tag:blogger.com,1999:blog-38600807.post3889264961046271699..comments2023-11-05T04:16:44.937-05:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Week 5 Efficiency RankingsUnknownnoreply@blogger.comBlogger11125tag:blogger.com,1999:blog-38600807.post-73004750251564489142008-10-08T16:23:00.000-04:002008-10-08T16:23:00.000-04:00Yup. What Josh said. But I'll have game probabilit...Yup. What Josh said. But I'll have game probabilities for week 6 up shortly so you don't have to do the math. But you can do hypothetical match-ups using that method.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-1669137018885834052008-10-08T12:16:00.000-04:002008-10-08T12:16:00.000-04:00Anonymous,To find out the probability for Team A p...Anonymous,<BR/><BR/>To find out the probability for Team A playing Team B, you would take both teams win% and put them into the log5 equation.<BR/><BR/>(A*(1-B)) / ((A*(1-B))+(B*(1-A)))<BR/><BR/>Example:<BR/><BR/>The Saints (58%) vs. The Raiders (52%)<BR/><BR/>(.58*(1-.52)) / ((.58*(1-.52)) + (.52*(1-.58)))<BR/><BR/>(.2784) / ((.2784)+(.2184))<BR/><BR/>=.5604<BR/><BR/>Saints win 56% of the time.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-6045688526764801272008-10-08T12:10:00.000-04:002008-10-08T12:10:00.000-04:00You can use R which can also be used into Excel. Y...You can use R which can also be used into Excel. You also have Gretl.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-81661936561622579192008-10-08T11:58:00.000-04:002008-10-08T11:58:00.000-04:00I absolutely love this site. Your work brings a b...I absolutely love this site. Your work brings a bit of science to a field that feels like it's filled with carnival con men. <BR/><BR/>Your site has caused me to start diving into learning more about regressions for the first time since my stats classes back at school. Would you have any suggestions on software to use for working on this type of problem. I've started playing around with R and elrm but was interested to see if there was anything else you would suggest.<BR/><BR/>Thanks again for the great content.johnbarthttps://www.blogger.com/profile/06785303602216595110noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-23731258086427653412008-10-08T11:34:00.000-04:002008-10-08T11:34:00.000-04:00I was curious about the following statement: "The ...I was curious about the following statement: "The GWP is the probability a team would beat the league average team at a neutral site." Is there a way for your audience to tweek these numbers if a team is playing a little bit better of a team then average? Not necessarily a good team but a team that's just better then the averageAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-84569980614427033772008-10-08T09:51:00.000-04:002008-10-08T09:51:00.000-04:00The official boxscore on NFL.com states that the r...The official boxscore on NFL.com states that the redskins fumbled the ball once. http://www.nfl.com/gamecenter/boxscore?game_id=29551&displayPage=tab_box_score&season=2008&week=REG2Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-35136757848599684242008-10-08T09:32:00.000-04:002008-10-08T09:32:00.000-04:00Yes, it accounts for all fumbles. But technically ...Yes, it accounts for all fumbles. But technically the Randle-El play was ruled a muffed punt and not a fumble because he was judged to not have possession.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-3642458386977451492008-10-08T08:33:00.000-04:002008-10-08T08:33:00.000-04:00Question on FUMRATE. I noticed that Washington ha...Question on FUMRATE. I noticed that Washington has a FUMRATE of 0. In week 2, however, Antwaan Randle-El fumbled a punt. Does this efficiency stat only account for fumbles that occur on offense and not on special teams?Brianhttps://www.blogger.com/profile/15394006910997218733noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-83928545592821883642008-10-08T06:56:00.000-04:002008-10-08T06:56:00.000-04:00Brian,I guess I should have been a bit clearer. My...Brian,<BR/><BR/>I guess I should have been a bit clearer. My problem with the soccer matches is the fact that there are ties. Since the probabilities are something like 40/35/25, is it still the same situation? I'm guessing it just makes the error higher, which isn't necessarily bad.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-87481884353756952332008-10-07T23:09:00.000-04:002008-10-07T23:09:00.000-04:00Josh-The Redskins have played very well against th...Josh-The Redskins have played very well against the most difficult schedule by far. They'll eventually have some turnovers and lose some games, but if you're a fan there's every reason to be optimistic. Their problem is their division.<BR/><BR/>For grading probabilities, one method is to use squared errors. For example, if you predict a .90/.10 game and get it wrong, that's a .9-0=.9 error. If you predict a .6/.4 game and get it right, that's a 1-.6=.4 error. You square all the errors then add them up, then take the square root. There's no "good" or "bad" error number. It's all relative depending on how predictable your subject is. It's useful for comparing various models. The idea is not only to identify the favorite, but to estimate the confidence level in the model.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-84251151415019405492008-10-07T22:43:00.000-04:002008-10-07T22:43:00.000-04:00Anytime the Redskins are at the top of a set of ra...Anytime the Redskins are at the top of a set of rankings, I should just leave well enough alone, but I'm curious as to how much a few turnover will hurt them. With them having fumble and interception rates at zero, even with regression to the mean factored in, they are still leaps and bounds ahead of the pack. Do you see a big drop after they pick up a few of each?<BR/><BR/>Also, I'm not sure if you ever click my link, but I put up soccer probabilities last week and I realized I had no idea how to grade the results. I'm not a big time stats whiz so I was wondering if you had any thoughts on how to determine the quality of the picks.Anonymousnoreply@blogger.com