3762 4138 25 4023 24 102 Drew Brees+ 3650 3978 26 3872 29 114 Boomer Esiason 5205 3976 27 4242 23 - 204 Trent Green+ 3740 3950 28 3962 27 - 96 Warren Moon* 6823 3949 29 3848 30 61 Jim Hart 5076 3948 30 3965 26 - 22 John Hadl 4687 3911 31 3975 25 - 5 Bert Jones 2551 3825 32 3818 32 10 John Brodie 4491 3809 33 3821 31 - 34 Roman Gabriel 4498 3743 34 3755 33 - 13 Cecil Isbell 818 3691 35 3440 38 285 Steve McNair 4544 3685 36 3717 34 - 109 Ken Stabler 3793 3674 37 2624 55 1118 Bob Griese* 3429 3636 38 3384 40 273 Charlie Conerly 2833 3565 39 2891 44 865 Y.A. Tittle* 4395 3377 40 4762 17 -1640 Donovan McNabb+ 4303 3327 41 3451 37 - 303 Rich Gannon 4206 3140 42 3506 36 - 384 Arnie Herber* 1175 3062 43 2995 42 - 71 Joe Theismann 3602 3031 44 2679 50 376 Mark Rypien 2613 2973 45 2389 57 573 Jim Kelly* 4779 2913 46 3621 35 -1022 Bernie Kosar 3365 2800 47 2365 58 442 Vinny Testaverde 6701 2729 48 2695 48 98 Randall Cunningham 4289 2722 49 2692 49 25 Daunte Culpepper+ 3042 2707 50 2806 47 - 115 Mark Brunell+ 4594 2637 51 3018 41 - 450 Matt Hasselbeck+ 3347 2634 52 2330 59 307 Jim Everett 4923 2498 53 2944 43 - 460 Billy Kilmer 2937 2457 54 2640 54 - 219 Steve Grogan 3593 2436 55 2669 52 - 312 Jim McMahon 2573 2410 56 1821 76 504 Earl Morrall 2689 2379 57 2643 53 - 253 Tommy Thompson 1424 2233 58 2671 51 - 440 Doug Williams 2507 2151 59 1882 71 192 Johnny Lujack 808 2130 60 2173 62 - 51 Brad Johnson+ 4326 2129 61 2264 60 - 222 Frank Ryan 2133 2126 62 1954 68 70 Bob Waterfield* 1617 2096 63 2061 65 29 Jeff Hostetler 2338 2082 64 1484 88 625 Chad Pennington+ 2395 2060 65 2251 61 - 191 Phil Simms 4647 2059 66 1754 77 334 Bobby Layne* 3700 2049 67 2546 56 - 958 George Blanda* 4007 2047 68 2131 64 - 623 Craig Morton 3786 2032 69 2814 46 - 939 Jake Delhomme+ 2434 2029 70 1598 85 463 Carson Palmer+ 2165 2022 71 1961 66 60 Milt Plum 2419 1976 72 2156 63 - 257 Tony Romo+ 1307 1974 73 1959 67 15 Philip Rivers+ 1428 1922 74 1849 72 76 Ben Roethlisberger+ 1905 1907 75 1416 93 546 Greg Landry 2300 1865 76 1932 70 - 78 Don Meredith 2308 1849 77 1827 74 - 7 Tom Flores 1715 1830 78 1830 73 0 Tobin Rote 2907 1814 79 761 124 1116 Ed Danowski 605 1800 80 1663 82 59 Bernie Masterson 409 1748 81 1653 83 99 Charley Johnson 3392 1707 82 1707 79 0 Joe Ferguson 4519 1691 83 1827 75 - 143 Doug Flutie 2151 1671 84 1666 81 - 35 Brian Sipe 3439 1662 85 1937 69 - 275 Neil Lomax 3153 1607 86 1614 84 - 10 Jeff George 3967 1599 87 1332 97 284 Marc Bulger+ 2924 1551 88 1440 91 108 Wade Wilson 2428 1547 89 1380 94 154 Steve DeBerg 5024 1525 90 1528 86 - 8 Bill Kenney 2430 1475 91 1469 90 8 Chris Chandler 4005 1437 92 1363 95 73 Bill Nelsen 1905 1365 93 1681 80 - 386 Danny White 2950 1317 94 1502 87 - 211 Jim Plunkett 3701 1309 95 363 160 1105 Norm Snead 4353 1309 96 1309 98 0 Dave Krieg 5311 1306 97 1479 89 - 213 Ron Jaworski 4117 1284 98 1710 78 - 405 Babe Parilli 3330 1240 99 1226 100 25 Bobby Thomason 1346 1238 100 1238 99 0
