PlayoffsPlease Posted January 26, 2019 Share Posted January 26, 2019 If you don't understand the difference between correlation and causation, I have lost you already. I took each team's spotrac cap allocation by positional group for each team, and then calculated the statistical correlation between the number of wins each team had. The summary is that there is no significant correlation between wins and any position, there is no significant statistical correlation for offense or defense or special teams as a whole. I also looked at the correlation for a teams "balance factor" . Balance factor is the difference between Offensive and Defensive cap allocation. Zero would be perfect balance. There is also no significant correlation between balance and winning. Finally I looked at total cap used. There is also no significant statistical correlation between total cap used and wins. Perfect correlation would be a value of one. Perfect negative no correlation would be a value of minus 1. Zero means perfectly uncorrelated. In the context that nothing reaches the level of being significant, the position with the worst negative correlation (ie more money spent = less wins) is RB. And the position with the best correlation is linebacker. As a group defense is slightly positive, offensive slightly negative. I have not spent too much time on the details. But I think the large number of teams eating bad QB contracts is the main thing that hammers offenses. That plus Rb. Interetingly Cap allocation to the "trenches" has no correlation at all. 1 Link to comment Share on other sites More sharing options...
Augie Posted January 26, 2019 Share Posted January 26, 2019 Is this what you do for fun in your free time? ? 3 Link to comment Share on other sites More sharing options...
Orlando Tim Posted January 26, 2019 Share Posted January 26, 2019 This is good info but is there anything about longer term trends? Did you do this yourself? Link to comment Share on other sites More sharing options...
White Linen Posted January 26, 2019 Share Posted January 26, 2019 So there's no correlation to causation? Got it. Link to comment Share on other sites More sharing options...
teef Posted January 26, 2019 Share Posted January 26, 2019 You think you’re better than me college boy? 5 Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 1 hour ago, Augie said: Is this what you do for fun in your free time? ? 50% of this board is people arguing which cliche is correct, when pretty much none of them are. 2 1 Link to comment Share on other sites More sharing options...
K-9 Posted January 26, 2019 Share Posted January 26, 2019 4 minutes ago, PlayoffsPlease said: 50% of this board is people arguing which cliche is correct, when pretty much none of them are. Yep. And 90% of the game is half mental. 1 Link to comment Share on other sites More sharing options...
hondo in seattle Posted January 26, 2019 Share Posted January 26, 2019 (edited) Interesting and maybe a little counter intiutive. I would have expected there to be a correlation between QB spending and wins. Seeing this makes me guess there’s no strong correlation between what a team pays for QB play and the quality of play they actually get. Nice work, PP. Shows another side of analytics. Edited January 26, 2019 by hondo in seattle Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 Just now, hondo in seattle said: Interesting and maybe a little counter intiutive. I would have expected there to be a correlation between QB spending and wins. Seeing this makes me guess there’s no strong correlation between what a team pays for QB play and the quality of play they actually get. The QB thing has multiple things going on. Teams like the Browns, Redskins and Vikings and 49ers have huge cap hits on QB with no production out of the spot. Other teams like the Chiefs and Rams have small cap hits and great production. Link to comment Share on other sites More sharing options...
entropyrules Posted January 26, 2019 Share Posted January 26, 2019 2 hours ago, PlayoffsPlease said: If you don't understand the difference between correlation and causation, I have lost you already. I took each team's spotrac cap allocation by positional group for each team, and then calculated the statistical correlation between the number of wins each team had. The summary is that there is no significant correlation between wins and any position, there is no significant statistical correlation for offense or defense or special teams as a whole. I also looked at the correlation for a teams "balance factor" . Balance factor is the difference between Offensive and Defensive cap allocation. Zero would be perfect balance. There is also no significant correlation between balance and winning. Finally I looked at total cap used. There is also no significant statistical correlation between total cap used and wins. Perfect correlation would be a value of one. Perfect negative no correlation would be a value of minus 1. Zero means perfectly uncorrelated. In the context that nothing reaches the level of being significant, the position with the worst negative correlation (ie more money spent = less wins) is RB. And the position with the best correlation is linebacker. As a group defense is slightly positive, offensive slightly negative. I have not spent too much time on the details. But I think the large number of teams eating bad QB contracts is the main thing that hammers offenses. That plus Rb. Interetingly Cap allocation to the "trenches" has no correlation at all. Did you run any tests for outliers that are skewing the data? looking at how close to zero all the R's are makes me think there won't be any but it would have been helpful to see the actual regression curves 1 Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 Just now, entropyrules said: Did you run any tests for outliers that are skewing the data? looking at how close to zero all the R's are makes me think there won't be any but it would have been helpful to see the actual regression curves Just took all the data in spotrac exactly as is. Of course their are outliers. But the truth is there are many strategies to pursue to end up between 6-10 and 10-6 Link to comment Share on other sites More sharing options...
