While the debate over whether Alex or Geno Smith ought to be the quarterback of the Kansas City Chiefs has been underway, I've been doing some thinking about some of the stats that I have seen being thrown around.
One of them is that in the last ten Super Bowls, 70% of the winning quarterbacks were drafted by their teams in the first round. Another is that 57% of Super Bowls were won by quarterbacks drafted in the first round - and 32% by QBs drafted #1 overall.
All three of those stats are true. Sorta. But I would like to take a closer look at them, and give you one more to consider.
In the last ten Super Bowls, 70% of the winning quarterbacks were drafted by their teams in the first round.
This one isn't quite true. Two of those Super Bowls were won by the New York Giants with Eli Manning under center. It is certainly true that Peyton's younger brother entered the league with the Giants, but the Giants didn't draft him. San Diego drafted Manning with the first pick, and promptly traded him for Philip Rivers, whom the Giants had drafted with the fourth pick. With this correction, the percentage should actually be 50% - not 70%.
(If you want to argue that this shouldn't make any difference, you have a point. But I maintain that in the context of the discussion about what the Chiefs should with the first overall pick, this is an important distinction. More about that in a little while.)
But in my mind, there is an additional flaw with this stat. If you're going to use it as a predictor of success in winning the Super Bowl - which is precisely what we're talking about - you shouldn't be counting the games. Instead, you should be counting the players. In other words, for the purposes of the stat, it shouldn't matter how many times a quarterback won a Super Bowl - only that they did win one.
That said, I do understand why some folks would want a stat that points to recent Super Bowl history. So how about this: Among the last ten quarterbacks to win a Super Bowl, 40% were drafted by their teams in the first round. This list would include Brad Johnson, Trent Dilfer and Kurt Warner because Manning, Ben Roethlisberger and Tom Brady have all won twice since 1999. While not as sexy as the original, flawed stat, it's still pretty compelling.
57% of all Super Bowls were won by quarterbacks drafted in the first round.
Again... true as far as it goes, but like the other three stats, it has the flaw of counting games instead of players. But applying this change doesn't alter the result: 57% of all Super Bowl winning quarterbacks (that's 17 of 30) were drafted in the first round.
32% of all Super Bowls were won by quarterbacks drafted with the first overall pick.
This time, if we change it as I have suggested, the result changes appreciably: 23% of all Super Bowl winning quarterbacks (7 of 30) were drafted with the first overall pick. Still, though, it's a pretty compelling stat.
Or is it?
Now we'll get to the meat of the point I want to make:
If you want to use Super Bowl history as a predictor of what might happen when you select a quarterback in the draft, you'd better be comparing apples to apples.
John Elway played his whole career in Denver, was selected with the first overall pick, and (eventually) won a Super Bowl.
But Denver didn't select Elway. The Baltimore Colts did.
Elway publicly threatened to continue playing professional baseball instead of playing for the then-lowly Colts, forcing the Colts to trade him to the Broncos - which had picked fourth in the draft. Essentially the Broncos and the Colts traded picks; Denver got Elway for guard Chris Hinton, whom they had expended their fourth pick to obtain... pretty much exactly how the Giants got their Super Bowl winner Eli Manning 21 years later.
Why should this matter in the argument over Geno and Alex Smith? Because I maintain that if you're talking about a 2-14 team that is picking #1 overall because they were the worst in the league, you'd better use the history that reflects the same situation.
A team that trades up to get the first overall pick - such as the Broncos and the Giants in the two examples I mentioned - has, by definition, a better team on which their new quarterback can play. I don't think anybody will disagree that a quarterback will have a better chance to succeed - that is, to win a Super Bowl - if he starts on a better team.
So in order to predict the likelihood that picking a QB with #1 will eventually result in a Super Bowl win for that player, you should be looking at quarterbacks who meet three tests: they were drafted first, played for the teams that drafted them until they won a Super Bowl, and were drafted by the league's worst team. This will be your best historical predictor for success.
Unfortunately for those who want the Chiefs to use their #1 pick on a quarterback, only three quarterbacks meet all three criteria: Troy Aikman, Terry Bradshaw and Peyton Manning. Pretty smart company, to be sure! Two of these guys are already in the Hall of Fame, and the third certainly will be.
But they represent just 10% of the quarterbacks who have won Super Bowls.
Furthermore, while Aikman won the Super Bowl in his fourth season and Bradshaw won in his sixth, Mainning didn't win his until his ninth season. So it isn't reasonable to expect your new quarterback to win a Super Bowl sooner than his fourth season - and more likely his sixth.
So if you want to use history to predict the likelihood of a Geno Smith win for the Chiefs in the Super Bowl, here's the stat you should be using:
10% of all Super Bowl winning quarterbacks won them for the team that drafted them, and was also the previous season's worst team. On average, it first happened in the player's sixth season.
23% of Super Bowl winning quarterbacks were drafted with the first overall pick.
Please understand... I don't bring these facts to the table in order to argue that the Chiefs shouldn't take Geno Smith (or another QB) with the first pick. I simply think that if you are going to argue that the Chiefs should go QB at #1 because that's the way for the team to get a Super Bowl winning quarterback, you should be aware of the true historical likelihood that it will happen.
I'm actually working on a detailed analysis of all quarterbacks selected in the modern draft, but it is an enormous amount of data to collect and analyze. It looks very interesting, but it likely won't be ready until the after the draft.
Until then, have fun arguing about this little tidbit!