I'm sure there will be lots of debates about this list in the comments. Note that this is just a ranking of quarterback statistics, and not a pure ranking of the best quarterbacks ever. Obviously a QB's statistics are affected by lots of things, but what this formula tries to do is combine all the main statistics a QB produces, adjust for era, and accumulate them in a way that puts the longevity guys and the high peak guys on an equal plane. As I said yesterday:
It’s important to remember that this is just a measure of each team’s passing game, assigned to the quarterback on the field for those plays. Obviously the quality of the offensive line, the ability of the receivers, the versatility of the tight ends and running backs, the philosophy of the coaches, the strength of the schedule, and good old randomness have a significant impact on the above numbers. The reason for these posts is to accurately measure quarterback statistics, and nothing else. Once we have strong measures of QB performance, we can then judge QBs based on how much of their success (or lack thereof) we want to assign to the QB and how much to other people/factors.
No, I don't think Ken Anderson was a better QB than Johnny Unitas; the fact that he ranks a hair above him doesn't mean the formula is junk, although it might mean that I'm still slightly overvaluing modern QBs (Unitas kills Anderson in the "SEARK" column). On the other hand, Sid Luckman comes in as a top-five QB of all-time on my list. Unitas might come in "only" at #8 because was unfortunate enough to play in a small league with HOF QBs like Van Brocklin, Layne, Jurgensen, Starr, Tittle and Tarkenton, along with some borderline guys like Brodie and Conerly, skewing the league average. Regardless, Anderson is the best quarterback eligible but not in the Hall of Fame, however.
Someone like Jim Kelly gets really hurt in these rankings because of his awful post-season performances in big games. I prefer to use replacement value (and not general value) for a QB like Kelly, as he is better than the 45th best QB in NFL history. He is one of the weakest QBs in the HOF, however.
I'll close with one last list. I looked at all QBs on a year-to-year and five-year basis. Who was the best QB in 1974? Who was the best QB from 1956 to 1960? The table below shows the yearly leader in converted yards over average (including post-season) and the five-year leader. Obviously for the past few seasons, the five-year leader is not a "true" five-year leader.
