Comments on: Jim Hart, Bernie Kosar, and Steep Declines http://www.footballperspective.com/jim-hart-bernie-kosar-and-steep-declines/ NFL History and Stats Sat, 30 Apr 2016 23:31:00 +0000 hourly 1 https://wordpress.org/?v=6.9.4 By: Steve O'Dell http://www.footballperspective.com/jim-hart-bernie-kosar-and-steep-declines/#comment-315705 Sat, 30 Apr 2016 23:31:00 +0000 http://www.footballperspective.com/?p=28857#comment-315705 Assumption here, of course, is that a QB’s performance and the team’s win percentage are one and the same. Gotta be a way to use DVOA or Elo to extract team performance from the QB’s?

]]> By: Chase http://www.footballperspective.com/jim-hart-bernie-kosar-and-steep-declines/#comment-315189 Tue, 05 Apr 2016 17:54:00 +0000 http://www.footballperspective.com/?p=28857#comment-315189 I wonder if there is a better way to present this data. I think this works fine for this sample, but what if I want to just measure big declines among all QBs (even QBs who finished above .500)? I think maybe adding 30 or 40 games of 0.500 ball to a QB’s play is the way to go, but not sure how to pick the right number.

By doing it this way, Bulger really stands out (of course, he’d stand out if you just used raw winning percentages, anyway). Add 30 games of .500 ball to his early games, and he still has a 0.623 winning percentage. Add 30 games of .500 ball to his late games, and his wining percentage is still only 0.326. That’s a decrease of nearly 30%, which is remarkable given that we have added a 15-15 stretch to both marks. That beats out Hill (who drops by 20.2%) and Kosar (21.2%) by decent margins.

I didn’t use winning percentage differential because of the small sample sizes (Culpepper, anyone?) but something like this makes sense to do as part of a follow-up post. I think. The issue is the number of games of .500 ball will be pretty arbitrarily picked.

]]>