Drafting vs. Recruiting Stars (As If There's a Difference)

Another quick thing on the draft before I get back to what I'm supposed to be doing...

One of the standing beliefs of SMQ is that, while far from perfect in the micro, recruiting rankings are very reliable indicators on which to base expectations. I've tried to show again and again that the guru predictions come out looking very solid overall, when you take the entire picture into account and not little slice like "Experts suck because Tom Brady." This fairly obvious point gets undermined on a pretty regular basis by our preternatural love of the underdog, who is not without his share of evidence.

Most of the guys under '0' come from I-AA schools and weren't rated out of high school/prep school/junior college at all; a couple others were walk-ons who weren't rated. The data on both groups was inconsistent (I-AA Domonique Rodgers-Cromartie was rated; I-AA Tim Hightower wasn't. Walk-on Jordy Nelson was rated; walk-on Owen Schmitt wasn't. Etc.). If you're wondering who the lone one-star guy was, as far as my accounting goes, it's Kansas receiver Marcus Henry, who went to the Jets in the sixth round.

As I pointed out in March, though, the raw numbers have to be taken in the context of the number of stars awarded to each category. Over the last five years, Rivals has consistently reserved its five-star designation for the top one percent of I-A recruits. Four-stars only make up a little over ten percent. A substantial majority (almost 60 percent) of the incoming rabble is rated two stars or lower. Their numbers on all-America teams and in the draft are naturally higher than the number of blue chips, but they are not even close to being proportional.

Taking the combined numbers from Rivals' 2003 and 2004 classes, I get 65 guys rated as five-stars, 509 four-stars, 1,512 three-stars and 3,206 two-stars or lower. Based on those numbers and the round-by-round breakdown above, the per capita odds of a player in each rating category hearing his name called last weekend broke down like this (kickers, punters and one long snapper excluded):

There are a couple problems here -- not every player drafted was a part of the 2003 and 2004 classes (I used 2005-06 JUCO ratings where applicable), and large segments of those classes have been and will be taken in other drafts. That's why it's phrased the way it is: 27.7 percent of players rated as five-star guys in 2003 or 2004 were drafted on this specific weekend. Over the course of several drafts, the percentage will be much higher.

Semantics can't hide the trend: five-star recruits have roughly a one-in-four chance of eventually being drafted, twice as high as four-star recruits, who themselves are about three times more likely to get picked than three-star prospects, and so on. Five-star kids are substantially more likely to get picked at any point in the first five rounds.

It's easy to forget how high the failure rate is for everyone. Very few make it to the league, and many fewer stick. This is the case across the board. The only reason we remember Ryan Leaf and Blair Thomas, and not the thousands of also-rans who barely set foot on a pro field, is that they had high expectations to begin with. Most of the time, by the nature of the business, high expectations will be wrong. Low expectations are hardly ever wrong, though you'll never hear the end of it on the rare occasion they are. In either case, the conventional wisdom has a better chance than flipping a coin.