5-year 5-year top QB Year N Best QB 2008-12 Kurt Warner (1427) Kurt Warner (1427) 2007-11 Peyton Manning (2230) Tom Brady (2124) 2006-10 Peyton Manning (3900) Peyton Manning (1670) 2005-09 Peyton Manning (5377) Peyton Manning (1477) 2004-08 Peyton Manning (7933) Peyton Manning (2556) 2003-07 Peyton Manning (8620) Peyton Manning (1664) 2002-06 Peyton Manning (7828) Trent Green (1158) 2001-05 Peyton Manning (6757) Kurt Warner (1572) 2000-04 Peyton Manning (6637) Jeff Garcia (1591) 1999-03 Peyton Manning (5287) Kurt Warner (2190) 1998-02 Kurt Warner (4268) Randall Cunningham (1570) 1997-01 Brett Favre (2518) Brett Favre (1078) 1996-00 Brett Favre (3323) Brett Favre (1608) 1995-99 Brett Favre (4790) Brett Favre (1510) 1994-98 Steve Young (6233) Steve Young (2697) 1993-97 Steve Young (6552) Steve Young (1591) 1992-96 Steve Young (7473) Troy Aikman (2113) 1991-95 Steve Young (7858) Mark Rypien (2003) 1990-94 Steve Young (7539) Jim Kelly (1537) 1989-93 Steve Young (5274) Joe Montana (2473) 1988-92 Joe Montana (5645) Joe Montana (1544) 1987-91 Joe Montana (6004) Bernie Kosar (1409) 1986-90 Joe Montana (6192) Dan Marino (1470) 1985-89 Joe Montana (6250) Jim McMahon (1299) 1984-88 Dan Marino (7032) Dan Marino (2923) 1983-87 Dan Marino (6826) Joe Theismann (1449) 1982-86 Joe Montana (5839) Dan Fouts (1738) 1981-85 Joe Montana (6696) Ken Anderson (2031) 1980-84 Joe Montana (5880) Dan Fouts (1195) 1979-83 Dan Fouts (5914) Roger Staubach (1204) 1978-82 Dan Fouts (5812) Roger Staubach (1311) 1977-81 Dan Fouts (4158) Roger Staubach (1392) 1976-80 Roger Staubach (4570) Ken Stabler (1653) 1975-79 Roger Staubach (5696) Ken Anderson (1505) 1974-78 Roger Staubach (4871) Ken Stabler (1370) 1973-77 Ken Anderson (4521) Roman Gabriel (1212) 1972-76 Fran Tarkenton (4325) Joe Namath (828) 1971-75 Ken Anderson (3668) Roger Staubach (1394) 1970-74 Fran Tarkenton (3246) John Brodie (1627) 1969-73 Fran Tarkenton (3453) Daryle Lamonica (1149) 1968-72 Daryle Lamonica (4263) Daryle Lamonica (1705) 1967-71 Daryle Lamonica (5308) Daryle Lamonica (1355) 1966-70 Daryle Lamonica (5090) Bart Starr (1926) 1965-69 Len Dawson (4469) Fran Tarkenton (757) 1964-68 Bart Starr (4882) Johnny Unitas (1153) 1963-67 Bart Starr (4510) Tobin Rote (1281) 1962-66 Len Dawson (4927) Len Dawson (1462) 1961-65 Len Dawson (3059) Sonny Jurgensen (1067) 1960-64 Sonny Jurgensen (2624) Milt Plum (1545) 1959-63 Johnny Unitas (2841) Johnny Unitas (1677) 1958-62 Johnny Unitas (3289) Johnny Unitas (1245) 1957-61 Johnny Unitas (4092) Johnny Unitas (767) 1956-60 Johnny Unitas (4719) Charlie Conerly (826) 1955-59 Charlie Conerly (2424) Otto Graham (1161) 1954-58 Otto Graham (2127) Otto Graham (966) 1953-57 Otto Graham (3591) Otto Graham (1463) 1952-56 Otto Graham (4456) Otto Graham (865) 1951-55 Otto Graham (5137) Norm Van Brocklin (1104) 1950-54 Otto Graham (5091) Otto Graham (1115) 1949-53 Norm Van Brocklin (4142) Johnny Lujack (1147) 1948-52 Johnny Lujack (2223) Charlie Conerly (1035) 1947-51 Sammy Baugh (2522) Sammy Baugh (1278) 1946-50 Sammy Baugh (2850) Sid Luckman (996) 1945-49 Sammy Baugh (4104) Sammy Baugh (1363) 1944-48 Sammy Baugh (3473) Frank Filchock (821) 1943-47 Sid Luckman (6385) Sid Luckman (3331) 1942-46 Sid Luckman (5949) Cecil Isbell (1523) 1941-45 Sid Luckman (6379) Sid Luckman (1426) 1940-44 Sid Luckman (5885) Sammy Baugh (643) 1939-43 Sid Luckman (5915) Frank Filchock (906) 1938-42 Cecil Isbell (3941) Ed Danowski (677) 1937-41 Sammy Baugh (1705) Sammy Baugh (1007) 1936-40 Arnie Herber (1984) Arnie Herber (1109) 1935-39 Arnie Herber (2223) Ed Danowski (1050) 1934-38 Arnie Herber (2620) Arnie Herber (568) 1933-37 Arnie Herber (2372) Harry Newman (1155) 1932-36 Arnie Herber (2533) Walt Holmer (458)
Looking at the five year rankings is like watching the evolution of the quarterback position. If we require three straight years of ranking as the top QB over a five-year stretch, you get a who’s who in QB history. We start with Arnie Herber (I was liberal with my definition of a QB) and the Packers, then Sid Luckman and the Bears, Sammy Baugh and the Redskins, Otto Graham and the Browns, and then Unitas and the Colts. Things stalled for a bit, then Lamonica and the Raiders take over, followed almost by Tarkenton and the Vikings but instead by Roger Staubach and America’s Team. He passed the crown to Dan Fouts and the Chargers, who was followed by consecutive 49ers QBs Joe Montana and Steve Young. Then Brett Favre took over, followed by the long reign of King Peyton Manning. He may end up having ten straight years of leading my five-year rankings by the time it’s all over.
