tag:blogger.com,1999:blog-38600807.post4889127313349957938..comments2018-06-02T14:19:34.554-04:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Team Efficiency Rankings - Week 4Unknownnoreply@blogger.comBlogger49125tag:blogger.com,1999:blog-38600807.post-87799581838206991562012-10-01T15:54:55.049-04:002012-10-01T15:54:55.049-04:00Brian;
a question about this weeks ratings?
Have y...Brian;<br />a question about this weeks ratings?<br />Have you accounted for the fact dallas (and other teams) has played two road games out of 3, and some teams like Chi have played 2/3 at home? This accounts for a large +,-8% swing in stats that I feel needs to be adjusted for I have Dallas 71% vs Chicago (using a system similar your methodology.) <br />They have better stats at neutral site, have played a tougher schedule and 2 on the road? thanks Dandanhttps://www.blogger.com/profile/11047193053856110602noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-79856155854568073262012-09-27T23:19:14.492-04:002012-09-27T23:19:14.492-04:00Re: Brian's test using only the first 3 Q'...Re: Brian's test using only the first 3 Q's: It would prob. be too subjective, but any game that would have a point differential of >14 entering the 4th Q AND ends with a final differential of >9 might be worth dropping the 4th Q of data. In other words, the trailing team might end up with a TD early in the 4th Q to cut the deficit to 10--if they make it less, then they prob. had a chance, even if a slight one, to tie the game. These stats probably relate well to the teams' overall strength--for example, the DEN-ATL week 2 MNF. If the trailing team never could get within a potential tying score (for example, TNF's CAR-NYG game), then there is a good probability to disregard the 4th Q, as the leading team is prob. more interested in running clock than giving 100% effort and strategy toward scoring, and the trailing team is prob. not using the most optimum strategies that would correctly demonstrate (statistically) their strengths and weaknesses.<br />Brian, have you ever tried something this subjective (although with objective parameters?Josephnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-66205986534057039422012-09-27T19:39:04.050-04:002012-09-27T19:39:04.050-04:00David Cooper, the regression coefficients are from...David Cooper, the regression coefficients are from years and years of data, so adding 16 games to the thousands in the regression would have minimal impact on the coefficients.Jameshttps://www.blogger.com/profile/01838293735141324662noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-72691205294944256152012-09-27T16:03:02.752-04:002012-09-27T16:03:02.752-04:00Do the coefficients for the regression change on a...Do the coefficients for the regression change on a weekly basis as more data from the season is input?Anonymoushttps://www.blogger.com/profile/13253045774470895232noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-50351433784010452762012-09-27T12:11:18.949-04:002012-09-27T12:11:18.949-04:00Thanks for your response Brian;
Maybe in the offse...Thanks for your response Brian;<br />Maybe in the offseason we could look at it?<br />It is a game theory question as well?<br />How do teams play with a lead?<br />It would take a lot of time...<br />In the meantime I say Denver is overrated. :)danhttps://www.blogger.com/profile/11047193053856110602noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-64037591089907247732012-09-27T11:51:12.327-04:002012-09-27T11:51:12.327-04:00I think it also has to do with the fact they have ...I think it also has to do with the fact they have the highest Opp GWP. You would expect their numbers to be below average if they were an average team. So their own GWP gets corrected for this.Eddienoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-55297920032377009982012-09-27T11:29:13.696-04:002012-09-27T11:29:13.696-04:00I suppose run success rate is much more highly wei...I suppose run success rate is much more highly weighted than rushing YPC in the past. Denver is rated much higher offensively than say NE, but really the only advantage they have statistically is 6% in success rate running the ball.Wynnnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-67564044286032112042012-09-27T11:26:08.812-04:002012-09-27T11:26:08.812-04:00Confused as to why Denver is rated so high, despit...Confused as to why Denver is rated so high, despite only being marginally above average in everything aside from DRUNSR%?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-85046747034193722052012-09-27T10:33:19.702-04:002012-09-27T10:33:19.702-04:00Brian, I believe that you previously argued that D...Brian, I believe that you previously argued that DINT rate isn't very predictive, is this no longer the case?Calnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-68237055554535120172012-09-27T09:32:15.065-04:002012-09-27T09:32:15.065-04:00Unknown-I did once test limiting all stats to the ...Unknown-I did once test limiting all stats to the first 3 quarters. I realize that throws a lot of baby out with the bathwater. The results were slightly less accurate than with all the data, including trash time.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-82355151547889366032012-09-27T00:44:59.550-04:002012-09-27T00:44:59.550-04:00Can anyone post the predictions from nfl forcast.....Can anyone post the predictions from nfl forcast..I can't get java to work...would be interested to see what Brian's WP translate to spreads?Unknownhttps://www.blogger.com/profile/09013481141139711324noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-53839479826225157652012-09-27T00:29:25.410-04:002012-09-27T00:29:25.410-04:00....I'm convinced this strategy would be very .......I'm convinced this strategy would be very effective...DanUnknownhttps://www.blogger.com/profile/09013481141139711324noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-91551442047159785432012-09-27T00:26:29.949-04:002012-09-27T00:26:29.949-04:00Brian;
Have you investigated the role 'score...Brian;<br /><br />Have you investigated the role 'score effects' play on<br />eff. stats?. Ex. Denver ran up many successful plays down by three scores against Atlanta. This makes their Opass stronger than it really is? I am referring to situations more than 'just garbage time'. Would it be possible to <br />calculate the average net yds per pass at different game score states and use this to adjust?<br /><br /> I know in Hockey teams leading change their strategy allowing more shots than when tied. I then adjust accordingly, <br /><br />DO you think teams in NFl do the same thing? Do they play less aggressive protecting big plays a nd sacrifice net passing yds? IF so, wouldn't it improve your model to adjust each teams stats based on the state of the score differential during the game? any thoughts? Unknownhttps://www.blogger.com/profile/09013481141139711324noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-67790947355018714452012-09-26T23:14:32.822-04:002012-09-26T23:14:32.822-04:00"Each team variable is regressed again to acc..."Each team variable is regressed again to account for how reliable each particular facet is throughout a season."<br /><br />Is this a seperate regression? Is the reliability not accounted for in the coefficient?Calnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-39424026876845936432012-09-26T23:10:26.166-04:002012-09-26T23:10:26.166-04:00@ annonymous
Dont forget that random luck has an ...@ annonymous<br /><br />Dont forget that random luck has an even distribution (the difference between "random luck" and "making your own breaks" being that if a team is "luckier" then another, it will show in the stats as success rates, turnover percentage because that "luck" is more a matter of skill. Case in point, the Texan's couldn't buy a break 2 years ago, because they sucked.)TMKnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-54830341908659389332012-09-26T23:00:18.081-04:002012-09-26T23:00:18.081-04:00One correction to the Randy Mossesque Q&A &quo...One correction to the Randy Mossesque Q&A "No one outside of Texas..." should read "No one outside of North Texas and the Rio Grande Valley..." give the rest of us some credit! On a serious note, appreciate the hard work. Alot of the math is beyond me, but I know enough to get the point and understand the models. You do a great job of explaining the math so that "we" can follow along. Keep up the good work!Thatmcaueykidnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-19057343182288804652012-09-26T22:41:55.103-04:002012-09-26T22:41:55.103-04:00Playoff odds are available on nfl-forecast.com.
