Have black QBs been discriminated in the draft? It's a complicated question that's probably impossible to fully answer. Akili Smith was a bust. Randall Cunningham was a steal. Those are easy, but how do we look at the whole group? Here's one way.
I looked at every QB selected in the first 224 picks of the 30 drafts from 1970-1999 (224 was chosen because that's how many picks there are in a 7-round, 32 team draft), and I then gave each QB drafted the pick value associated with that draft pick (a sample pick value chart can be found here). So Tim Couch, Peyton Manning, Drew Bledsoe, Jeff George, Troy Aikman, Vinny Testaverde, John Elway, Steve Bartkowski, Jim Plunkett and Terry Bradshaw were all assigned 3,000 points, since that's what the number one pick is worth. Trent Green -- pick 222 of the 1993 draft was assigned 2.6 points, and every other QB fell somewhere in between. There were 289 quarterbacks that met these criteria.
Now we need to give each QB a grade for each season played, and then add all the value from each season to get a career total. In Part II, I created the following formula:
(Passing Yards [plus] Passing TDs *10 [plus] RYO4.0 [minus] Passing INTs * 45 [minus] Sack Yards lost) / (Pass attempts + Sacks)
That formula is essentially Adjusted Net Yards per Attempt, with a bonus for every rushing yard over four yards per carry by QBs. If you subtract the league average ANY/A from an individual QB's ANY/A (with the rushing bonus), you can see how much value a QB added per pass attempt (defined as pass attempts plus sacks) in a season. If you multiply that differential by total pass attempts, you can see how much total value a QB added in a season. And then if you add up each season's value, you can see how much value a QB added in a career. A less refined formula was the basis for the Best Passer of All Time post.
I'm going to make one small change this time, though. When deciding who the best QB ever is, quality reigns supreme. When deciding if a player was a good draft pick, quantity becomes more important. A QB that is league average for ten straight years gets a zero on the Best QB Ever list; that may or may not seem fair to you, but either way, we know he is far from the best QB ever. But a QB that starts for ten straight years and is perfectly average has certainly been a better player than a guy who never made an NFL roster. So to compensate for this, I'm going to subtract 0.75*LgAvg ANY/A instead of the entire league average net adjusted yards per attempt. By lowering the baseline, we can reward quantity a bit more. The 3/4 mark of the ANY/A for the 2007 would have been 3.83, which is where Rex Grossman and Brodie Croyle landed. I'm comfortable with not giving a QB credit if he wasn't any more productive than the 2007 version of Brodie Croyle.
The best season by this metric -- not that I necessarily think that this metric is the most precise to determine the best QB season ever -- is Dan Marino's 1984 season. He averaged 8.11 adjusted net yards per attempt, received no rushing bonus, and had 577 total pass attempts. The league average ANY/A was just 4.63 that season, and 3/4 of that is 3.47. Marino's 2677 value added number was the highest of any season from 1970-2007, with Brady ('07) and Manning ('04) coming in right behind him. A system that ranks Marino's '84 season, Brady's '07 season and Manning's '04 season as the top three seasons by a QB since the merger at least passes the sniff test for me.
Under this system, among players drafted in the regular NFL draft in 1970 or later, Marino tops the career value list. At the bottom? Kim McQuilken. McQuilken was so awful, he made me have to consider what to do with the large amount of negative value he accumulated. I decided (after spending about 30 seconds) that anyone whose career value was negative would be given a zero. Players that couldn't even play an NFL game don't deserve much credit, so I decided not to rank any of them ahead of McQuilken, Ryan Leaf, Gary Huff or any other player who had a career value in the negatives.
Okay, so now what? We can run a simple regression analysis to tell us how many points of career value we should expect each draft pick to be worth. The formula is
Expected Career Value = 827 + 1.63 * Draft Value
The first pick in the draft (3,000 draft value) is therefore expected to have about 4,887 more points in value than the last pick in the draft (2 draft value). We can also run a regression analysis to predict games played in the NFL, using draft value as our variable:
Expected games played = 37.66 + 0.036 * Draft Value
This tells us the first pick should play over 100 more games in the NFL than the last pick. That sounds about right to me, especially when you consider that outliers like Elvis Grbac (pick 219) and Trent Green (pick 222) - two of the five lowest drafted QBs in the study - each started over 100 games.
Okay, with that out of the way, let's get to the black QBs. There were 17 drafted between 1970 and 1999. The table below shows each QB's draft pick, draft value based on that pick (drv), college, year they were drafted, the expected career value of a player with that draft position, the actual value that QB accumulated, the expected number of games played from a QB with that draft position, and the actual number of games played by the QB:
pick drv college year expval actval expgp actgp Parnell Dickinson tam 183 18 missvalleyst 1976 857 0 38 8 Vince Evans chi 140 36 usc 1977 886 887 39 100 Doug Williams tam 17 950 gramblingst 1978 2376 4639 72 88 Reggie Collier dal 162 27 somississippi 1983 870 3 39 6 Randall Cunningham phi 37 530 unlv 1985 1691 7160 57 161 Rodney Peete det 141 36 usc 1989 885 2375 39 104 Andre Ware det 7 1500 houston 1990 3273 78 91 14 Jeff Blake nyj 166 25 eastcarolina 1992 868 4687 39 120 Steve McNair oti 3 2200 alcornst 1995 4415 9430 116 162 Kordell Stewart pit 60 300 colorado 1995 1316 2128 48 126 Tony Banks ram 42 480 michiganst 1996 1610 2021 55 97 Charlie Batch det 60 300 eastmichigan 1998 1316 1996 48 71 Donovan McNabb phi 2 2600 syracuse 1999 5067 7758 131 118 Akili Smith cin 3 2200 oregon 1999 4415 0 116 22 Daunte Culpepper min 11 1250 centralflorida 1999 2866 6851 82 92 Shaun King tam 50 400 tulane 1999 1479 484 52 34 Aaron Brooks gnb 131 41 virginia 1999 894 4182 39 93
For each player, I bolded the higher number between the expected and actual totals. For Vince Evans, I bolded both his actual and expected value, since they were nearly identical. I did the same for McNabb's actual and expected games played, because while he's played fewer than we'd expect so far, I'd wager he's a good bet to end up playing more games. I think by any reasonable standard (i.e., outside of Suepr Bowl victories), McNabb has been a better than average number two pick.