This entry was posted on Tuesday, August 11th, 2009 at 7:01 am and is filed under Best/Worst Ever. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.
AFL QBs, and especially early AFL QBs, are overvalued slightly in this formula. We're all enjoying Jason's AFL-NFL series, and when that is over, I'll have an AFL modifier to downgrade some of those QBs. Here are the league-wide numbers (excluding sack data) over the '60s:
So by comparing AFL QBs to only AFL QBs, and not using a modifier, they're slightly overvalued (to whatever extent the modifier should be).
Two points--and bt=y the way I am enjoying the analysis--1) Should this measure be seen as an indicator of "Best QB careers" as opposed to "BesT QB's or "Best Team Passing Performances?" This would mean that someone could be "better" (Unitas vs. Anderson) but that the latter had a stronger overall career by the numbers; 2) I think it is a problem not to include the AAFC data, despite what the NFL does. The AAFC was a better league in comparison to the NFL than was the earlier AFL, based at least on the better teams in the AAFC (obviously Cleveland). While this does involve some work, I think it is more valid to include those #'s.
Hey Tim,
Glad you're enjoying the analysis.
I'm not sure if there's a meaningful difference between best QB and best QB career (although I understand what you're saying). Best team passing performances is certainly the most accurate, but we know that isn't very catchy.
After Jason is done with the AFL, we'll get him right on the AAFC. I'm sympathetic to your argument, though.
All very well and good, but what I'm waiting for is a re-examination of the worst QBs of all time
How far "up" the charts did Trent Dilfer's 2007 move him? Has Jon Kitna moved into the "top" 10? I have to know!
For you, Jason.
Rick Mirer, Ryan Leaf, David Carr, Joey Harrington, Mike Phipps (post-season included!), Trent Dilfer (uh, still post-season included), Mike Taliaferro, Gary Huff, Akili Smith and Kim McQuilken round out the bottom ten (from worst to best).
For those new here, this is just most career yards below average. So it's really the list of QBs who were the farthest below average for the longest (the best QBs ever are the ones who were the farthest above average for the longest), and not really the worst QBs ever. Except for Leaf and McQuilken, of course.
I'm not sure why you prorate just half the games. Why not use 3/4 of the 16 games-how many games they played. For a 9 game season, 12.5 games seem a little on the low side. but for 3/4 proration you get 14.25 games which seems a lot better than 12.5
I still think there needs to be some type of correction for weather; especially in the playoffs portion of this. Guys like Favre and Brady obviously get hurt relative to guys like Manning and Montana. You simply can't put up the same statistics in a blizzard as you can indoors.
I think a bonus for winning would also be interesting, since it is so directly associated with the best QB discussion. Maybe add 100 yards for every win after #8?
pm,
It's a tough balance to strike. Remember that we're not pro-rating raw numbers like touchdowns. In that case, a 3/4 pro-rating or even a straight pro-rating (i.e., using per-game stats) is reasonable.
