UPDATE 3/1: Apologies to KC Phenom, who posted a much better analysis on this same topic a few hours before I posted this.

Before I start, I have to admit I got the idea for this post from Wait and See's excellent post about where today's starting QB's were drafted. And in that post, KSU-Chief commented, "The one thing that would make it better is the fail rate in each round..." Which is my long-winded way of saying that this post contains almost no original thinking. So with that rousing introduction...

I went to DraftHistory.com to pull the list of all quarterbacks who were drafted from 1990-2009. Why 1990-2009? I wanted a large sample size, so 20 years seemed like a good swag. And I wanted to exclude quarterbacks whom it would be too early to evaluate, so I excluded the last two drafts.

That left me with 157 quarterbacks to consider. My next question was to figure out how to categorize each quarterback as a success or not. There are lots of ways to do this (points for stating the obvious?), but in the interest of expediency, I decided anyone who had a QB rating of at least 78.0 and at least 1500 pass attempts (the minimum to qualify for the NFL career QB rating records) counted as a success. And yes, QB rating is a flawed metric, and 78.0 was pretty arbitrary - but I figured it would work for the purposes of this rough analysis.

Those criteria gave me the following breakdown by round:

ROUND    QBs DRAFTED    QB SUCCESSES    SUCCESS RATE
  1          34              17             50.0%
  2          14               1              7.1%
  3          16               2             12.5%
  4          15               3             20.0%
  5          22               0              0.0%
  6          29               3             10.3%
  7          27               1              3.7%
 TOT        157              27             17.2%

Some observations:

• Half of 1st round QB's are successful, at least as defined by the irrefutable Rule of 78.0.
  
• After the 1st round, the success rate plummets to 8.1% (10 successes out of 123 draft picks) - and there seems to be little correlation between round and success rate.
• The 5th round is a cold, cruel place for quarterback dreams.
• During 1990-2009, it seems like teams preferred drafting a QB in the later rounds if they didn't take one in the first round.
• The seeming outsized success of 4th round picks should be taken with a grain of salt. The 3 "successful" QB's are Aaron Brooks (78.5), Kyle Orton (79.4), and David Garrard (85.8). After seeing that, I almost changed my criteria to 79.5, but it's late, and I'm feeling lazy.
  

My main takeaway is that this supports what pretty much everyone else has been saying. If you need to draft a quarterback, taking one in the first round is the way to go.