There’s been a lot of talk about how hard it is to predict when a college QB will be a stud in the pros. Football Outsiders put together a model to predict NFL success for college QBs. To summarize how it works, they start by noting that if you just use completion percentage and games played, you’re most of the way there. Those two stats alone are pretty good, but less so in recent years. Building on that, FO used recent data to make the model better at describing the last 15 years or so of QBs. Now, the model uses: games played, completion percentage (though now you get more of a penalty for dropping from 50% to 48% and less reward for going from 70% to 72%), change in QB rating from junior to senior year (which is really important, apparently, cause a big drop means you have flaws that people have figured out), BCS-level schedule, and run to pass ratio (to punish running too much and taking sacks). That’s the gist of the article, but it’s worth a read, and at the end, it shows the top five best and worst of all time, as well as predictions for last year’s crop of QBs (hint–it doesn’t look good).
Last year, Andy Dalton was number 1. After his sophomore year, Andrew Luck was about where Dalton was. Now he has a ton more games played, and his career looks basically like Drew Brees’, but with a better completion percentage. As such, I imagine he’ll be up there with dudes in the top five of all time (Phillip Rivers, Drew Brees, Carson Palmer, Peyton, and Chad Pennington, in that order).
Two takeaways here and one caveat:
1) This tells me that Andrew Luck is, in fact, better than most 1st picks. He’s more likely in Phillip Rivers territory than Drew Bledsoe/Vinny Testaverde territory, though you still never know.
2) As Nate said, “once in a generation” QBs actually come around every two years or so. Given that, the Colts should probably trade Luck for many picks and wait to get players with similar stats but with later picks, like Drew Brees or Andy Dalton. Even if Luck looks better than every other 1st pick, he’s not so much better that he’s worth multiple first round picks.
1) I can’t overstate how important this is: FO’s model is descriptive, not predictive. Anyone can look at 10 years of QBs and making a fancy model to predict the best ones–you can always look backwards and craft a story that “explains” everything. If you keep fiddling with the numbers, you’ll end up with a “perfect” model. So don’t salivate over seeing Luck with Brees/Manning/Rivers. If FO couldn’t come up with a model to show that, they wouldn’t have published it. That said, I think this still bodes well for Andrew Luck. I might do another post soon with the other QBs this draft.