But we're pro-rating value above average. I suspect if we looked at all QBs who were say, 500 yards over average after 9 games, they probably would end up closer to 694 (using 12.5) than 791 (14.25) or 888 *using a standard pro-rating method). But that's an empirical question, and one we can answer. Originally, the "split the baby" method was a kludge to avoid giving too much credit to having an awesome 9 games or whatever. But you're on the right track that we can be more precise there.
Andrew,
Check back tomorrow please.
I forgot about Sid Luckman
2200 yards and 28 passing TDs in 10 games in the pass-unfriendly 40's, that's pretty hard core.
All I know is that Drew Bledsoe threw for the seventh most yards ALL-TIME!
How can he not be among the 100 greatest QB's ever? This list is a total joke. Bledsoe's career numbers are better than Fouts yet Fouts comes in at number six?
The HOF may recognize AAFC for induction reasons, but they don't recognize the statistics, and I agree fully with it. Look at the drop Otto Graham had from the AAFC to the NFL. Frankie Albert is another good example. He had some amazing numbers in 1948-49 (TD%'s of 10.4 and 11%), but he only lasted 3 years in the NFL and really wasn't too good.
Glenn Dobbs led the AAFC in attempts and completions for two of it's four seasons, but no NFL team ever picked him up.
HOF TB Ace Parker played his last year in the AAFC and had arguably his best season throwing the ball.
I'm pretty sure Tittle and Graham are the only QBs that were able to have NFL success as well, but I still wouldn't count their AAFC stats.
And I have to say Unitas coming behind Fouts and Anderson would worry me. He might have been pretty bad the last 6 years of his career, but Unitas the was the man for his first 12 seasons (except 1961). Hardly any QB can make claim to having that many good seasons. He led the league in major passing categories several times and did win 3 MVPs. He also has the most 1st team All Pros.
Excluding regular season stats, here are the year-by-year grades for those QBs. The categories should be self-explanatory, just note that the ValueP category is the score I'm using (Value is the regular value, and ValueP is the pro-rated version):
Unitas:
Anderson:
Fouts:
You obviously need to do the 100/95/90 thing, but that should explain their career regular season ratings. I'd also say that of Unitas' first 12 seasons, he was also bad in '61 and '62 (below average), not particularly impressive in '66, and just okay as a rookie in '56. Obviously Unitas was the man, but just trying to explain his ranking here. Feel free to investigate the issue further, here, and maybe you'll find some flaw in the formula that explains Unitas ranking below Fouts and Anderson.
I definitely see some things that I'd like you to clarify for me.
First, the RK is where that QB ranks for that season right? And in the 60's, you are combining AFL and NFL (but keeping the stats separate for Value) right?
I thought you were only using all QBs with 100 attempts for each year, but it seems like you're not (hence Unitas being ranked 66th in a season he threw 32 passes). I think you can create a lot of trouble by using everyone that threw a pass. VALUE is basically a rate statistic, so like all rate statistics, a minimum number of attempts to qualify should be added. You can go with 100 to make it easier for all seasons, or you can go with what the NFL used.
1985-present: 14 attempts/game
1960-1984: 10 attempts/game
1950-59: 100 attempts
1940-49: 6 attempts/game
1932-39: 5 attempts/game
As for the seasons that don't qualify, well, I wouldn't want to just throw them away completely. Maybe take all the non-qualifying seasons in a year, calculate the stats on those based on those averages, then give small weights to those when calculating career numbers.
What jumps out to me is Unitas ranking 36th in 1962. Now I agreed with you that 1961 wasn't too hot a year for Unitas, but I can't imagine how a season like 1962 can be so low statistically.
