YattaOkasan Posted July 15, 2016 Posted July 15, 2016 (edited) Bored on a Friday. I've been seeing a lot of discussion on Rex's sub par points allowed ranking with the Jets. Additionally I saw little discussion on how this may be related to fact that his offense was not very good (with respect to TOs). Below is the linear fit of turnovers to points allowed for the 2015 season. Not surprising there is a positive relationship although not statistically significant (its close). There are a couple of huge outliers like the NO saints and NY Giants terrible defense. On the other side, this makes Denver's D was very good at keeping the other team out of the endzone. I am thinking of expanding this data set to include previous seasons as well. Thoughts???? I can share my data set of course if people would like. Apparently don't know how to get the report into the post Edited July 15, 2016 by YattaOkasan
John from Riverside Posted July 15, 2016 Posted July 15, 2016 This might be a very simplistic way of looking at things....but If your offense is soooo BAD doesnt it stand to reason that your not going to do well in points allowed? Defense cant hold up all day long when they turn the ball over......cant score....cant get first downs?
YattaOkasan Posted July 15, 2016 Author Posted July 15, 2016 That was sorta what I was getting at. There does seem to be a correlation. I am interested in getting some more data. Check back in a couple days.
NoSaint Posted July 15, 2016 Posted July 15, 2016 This might be a very simplistic way of looking at things....but If your offense is soooo BAD doesnt it stand to reason that your not going to do well in points allowed? Defense cant hold up all day long when they turn the ball over......cant score....cant get first downs? or your defense is so bad that you have to play riskier ball, for some cases. (see the aforementioned saints - id be curious if they were removed, and maybe others that might be in the same boat, what the correlation would look like)
John from Riverside Posted July 15, 2016 Posted July 15, 2016 or your defense is so bad that you have to play riskier ball, for some cases. (see the aforementioned saints - id be curious if they were removed, and maybe others that might be in the same boat, what the correlation would look like) Yeah I dont know...for me that would go the opposite direction. If your D cant stop anyone then you HAVE to limit turnovers....because that is just the death blow. When you open up your offense that is when turnovers are more prone (in my opinion)
Luxy312 Posted July 15, 2016 Posted July 15, 2016 Did you try log or square root data transformation for the bivariate fit? It might get rid of some of the noise of the outlying data points. Just a thought.
YattaOkasan Posted July 15, 2016 Author Posted July 15, 2016 Did you try log or square root data transformation for the bivariate fit? It might get rid of some of the noise of the outlying data points. Just a thought. Good idea. Honestly the time consuming parts were data aggregation and figuring out how to post the results. Was interested to see what the group thought. I was also hoping additional data would help limit the outliers (that NO defense was terrible, as mentioned on this forum).
Luxy312 Posted July 15, 2016 Posted July 15, 2016 Good idea. Honestly the time consuming parts were data aggregation and figuring out how to post the results. Was interested to see what the group thought. I was also hoping additional data would help limit the outliers (that NO defense was terrible, as mentioned on this forum). One other random thought, but why not work with net turnovers and net points. If the offense is sloppy and giving up the ball, but the defense turns around and gets it back, isn't that a net zero from a turnover perspective? I would be serious money that if you did net turnovers and net points that you might even get close to a meaningful R Square. After all, aren't you basically trying to segregate the effects of turnovers on scoring overall and scoring allowed.
NoSaint Posted July 15, 2016 Posted July 15, 2016 (edited) Yeah I dont know...for me that would go the opposite direction. If your D cant stop anyone then you HAVE to limit turnovers....because that is just the death blow. When you open up your offense that is when turnovers are more prone (in my opinion) maybe misunderstanding? My point was teams like the saints have to air it out to score. Safe, ground and pound field position games dont work the opposition is getting 7 most possessions regardless of where they get the ball. you take risks, and hope you hit them. theres obviously a wide range in between the extremes there. but i was talking about seeing the correlation after removing a teams like that. Edited July 15, 2016 by NoSaint
YattaOkasan Posted July 15, 2016 Author Posted July 15, 2016 One other random thought, but why not work with net turnovers and net points. If the offense is sloppy and giving up the ball, but the defense turns around and gets it back, isn't that a net zero from a turnover perspective? I would be serious money that if you did net turnovers and net points that you might even get close to a meaningful R Square. After all, aren't you basically trying to segregate the effects of turnovers on scoring overall and scoring allowed. I can probably do that as well. TO differential was more readily available. My reason for not using it was looking into TOs it appeared that Ints forced, Fumbles forced, and fumbles given away all seemed to "random." I think the data I saw showed that there was no correlation between the Ints forced, fumbles forced, and fumbles given from the 1st half of a season to the 2nd half.
Recommended Posts