Re: My Article on Projecting Pass Rushers in the NFL Draft
It seems that the bulk the the previous two drafts have been bucking your model, any hypothesis as to why this may be?
It actually performed quite well for this year's draft. It said Von Miller would be awesome, Von Miller is awesome. It said that Ryan Kerrigan would be quite good and he has been (as is the case with Clayborn who wasn't included in the update). It missed a bit on Aldon Smith, obviously, but very few people liked Smith better than Quinn pre-draft, and SackSEER did. It wasn't perfect, but for the 2011 draftees, overall, it's about as accurate as it should be this early in their careers.
Your point about the 2010 Draft, however, is well made. It just had a really, really bad year. The short answer is that these things are never 100% accurate and you'll have some misses, and it just so happened that more of them happened in 2010 than in any other year.
That said, I have been working on some major improvements in the meantime. The new model will incorporate all 260 edge rushers drafted in any round from 1998 to now and there are a few other tweaks, as well. I really feel that the changes will stabilize the model and it will perform much better over the next couple of years than it has so far. The irony, however, is that none of the improvements would have helped me much on Jason Pierre-Paul. It's hard to emphasize enough how many historical trends he bucked to be as good as he is. Out of those 260 guys, nobody else has come out in the draft with that little experience playing NCAA football, with such little production in terms of sacks, and with such a poor combine (yes, I know that he's hyped as an athletic freak, but his combine performance was poor for such a highly regarded prospect), and has had half of his success. It's really odd. Kudos to those who liked him by looking at the tape.