In the NFL in 1962, Unitas ranked 2nd in attempts, 2nd in completions, 5th in comp. %, 4th in yards, 11th in YPA, 2nd in TD passes, 6th in TD%, 4th most INTs, 9th in INT% and 7th in passer rating. He only fumbled 5 times and I don't believe you're using sacks for 1962. 16 QBs qualified for rate statistics, so his rating puts him above average. In the AFL that year, Len Dawson had an amazing season and Babe Parilli was really good as well. But the 3rd rated passer was Frank Tripucka, and he wasn't better than Unitas. Not to mention you only had 8 passers that qualified for rate stats. There were 7 other QBs to try 75-139 attempts that year, and those too were clearly inferior seasons to Unitas.
So basically, I have no clue how Unitas can come out 36th in a (Pro Bowl) season where only 2 AFL QBs were better, and maybe at most (being generous here) 12 NFL QBs were better. I'm stumped.
If it's not too much trouble (or if it's easier, e-mail me it), I'd love to see the 35 QBs ranked ahead of Unitas in 1962. If John David Crow and his back-up Lamar McHan are ahead of him (QBs with 20 attempts that year), then I think that's a fatal flaw in this methodology.
(Sorry if that came out rude, I am trying to help)
Hey Scott... I posted the numbers for discusssion, so I'm glad you're investigating for me. You are correct on what the RK is, although I'll discuss it more at the end.
I only listed QBs with 100 attempts in '08 because of space concerns. VALUE is *not* a rate statistic; there's no need for a minimum number of attempts, because QBs can't deviate far from average with a small number of attempts. VALUE simply measures how far form average a QB truly is. Now, I agree that comparing QBs with a small number of attempts to a large number of attempts feels weird -- that's why I like using REPL a lot of the time, which limits what you can do on a bunch of attempts. But if a guy stinks for 32 attempts, that should count. Also, remember that since that was just Unitas' 16th best season, only 25% of his value for that season counts. So his -297 gets converted into just -74.
As for 1962, it's very close, but Unitas ranks a hair below average. He averaged 7.6 Y/A, the league was at 7.8. He averaged 5.9 PTD/ATT, the league 5.6, but he averaged (of course) 5.9 TTD/ATT, the league 6.1. He averaged 5.9 INT/ATT, the league 6.0. So in this case, it looks like Unitas was ever so slightly above average normally, but by including rushing and passing TDs together, he was ever so slightly below average. If you're a Unitas fan (how can you not be?), this might seem like a reason to throw out rushing TDs. But honestly, there would be almost no difference to Unitas' career score if he was slightly above average in '62 as opposed to slightly below average. (Moving him from -36 in VALUEP to +36 in VALUEP obviously wouldn't do much.) That's good, since those things obviously aren't very different.
But what about that 36 rank? That's where the below average thing comes. Every single QB that didn't throw a pass would necessarily rank above Unitas once we call him below average. Every QB with 1 or 2 or 3 passes, too. Now it's silly to say Unitas was the 36th best QB that season, because he wasn't. If I was grading Unitas on the season, I'd use replacement value, and Unitas ranks 11th using replacement. That's much more in line with what you were saying.
Remember, that "36" just means he was the 36th best as producing yards over average. It doesn't mean anything in any real context. The "SEARK" column ignores everyone who finishes out of the top 10, so there's no difference between 11 and 36. And the VALUE column simply isn't changed all that much by dropping from slightly above average to slightly below average -- it's changed exactly the same by dropping from slightly above unbelievable to slightly below unbelievable, if that makes sense. Now the "REPL" score changes, but there still, his "36" in VALUE is irrelevant to that. That's also why Fouts was the 100th best QB in '87; no, he wasn't the 100th best QB -- he was just the 100th best QB at producing value over average. Every QB who didn't take a snap was better at that than him.
Anyway, sorry for the length of the post, but I am doing my best to explain a nuanced topic. I hope that helps.
I feel if it's something you can't count to come up with, then it's a rate statistic. You do have a formula with a denominator using total attempts to help calculate it, so I think it is a rate stat.
"because QBs can’t deviate far from average with a small number of attempts"
I don't know about that. I showed you last time that Roethlisberger lost a good 460 value points by what he did in 10 quarters of play (87 passes) last year.
