tag:blogger.com,1999:blog-38600807.post3716283284682366999..comments2009-09-23T12:54:46.044-04:00Brian Burkenoreply@blogger.comBlogger8125tag:blogger.com,1999:blog-38600807.post-46449837685925114262009-09-23T12:54:46.044-04:002009-09-23T12:54:46.044-04:00Hey I just found this site. I was wondering how you found that the Burress TD catch had a +0.41 WPA. It looks to me that the Giants&#39; win probabiliy was 0.82 after the third down conversion to Steve Smith, and after the TD to Burress (on the very next play) it went to 0.92, making it a +0.10 WPA play. Perhaps I am misunderstanding the graph?<br /><br />Thanks for the site, this is great stuff.Davenoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-57305874769451897032009-06-04T11:13:03.361-04:002009-06-04T11:13:03.361-04:00I&#39;m just talking about QB&#39;s, RB&#39;s, TE&#39;s, and WR&#39;s. Those are the only players you have to do it for. You could also do it for team offenses and defenses.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-19659246555050705062009-06-03T22:27:39.595-04:002009-06-03T22:27:39.595-04:00Sure. But the complex team aspect of football makes it less meaningful than for baseball. It&#39;s as if every player on the field gets an &quot;assist&quot; for every play. Who gets credit for the WPA of a sack, the guy who drew the double-team block or the guy who was left unblocked? But it can be done, with some caveats.Brian Burkehttp://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-74776687058784490132009-06-03T21:51:01.460-04:002009-06-03T21:51:01.460-04:00The above post is about individual stats using WPA.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-10846714162667282402009-06-03T21:50:32.270-04:002009-06-03T21:50:32.270-04:00Can you do a WPA stat like in baseball. Find out how many probablility points a player lost and how many they gained on a play. For example, A receiver makes a 15 catch, he adds 3% to +WPA, he gets +0.03 +WPA points, the QB throws an incomplete pass to him and the WP losses by 2%, he gets a -0.02 -WPA points. Overall het gets a 0.01 WPA.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-83182016575457936632009-06-03T21:14:02.996-04:002009-06-03T21:14:02.996-04:00Jon-Thanks for the suggestion. The log scale would be scientifically useful, but could be too abstract for many people. Besides, the same play at the end of a close game, compared to an early play or play at the end of a blowout, will have a larger WP swing and also be more exciting.Brian Burkehttp://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-81385763971917600762009-06-03T20:16:44.654-04:002009-06-03T20:16:44.654-04:00Thanks, Brian. I;ve been digging these WPA charts ever since studes started using them at Hardball Times for baseball. I wonder if they would look better if they were on a log scale instead of linear scale. It seems like the right side might overstate how much excitement in a close game is at the end.Jonhttp://www.blogger.com/profile/07465013255057574351noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-76700305047546927682009-06-03T12:21:34.684-04:002009-06-03T12:21:34.684-04:00(ER) Entertainment RatingBrianhttp://www.blogger.com/profile/16462424620403517405noreply@blogger.com