tag:blogger.com,1999:blog-38600807.post9155321443669744959..comments2018-06-02T14:19:34.554-04:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Playoff Probabilities: Week 10Unknownnoreply@blogger.comBlogger12125tag:blogger.com,1999:blog-38600807.post-5128739250524617372011-11-12T15:23:56.630-05:002011-11-12T15:23:56.630-05:00If you know, or if it's not too much trouble t...If you know, or if it's not too much trouble to run it, how does the Bears-Lions game look in terms of leverage?Zedhttps://www.blogger.com/profile/18295422294055414260noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-17106408911910323012011-11-12T10:57:55.055-05:002011-11-12T10:57:55.055-05:00bigmouth: I've never heard that before either,...bigmouth: I've never heard that before either, and it's certainly not in the first chapter of my Intro to Stats book.<br /><br />That said, I suppose I could understand the argument if you're in a situation, as here, where a team could feasibly have a 100% probability and you want to differentiate between them and a team having a probability of less than 100% but greater than 99.5. Likewise for differentiating between a 0% probability and non-zero probabilities that have been rounded down.<br /><br />For the record, neither Green Bay nor San Francisco has clinched a berth yet, and the playoffs remain a mathematical possibility for every team (yes, even the Colts).Josh Katzhttps://www.blogger.com/profile/06646400031653670129noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-34657405343629832392011-11-11T16:35:03.857-05:002011-11-11T16:35:03.857-05:00Tangentially related question that arose on Niners...Tangentially related question that arose on Niners Nation. Some posters were quite insistent that it's bad practice in statistical analysis to round up probabilities to 100%. Someone even claimed it's in the first chapter of any Intro to Stats textbook.<br /><br />Yet I've never, ever heard this before. Anyone else?bigmouthhttps://www.blogger.com/profile/00336001437381184261noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-66504108610186425582011-11-11T15:32:39.442-05:002011-11-11T15:32:39.442-05:00robbbbbb --
You should run my software. It's ...robbbbbb --<br /><br />You should run my software. It's easy and free.<br /><br />Here are odds fro the worst 5 teams and the first 5 picks:<br /><br />Indy: 82%, 11%, 4%, 2%, 1% <br />Miami: 7%, 25%, 17%, 12%, 9% <br />Arizona: 4%, 18%, 16%, 13%, 11% <br />Seattle: 2%, 12%, 13%, 11% 10%<br />St. Louis 1%, 9%, 10%, 10%. 9%Chrishttp://nfl-forecast.comnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-8958389036534952432011-11-11T15:03:43.340-05:002011-11-11T15:03:43.340-05:00This is fun stuff; thanks.
Can you please start p...This is fun stuff; thanks.<br /><br />Can you please start putting some information in on draft position? Those of us who are fans of also-ran teams (I'm looking at you, Seahawks) would love to see some data on what the draft's likely to look like.<br /><br />Stuff on #1 pick probabilities is fun, but maybe a little deeper look (top 3? top 5?) would be informative and fun, too.<br /><br />Thanks much!robbbbbbhttp://robbbbbb.livejoural.com/noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-24610456868647097582011-11-11T09:38:29.507-05:002011-11-11T09:38:29.507-05:00OAK now jumps ahead of SD with their win last nigh...OAK now jumps ahead of SD with their win last night. Total chance of making playoffs: OAK: 46%, SD: 39%.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-75616200392302545952011-11-10T17:12:45.099-05:002011-11-10T17:12:45.099-05:00Mistake! Sorry guys, I realized the OppGWP chart e...Mistake! Sorry guys, I realized the OppGWP chart erroneously included the games from Week 9... Chart now updated to reflect the real values.Josh Katzhttps://www.blogger.com/profile/06646400031653670129noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-16292086736879112762011-11-10T16:16:35.016-05:002011-11-10T16:16:35.016-05:00good point anonymous...
I can see the merits of ...good point anonymous... <br /><br />I can see the merits of both sides. Median will basically tell you how many games are "tough opponents" vs "easy opponents".<br /><br />On the other hand the Mean gives you and average of how good or how bad each team is... obviously a GWP team of .9 is a much tougher matchup than one at 0.6, and your chances of winning that individual game are much less, but then looking at the average it skews the rest of the games.<br /><br />I think you probably need both pieces of information to get the real answer.Mattnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-15538073750268487032011-11-10T16:11:04.679-05:002011-11-10T16:11:04.679-05:00James: That's a great point. I hadn't real...James: That's a great point. I hadn't realized how often San Francisco makes it in as the #2 seed.<br /><br />Brian: Nice! I must have seen that graph at some point in the long, long ago and filed it away for inspiration to strike.<br /><br />Anon: As for mean vs. median (though to be honest, I didn't give it that much thought beforehand), while it's true that as you say, a game against a strong team can lose you at most 1 game, you are much more likely to lose that game, so it has a greater impact on a your expected wins.<br /><br />For example, if Team A plays teams with GWPs of .450, .500, and .900, their number of expected wins is going to be much lower than Team B playing teams with GWPs of .450, .500, and .550 (even though the medians are identical).Josh Katzhttps://www.blogger.com/profile/06646400031653670129noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-43063085963187881592011-11-10T15:20:05.794-05:002011-11-10T15:20:05.794-05:00Why average OppGWP instead of median? Average make...Why average OppGWP instead of median? Average makes sense for normalizing stats that might scale more smoothly with opponent strength (YPA, SR%, etc.) but for # of games won, a game against a particularly strong team can lose you at most 1 game, regardless of how strong they are (and vice versa for particularly weak teams.)<br />A quick spreadsheet to show the difference: https://docs.google.com/spreadsheet/ccc?key=0At2xGu_ne0lcdGZjZjBlN0pLNTkwakJLMHNGa3NFS0EAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-69142250538692997952011-11-10T15:17:35.588-05:002011-11-10T15:17:35.588-05:00Man, that heat graph really takes me back! Check o...Man, that heat graph really takes me back! <a href="http://www.advancednflstats.com/2007/03/predicting-playoff-races.html" rel="nofollow">Check out this one from the 2006 season.</a><br /><br />Fantastic post. I was just scribbling together a post about CIN and Josh just did all the hard work for me.<br /><br />An interesting thing about the PIT-BAL race is that they've both won a strenth-of-schedule game so far (BAL over NYJ, PIT over NE), but despite being in 1st place last year, PIT draws KC and BAL draws SD. I think SD is the much better team despite the fluky loss to KC.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-33678927591944170622011-11-10T15:08:12.563-05:002011-11-10T15:08:12.563-05:00Considering how likely it is that Green Bay will g...Considering how likely it is that Green Bay will get the #1 seed and San Fransisco the #2 seed in the NFC, that makes a significant impact on the other playoff teams' postseason success depending upon seeding pairings.<br /><br />Specifically, the #3 seed is vastly more desirable than the #4 seed, and the #5 more than the #6 (as the #6 is locked to play GB if they advance, whereas #5 has a chance to avoid GB). Obviously this is the way the seedings are intended to be, but I think this year it has the potential to be a much more significant difference than normal.Jameshttps://www.blogger.com/profile/01838293735141324662noreply@blogger.com