If Unitas is ranking 11th in replacement in 1962, then I think I like replacement better. Looking at the overall list again strengthens my liking of replacement (even though a few issues come up there like Otto Graham's big fall).
"Every single QB that didn’t throw a pass would necessarily rank above Unitas once we call him below average. Every QB with 1 or 2 or 3 passes, too"
I don't know, something still feels unsettling. Guess I'll just have to wait for the next version you come up with.
Thanks Chase (#3) for the response. After reading all the other posts, especially about Unitas' relative standing, I still think the concept of "Best QB career" applies. Also, thanks to Scott (# 11) for clarifying the AAFC stat issue related to being included with NFL statistics. Certainly there was a big dropoff for many players when they went to the NFL. However, my thoughts on including them were relative to the AFL statisitics being included, especially from the early years. Case in point, Frank Tripucka, 8 years removed from the NFL, throws 24 TD's in his first AFL season with Denver. His career high in the NFL was 9. I think the argument about AAFC #s can be thought of in reverse for the AFL. That is, there were many players who performed in the NFL who then had significantlyy higher performances when in the AFL.
Oh and just another suggestion related to minimum attempts (I'm big on that), if you're going to do a "Greatest QB of All-Time" series, you might want to bump up the minimum attempts to something like 1000. This gives you 195 QBs to look at (after AAFC stats are left out). That's how I started when I recently did a non-statistical top 50 all-time. If you're going to be among the best, we need to see you play a fair amount. 1000 attempts, that's 10 seasons averaging 100, 5 seasons averaging 200, 2 seasons averaging 500, etc. Not a really big requirement and I don't think you really leave out anyone great. Actually I think I had Benny Friedman, Greg Cook, Matt Ryan, Aaron Rodgers and Ace Parker as the best QBs under 1000 attempts. And unless they have a Cook-like injury, Ryan and Rodgers will get (well) over 1000 attempts.
I just noticed you had Bernie Masterson and Ed Danowski in the list and that's why I brought it up. And I considered a guy like Cecil Isbell, but when you think about World War II and the Don Hutson factor, I don't think it's much of a loss to exclude him as well.
Scott,
It's not a counting statistic or a rate statistic. It's a blend of the two. Say a league average of 7.0 Y/A. This puts a guy with 500 passes at 9.0 Y/A at +1000. It also puts a guy with 400 passes at 9.5 Y/A at +1000. It starts out as a rate statistic (with attempts in the denominator) but then the result is multiplied by your number of attempts. So the attempts part cancels out, and it stops being a rate statistic.
I like replacement better for Unitas' 1962 season, as well. Generally, I like replacement for QBs outside of the top ten in a season and QBs outside of the top thirty in a career. However, for the best few QBs in a season or a career, I think VALUE is much better. Eli Manning is a great guy to illustrate the point.
In '05, he ranked 18th overall and 11th in replacement. In '06, he ranked 67th in value and 20th in replacement. In '07, he ranked 83rd overall and 26th in replacement. Last year he was 16th overall and 15th in replacement. Using replacement gives us a much better view of Eli's career, although he still is mediocre using replacement (but using Value would say he's one of the worst QBs ever).
I understand why you think things feel unsettling. Honestly, I'd just ignore the VALUE ranks when they're lower (larger) than 10. GQBOAT-IV won't change this issue, because comparing QBs to league average is the central part of the formula. Replacement is good, but it's not good for the best QBs. Replacement says Favre>>Young, Esiason>>Warner, and things like that. It rewards compilers and kills the guys with short but great careers. That's fine when we're talking about the 40th best QB ever (we care about compilers then) but not the 15th best.
Replacement also really harms older QBs (on a career list). Replacement just gives guys with more attempts more value, and obviously attempts are much higher now than they were for most of history.
As for Drew Bledsoe, he ranks 126th using value and 41st using replacement. I think that #41 is too high, but it's only because the older QBs don't rank well in replacement (Otto Graham is #42).
