tag:blogger.com,1999:blog-38600807.post2935525445731951171..comments2023-11-05T04:16:44.937-05:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Playoff Projections - Week 12Unknownnoreply@blogger.comBlogger5125tag:blogger.com,1999:blog-38600807.post-16894986478400999872013-12-05T14:29:42.723-05:002013-12-05T14:29:42.723-05:00Toby - The model says the Eagles have a 66% proba...Toby - The model says the Eagles have a 66% probability against an average team (on a neutral field) and Dallas is a below average team. If you go through the math, the model says that the Eagles would have more than a 70% chance against Dallas on a neutral field. Because the game is in Dallas, the odds are more like 64%.<br /><br />Pointing to a random game and saying "I am skeptical about the model's output" isn't much of an argument. The Eagles have better efficiency numbers in several important categories. Those efficiency numbers are the basis of the model. If you want to argue against the model's probabilities for that particular game, you are probably going to have to point out systematic issues with Brian's logistic regression methodology.Anonymoushttps://www.blogger.com/profile/08342283313661814313noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-63716808374785163402013-12-02T11:39:32.145-05:002013-12-02T11:39:32.145-05:00Rob,
I disagree with your argument here. Using th...Rob,<br /><br />I disagree with your argument here. Using the model for a reference here is not very convincing. The model assumes a .66 win probability given any random opponent. If you were referring numbers in terms of DAL vs PHI you might have something to say ... I am totally skeptical of the 65/35 argument for PHI vs DAL, this makes the model lose credibility in my opinion.Tobyhttps://www.blogger.com/profile/16370333654512823284noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-55252462700966392032013-11-29T11:07:43.751-05:002013-11-29T11:07:43.751-05:00"there is no way that week 17 match is anythi..."there is no way that week 17 match is anything like a 66/34 proposition in their favor."<br /><br />The ANS win probability model disagrees with you: http://www.advancednflstats.com/2013/11/team-efficiency-rankings-week-12.html<br /><br />It thinks the Eagles are the 4th best team in the league (GWP = 0.66) where the Cowboys are the 22nd (GWP = 0.43). Look at the efficiency numbers in the same link. The Eagles have massively better pass Y/A and run success rate numbers, and everything else is comparable.<br /><br />J.D., you are right. The model does not account for "throwaway games." The only inputs into the equations are team efficiency numbers - nothing about current week number, playoff seeding, etc. is considered.Anonymoushttps://www.blogger.com/profile/08342283313661814313noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-17924644586300053542013-11-29T00:34:13.885-05:002013-11-29T00:34:13.885-05:00I bet this model doesn't take into account the...I bet this model doesn't take into account the possibility that in the last week or two of the season, some teams may have their playoff seeds locked up and therefore won't play at 100% (because of resting players they wouldn't otherwise). For instance, if the Seahawks beat the Saints and then have #1 seed wrapped up going into Week 16, would they still be just as strong against the Cardinals? I bet the model assumes they will.J.D. Krullnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-50154005536354108762013-11-28T20:59:07.062-05:002013-11-28T20:59:07.062-05:00I fail to see how Philly can be favored so decisiv...I fail to see how Philly can be favored so decisively for the NFC East (even before Dallas won today). In order for then to not need to win in Dallas in week 17 they need to be 2 games ahead going in, which seems unlikely. Even accounting for rating Philly better than Dallas there is no way that week 17 match is anything like a 66/34 proposition in their favor.Anonymousnoreply@blogger.com