E...Playoff odds are available on nfl-forecast.com.<br /><br />Early SB favorites are the Texans over the Falcons.Chrishttp://nfl-forecast.comnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-63164283805791478142012-09-26T22:17:44.363-04:002012-09-26T22:17:44.363-04:00There are 2 official passing stats: net passing ya...There are 2 official passing stats: net passing yards and total passing yards. 'Net' means sack yards are included in the sum. Interceptions are always considered attempts.<br /><br />My own 'opass' is (total yards - sack yards) / (attempts + sacks)Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-32284369445242629392012-09-26T18:37:14.507-04:002012-09-26T18:37:14.507-04:00Brian;
Do you know if the official pass attempted...Brian;<br /><br />Do you know if the official pass attempted stats include (sacks and int) If not, do you choose to add them to attempted total before calculating opass eff. thanks<br /><br /><br /><br />Unknownhttps://www.blogger.com/profile/09013481141139711324noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-60249178741394418462012-09-26T17:46:43.164-04:002012-09-26T17:46:43.164-04:00Re: Challenge #2:
Just throwing it out there: The...Re: Challenge #2:<br /><br />Just throwing it out there: The efficiency model doesn't take into account good *strategic* coaching. You know that, but not all the readers do.Jameshttps://www.blogger.com/profile/01838293735141324662noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-42145272501591514052012-09-26T15:52:20.541-04:002012-09-26T15:52:20.541-04:00This is awesome. I recently gained a newfound exc...This is awesome. I recently gained a newfound excitement for football and the NFL through statistical prediction and analysis and have been searching for somewhere that did this type of number crunching. I look forward to seeing how this season progresses and might take a crack at tweaking the model presented.Anonymoushttps://www.blogger.com/profile/13253045774470895232noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-62265477462843926302012-09-26T14:42:25.665-04:002012-09-26T14:42:25.665-04:00Washington's low ranking makes perfect sense i...Washington's low ranking makes perfect sense if you look at the numbers. They've played the league's weakest schedule but are ranked just 18th on offense and 32nd on defense. They're decent at throwing the ball, and they run it reasonably well, but they fumble a lot and they have the league's worst penalty rate.kjbhttp://broomonthewarpath.sportsblog.com/noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-29345612490988964552012-09-26T13:33:24.664-04:002012-09-26T13:33:24.664-04:00Last week means, following week 2 of this season. ...Last week means, following week 2 of this season. I usually blank that out, but left it in because I was so pressed for time. It's illustrative just how unsettled these estimates are so early in the season. Keep in mind we have 50% more data than we did just one week ago. But as always, lots of grains of salt.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-32622871497154155142012-09-26T13:16:55.441-04:002012-09-26T13:16:55.441-04:00Gotta say, GB and WAS leading the way in Penalty r...Gotta say, GB and WAS leading the way in Penalty rate after TERRIBLE GAMES from the same officiating crew (albiet for only 1/3rd of those games) immediately screaqms out to me that the ref situation is heavily affecting these rankings. <br /><br />PS. I am a very biased Packer fan, but for one time only I feel okay about being thinking this for now without necessarily a great reason. We'll see how things play out..Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-30601784768809484032012-09-26T13:06:00.962-04:002012-09-26T13:06:00.962-04:00I'm wondering if Carolina's high ranking i...<br /><br />I'm wondering if Carolina's high ranking is because of their high passing YPA. I recall past articles here emphasizing how important passing offense is to winning, more so than defense or rushing offense. So having a strong passing attack may contribute greatly toward a high ranking, while having a weak defense doesn't hurt so much. Carolina's high INT and fumble rates don't ding them much because those stats are said to be more random, while their low penalty rate does help them out because penalty rate is supposed to be more predictive.<br /><br />Using “traditional” statistics, Carolina is tied for #1 in the NFL in raw yards per attempt with Cincinnati. They’ve scored far fewer points and thrown more interceptions and fewer touchdowns. Yet they are ranked better than Cincinnati. I wonder if Carolina isn’t so much “good” as they are good in just the right ways to achieve a high ranking by the WP model.Lucanoreply@blogger.com