Enjoying the discussion, Scott.
Re: Johnny Lujack
Funny how he doesn't get any All-Pro recognition in 1949 when he led the league in attempts, completions, yards, TDs, YPA while rushing for 2 more TDs and Chicago had a 9-3 record.
But a year later, he throws 4 TDs, 21 picks, a 41.0 passer rating (ranked 16th out of 18), the same 9-3 record, and he's suddenly All-Pro/Pro Bowl. Of course he did have 397 rushing yards and 11 rushing TDs (a league-high). And this is in the days of the TB and passing wasn't accepted yet as it would be after Unitas showed people how to play the position right. But that's still a pretty funny example of how things were valued 60 years ago compared to today.
"Generally, I like replacement for QBs outside of the top ten in a season and QBs outside of the top thirty in a career."
I actually think the top 30 for replacement jives better with conventional wisdom better than the top 30 value list.
"Using replacement gives us a much better view of Eli’s career, although he still is mediocre using replacement (but using Value would say he’s one of the worst QBs ever)."
I understand what you're trying to do with value, I really do, but I think when the results include QBs with 1 pass attempt being "better" at making value over average of a QB with 500 attempts, it all looks a bit silly. I would just make different lists for QBs with >100 attempts and everyone else.
And when you compute your league average numbers, are you taking out passes from RB's, WR's, P's, etc.?
Scott,
I don't like to arbitrarily include a minimum, so that's why I didn't want different lists. And really, there's no significant difference (except in the ranks, which I don't think is important) between a QB with 500 attempts being +40 or -40. That won't change his career rating much.
I think replacement simply rewards the compilers too much, but I can see how people would disagree. I'm not sure how much faith I put into conventional wisdom
. I'm comfortable with not giving a QB credit if he was not above league average in a season when I'm trying to decide the best QBs ever.
Yes, I exclude passes from all non-QBs, except in the early days of NFL history where it's very tough to separate out the QBs and RBs and TBs. There I just took my best guess, and tried to be more inclusive of "QBs" as it was really a totally different position back then.
I was coached one summer by Benny Friedman. No mention of him. Is it that your calculations start after his retirement...or didn't he make the cut?
R. Rosenthal,
Friedman played most of his career in the days before anything besides TDs was recorded. Much of what I've read indicates that he was every bit the equal (if not superior) of Luckman and Baugh. He certainly would make the cut if we had his numbers. Rest assured, you were coached by an all-time great.
Tony Romo?!?! You've got to be kidding. How does he come in at 73rd on a list of best post-season QB's of all time if he has never won a playoff game? The Cowboys haven't even won a playoff game since 1996. Romo can't even handle a hold on a field goal attempt during a playoff game, his name should not even be considered for this list.
Tom Brady is #19 before you add the playoff #'s and Aikman is 39th. Aikman has 16 games (11-5) has a QB rating of 88.3 (23 TD's and 17 Ints) and Brady plays in 17 games (14-3) has a QB rating of 88 (26 TD's and 12 INT's). Those numbers boost Aikman up to #20 overall (19 spot elevation), but drop Brady back to 21 (2 spot demotion)? And is there a reason Ken Anderson is so high on either of these lists?
Because Ken Anderson is awesome.
@ChaseTwo major issues that I see.
#1: Negative value. Unitas loses 12 or 13 hundred points to negative value 'seasons', Fouts loses less than 400.
#2: Big attempts bias in favor of modern QBs. Unitas was better in 1959 than Fouts was in 1981 on a per pass basis (CY/A - LGavg), and led the league in attempts. And Fouts blows him away in 'value' there. Apparently Unitas led the league in attempts, but wasn't valuable because he didn't throw the ball very much.
Off this specific topic, I seem to recall investigations into weather adjustments being made recently, and a conclusion being that domes add one Y/A. Would that adjustment hurt Peyton here as badly as it looks like it would? 6000 attempts, an extra Y/A on half of them would seem a signficant adjustment relative to others near the top. Just wondering.