tag:blogger.com,1999:blog-38600807.post7671331034749979859..comments2023-11-05T04:16:44.937-05:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Super Bowl XLVIII Game ProbabilityUnknownnoreply@blogger.comBlogger11125tag:blogger.com,1999:blog-38600807.post-32237472632796245282014-02-23T23:02:50.201-05:002014-02-23T23:02:50.201-05:00wooooooooooooooooooo hawks!wooooooooooooooooooo hawks!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-51102907561932649312014-02-01T12:05:30.088-05:002014-02-01T12:05:30.088-05:00Analytics favor Seattle; fans going by gut favor D...Analytics favor Seattle; fans going by gut favor Denver. Don't underestimate gut. Football has a huge chance element including officiating vagaries and creative game plans. My gut tells me Den cannot hold Lynch in check and contain Sea over the top. Will Carroll roll the dice? Zevgadolhttps://www.blogger.com/profile/09825529527014050056noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-82579725237469491332014-01-31T13:02:43.194-05:002014-01-31T13:02:43.194-05:00Vegas definitely uses ANS-style models to set line...Vegas definitely uses ANS-style models to set lines and spreads. Here's a rather good article on it:<br /><br />http://msn.foxsports.com/college-football/story/why-vegas-doesn-t-believe-auburn-is-nearly-as-good-as-fsu-121213Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-36569743933210617532014-01-31T12:37:39.060-05:002014-01-31T12:37:39.060-05:00> ... It doesn´t matter. The "keep the opp...> ... It doesn´t matter. The "keep the opposing offense off the field" <br />> approach failed more than once. ...<br /><br />I was talking about the past (i.e. this last season) and not the future. Seattle looks worse than Denver in stats like point differential because they've played a smaller number of possessions. If we compensate for that, the teams' point differentials are virtually identical.<br /><br />> ... I saw lower discrepancies between Brian's model and Vegas lines as <br />> the season went on ...<br /><br />Vegas is more sophisticated about getting data than most computer models. It generally takes a while for the computer models to come up to speed at the start of the season.Natenoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-13601132497213548402014-01-31T11:33:52.199-05:002014-01-31T11:33:52.199-05:00J.D., the problem with that is the assumption that...J.D., the problem with that is the assumption that the majority of gamblers on all games are professional gamblers. While that may very well be true to some extent during the NFL regular season (for example, I saw lower discrepancies between Brian's model and Vegas lines as the season went on, and while this could be random variation, it could be gamblers incorporating similar models to Brian's), the Super Bowl is an entirely different game. It's been estimated that 90% of Super Bowl betting handle comes from recreational bettors, and it certainly isn't surprising to see recreational bettors simplify the matchup as one between the two QBs and, as a result, back Peyton Manning.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-39851495111715845082014-01-30T19:36:24.933-05:002014-01-30T19:36:24.933-05:00Vegas started with a 2 pt line for SEA. So I think...Vegas started with a 2 pt line for SEA. So I think Brian and they agree. Later when the bets come in, it´s more of a public line b/c they (Vegas) have to even out the wagers on each side, thus to shift the line.<br /><br />Karl, GermanyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-31248522251798514872014-01-30T17:54:11.543-05:002014-01-30T17:54:11.543-05:00I think they already did.I think they already did.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-11766418432310482242014-01-30T17:48:16.268-05:002014-01-30T17:48:16.268-05:00It seems to me that the area in which analytics is...It seems to me that the area in which analytics is *least* needed is in looking at single-game win probabilities. Unlike other areas, which are neglected by "the establishment", future game results are under an incredibly sharp microscope, because gamblers are always looking for any edge they can get. Thus, the games are always exhaustively analyzed, both for things that can be numerically measured and things that can't. If Brian came up with something that was better than what the gamblers had, they would quickly steal/copy/co-opt it, and then it would be incorporated in the betting lines.J.D. Krullnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-14794118700795223912014-01-30T16:44:46.055-05:002014-01-30T16:44:46.055-05:00Nate, the winner of the game is also based on a &...Nate, the winner of the game is also based on a "per game basis".Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-61971436662244758052014-01-30T12:52:48.652-05:002014-01-30T12:52:48.652-05:00It doesn´t matter. The "keep the opposing off...It doesn´t matter. The "keep the opposing offense off the field" approach failed more than once. The best example is still the PIT-MIA 1984 game.<br />What if SEA playing it slow doesn´t work? Speak many 3-and-outs.<br />And comparing Pts per drive, I don´t know where you get the numbers from.<br />I looked at PFR. Denver with 2.83 PPD, SEA with 2.12 PPD. Defense; Denver with 1.83, SEA with 1.17. All adds up to an average of 2.00 for DEN, 1.975 for SEA.<br />About SoS: It depends on how you look at it. I compared all 18 games of each team and what their expected points are on average and median. No major difference, but wherever I look, DEN ends up as slight favourite. I see DEN as 1-point-Fav. (24-23) after calculating (almost) every imaginable predictive numbers... but as Brian pointed out yesterday, it will all come down to officiating (thanks football gods, the weather won´t play tricks with us), as usual if we look at the SBs since 2001. <br /><br />Karl, GermanyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-74312155226193982712014-01-30T12:01:40.296-05:002014-01-30T12:01:40.296-05:00I think one of the hidden differences is the dispa...I think one of the hidden differences is the disparity in pace: Denver plays some of the fastest football, and Seattle some of the slowest. That means that Denver is going to show better in any stat that measures per-game performance. Both teams net about 1 point per drive, but Denver has 20 more drives on the year, which nicely offsets their advantage in point differential.<br /><br />That said, by any reasonable model I come up with, the difference between the two teams is small enough that confidence in one, or the other's superiority is going to be very limited.Natenoreply@blogger.com