As you can tell, five of the 17 QBs underperformed their expected value, and eleven of them overperformed. A decent theory would be that Andre Ware set black QBs back a bit. He was a famous bust for the Lions, a waste of the first ever top ten pick spent on a black quarterback. After him, three black QBs had underperformed significantly (Dickinson, Collier and Ware), three overperformed significantly (Williams, Cunningham and Peete), and one performed dead on. That sounds like almost any other group of draft prospects.
But after Ware, the next six QBs overachieved. Did Jeff Blake fall so far in the draft because of Andre Ware? Did Kordell, or Tony Banks, or Charlie Batch? It's tough to say that McNair or McNabb fell, although McNabb was drafted behind Tim Couch. Regardless, all six black QBs outperformed their draft status.
From a games played standpoint, the numbers are similar. Five of the 17 QBs played fewer games than expected, 11 played more games than expected, and one is still up in the air. In the 1990s, black QBs on the whole were undervalued in the draft, whether due to conscious racism, unconscious race bias, the Andre Ware effect, small sample size, luck, the teams they were drafted on, or any combination of factors.
That 1999 draft featured five black QBs, three of whom went on to become very good. What's been the case since then? Have we seen a reverse-Ware effect? We could call it the McNabb-Couch theory, if black QBs have since been drafted too early and underperformed.
There have been fourteen QBs drafted in the top 224 picks since 2000.
name college year team pick drv expval actval expgp actgp Tee Martin Tennessee 2000 pit 163 26 870 0 39 3 Spergon Wynn SW. Texas St 2000 cle 183 18 857 0 38 10 Jarious Jackson Notre Dame 2000 den 214 6 836 0 38 5 Michael Vick Virginia Tech 2001 atl 1 3000 5720 3256 145 74 Quincy Carter Georgia 2001 dal 53 370 1430 737 51 38 David Garrard East Carolina 2002 jax 108 78 954 2026 40 40 Rohan Davey LSU 2002 nwe 117 60 925 16 40 7 Byron Leftwich Marshall 2003 jax 7 1500 3273 2150 91 50 Seneca Wallace Iowa State 2003 sea 110 74 948 172 40 24 Jason Campbell Auburn 2005 was 25 720 2001 856 63 20 Vince Young Texas 2006 oti 3 2200 4415 565 116 30 Tarvaris Jackson Alabama State 2006 min 64 270 1267 196 47 16 JaMarcus Russell LSU 2007 rai 1 3000 5720 0 145 4 Troy Smith Ohio State 2007 rav 174 22 863 143 38 7
The three QBs in the 2000 draft -- all very late round picks -- ended up flaming out. Who knows what to say of Vick's career, but it seems unlikely that he'll ever live up to expectations anymore. Quincy Carter was considered a reach when Jerry Jones grabbed him seven years ago, and he fell a little short of typical second round expectations.
David Garrard looks like a great pick, although we wouldn't have said that a year ago. Rohan Davey was a bust. Leftwich's career isn't over, but I don't think he'll ever reach the lofty expectations of a top ten pick. Seneca Wallace has played well in limited time: he might be the eventual replacement for Hasselbeck up in Seattle, and I think he'd succeed in that role.
Jason Campbell and Vince Young have played well in spurts, and both at least appear on their way to living up to their draft status. It's too early to make any calls of them, and it's obviously way too early to tell what will come of Jackson, Smith, and especially Russell.
I'd say the five black QBs from the '00 and '01 drafts were busts, although I wouldn't argue vehemently if you don't want to classify Vick as one. Garrard's a check in the outperformed box and Davey's a check in the underperformed one. Leftwich and Seneca Wallace seem like tough ones -- I think each QB has a decent chance to end up as an overachiever, especially Wallace in the games played section (assuming he ever starts).
I don't think there was a McNabb-Couch effect, because McNabb and Couch had similar career results by the end of the '00 season, and only the '00 and '01 drafts overrated the black QBs. But Steve McNair famously lead his team to the Super Bowl in January of 2000, which may be as good an explanation (outside of small sample size) as any for that short spike.
It's anyone's guess as to what will become of Campbell, Young, Jackson, Russell and Smith. Outside of Russell (who's barely played), the others have shown enough to make you think they might one day live up to their draft status. None of them are busts yet. To conclude, it doesn't appear as if race drove down the draft stock of those black QBs lucky enough to get drafted in the '70s and '80s. While there were talented players that were never given a chance to play, went to another football league, or were unfairly converted to another position, the ones that were drafted to play the position were not noticeably undervalued. Conversely, from '92 to '99, eight of the ten black QBs were underrated. After five teams went the other way in the '00 and '01 drafts, it seems -- as it should be -- that black QBs are now being rated based on nothing else but their pro potential.
This entry was posted on Friday, February 22nd, 2008 at 6:22 am and is filed under History, NFL Draft, Statgeekery. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.