Foxx Posted January 26, 2019 Share Posted January 26, 2019 (edited) 2 hours ago, PlayoffsPlease said: If you don't understand the difference between correlation and causation, I have lost you already. ... so... not understanding the difference between causation and correlation led you to opine about a correlation effect in which you found no causal reaction? did i get that right? Edited January 26, 2019 by Foxx Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 Just now, Foxx said: so... not understanding the difference between causation and correlation led you to opine about a correlation effect in which you found no causal reaction in? did i get that right? no 1 Link to comment Share on other sites More sharing options...
Foxx Posted January 26, 2019 Share Posted January 26, 2019 Just now, PlayoffsPlease said: no so, no correlation at all then. gotcha. 1 Link to comment Share on other sites More sharing options...
machine gun kelly Posted January 26, 2019 Share Posted January 26, 2019 3 hours ago, PlayoffsPlease said: If you don't understand the difference between correlation and causation, I have lost you already. I took each team's spotrac cap allocation by positional group for each team, and then calculated the statistical correlation between the number of wins each team had. The summary is that there is no significant correlation between wins and any position, there is no significant statistical correlation for offense or defense or special teams as a whole. I also looked at the correlation for a teams "balance factor" . Balance factor is the difference between Offensive and Defensive cap allocation. Zero would be perfect balance. There is also no significant correlation between balance and winning. Finally I looked at total cap used. There is also no significant statistical correlation between total cap used and wins. Perfect correlation would be a value of one. Perfect negative no correlation would be a value of minus 1. Zero means perfectly uncorrelated. In the context that nothing reaches the level of being significant, the position with the worst negative correlation (ie more money spent = less wins) is RB. And the position with the best correlation is linebacker. As a group defense is slightly positive, offensive slightly negative. I have not spent too much time on the details. But I think the large number of teams eating bad QB contracts is the main thing that hammers offenses. That plus Rb. Interetingly Cap allocation to the "trenches" has no correlation at all. I appreciate you’re homework and thought put into this breakdown, but don’t insult some very intelligent posters on this board regarding you’re first statement. Link to comment Share on other sites More sharing options...
JOE IN HAMPTON ROADS Posted January 26, 2019 Share Posted January 26, 2019 you lost me at "allocation" Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 21 minutes ago, JOE IN HAMPTON ROADS said: you lost me at "allocation" Fair enough. I am not clear on what the proper technical term is. But this is the total amount of cap hit for the position, regardless of who plays. For example the cap allocation to QB for the Browns is 24.2 Million. Link to comment Share on other sites More sharing options...
formerlyofCtown Posted January 26, 2019 Share Posted January 26, 2019 3 hours ago, PlayoffsPlease said: If you don't understand the difference between correlation and causation, I have lost you already. I took each team's spotrac cap allocation by positional group for each team, and then calculated the statistical correlation between the number of wins each team had. The summary is that there is no significant correlation between wins and any position, there is no significant statistical correlation for offense or defense or special teams as a whole. I also looked at the correlation for a teams "balance factor" . Balance factor is the difference between Offensive and Defensive cap allocation. Zero would be perfect balance. There is also no significant correlation between balance and winning. Finally I looked at total cap used. There is also no significant statistical correlation between total cap used and wins. Perfect correlation would be a value of one. Perfect negative no correlation would be a value of minus 1. Zero means perfectly uncorrelated. In the context that nothing reaches the level of being significant, the position with the worst negative correlation (ie more money spent = less wins) is RB. And the position with the best correlation is linebacker. As a group defense is slightly positive, offensive slightly negative. I have not spent too much time on the details. But I think the large number of teams eating bad QB contracts is the main thing that hammers offenses. That plus Rb. Interetingly Cap allocation to the "trenches" has no correlation at all. There are far too many variable for there to be any correlation. I think where you may find some positive correlation is with successful drafting. Link to comment Share on other sites More sharing options...
Augie Posted January 26, 2019 Share Posted January 26, 2019 If we use all that cap space on giving our linebackers HUGE raises, that has to lead to fantastic things! 1 Link to comment Share on other sites More sharing options...
PlayoffsPlease Posted January 26, 2019 Author Share Posted January 26, 2019 (edited) 19 minutes ago, formerlyofCtown said: There are far too many variable for there to be any correlation. I think where you may find some positive correlation is with successful drafting. I don't think there are two many variables. Compared to other complex systems, there really are not. I think the problem is that there are multiple strategies that have equal chance for success, and I suspect coaching and adaptability to available assets is more important than having the right "philosophy" or "focus" . I The Rams and NE have nearly identical spends at every position except for QB and Defensive Line. Essentially the Rams have invested the difference in Brady and Goff in Donald. (Oline too) 7 minutes ago, Augie said: If we use all that cap space on giving our linebackers HUGE raises, that has to lead to fantastic things! either you are making a lame facetious remark, or you don't get the concepts at all. In either case. ? Edited January 26, 2019 by PlayoffsPlease Link to comment Share on other sites More sharing options...
Recommended Posts