scott91575
Resident hater
If there is another thread on this, please merge.
Got into the usual debate about recruiting rankings not mattering, etc. So I decided to throw together a spreadsheet of Rivals recruiting rankings average (2005-2010) vs. Sagarin rankings average (2007-2012)
https://docs.google.com/spreadsheet/ccc?key=0AkyU81WIxFmCdG5OVmJtVU5zMzlWSXRXUkNxdjByT0E#gid=0
Notes:
1) This is not a complete list of all teams. I started with the top 40 teams from last years Sagarin ratings and then added a handful of teams after that. So this is not all D1 teams. When I have some more time I will try to complete it.
2) I used Sagarin over normal rankings so I could get numbers higher than 25, and I consider Sagarin one of the more reputable computer rankings
3) I used Rivals because, well, it was just easier and they have good records where all team rankings are laid out on the same page
4) I chose those years for some sort of overlap of talent. Obviously the 2007 and 2008 classes have the most impact since the players from those classes played in all of the years for the rankings. Yet I feel this was the best compromise.
5) I did not want to go any earlier than 2005 for Rivals since their ratings get a little wonky. I also did not want to use any Sagarin rankings earlier than 2007 since most of the impact players from that year would have been from the 2002-2004 recruiting classes.
6) I used Google docs since it was the only thing I had on hand and it was sort of a spur of the moment thing.
Correlation:
I did a quick correlation number. For those not used to it, 1 and -1 are perfect correlation. 0 is no correlation. For data like this, >0.8 or <-0.8 is good correlation, and between -0.5 and 0.5 is weak.
Correlation# = .65
So, not good correlation, but not irrelevant either. In other words, not the only factor in success, but certainly an important one. I know, duh! Yet it's nice to put it in data form.
The most underperforming team is Michigan and the most over performing team is Boise State. Rivals rankings and Sagarin rankings matched up perfectly for Rutgers.
When I am bored enough I will try and fill it out some more. I believe the correlation will actually get better, but I won't go into all of the reasons.
Pretty lame, I know, but next time you have the ever so tiresome debate here is at least a little bit of data.
Got into the usual debate about recruiting rankings not mattering, etc. So I decided to throw together a spreadsheet of Rivals recruiting rankings average (2005-2010) vs. Sagarin rankings average (2007-2012)
https://docs.google.com/spreadsheet/ccc?key=0AkyU81WIxFmCdG5OVmJtVU5zMzlWSXRXUkNxdjByT0E#gid=0
Notes:
1) This is not a complete list of all teams. I started with the top 40 teams from last years Sagarin ratings and then added a handful of teams after that. So this is not all D1 teams. When I have some more time I will try to complete it.
2) I used Sagarin over normal rankings so I could get numbers higher than 25, and I consider Sagarin one of the more reputable computer rankings
3) I used Rivals because, well, it was just easier and they have good records where all team rankings are laid out on the same page
4) I chose those years for some sort of overlap of talent. Obviously the 2007 and 2008 classes have the most impact since the players from those classes played in all of the years for the rankings. Yet I feel this was the best compromise.
5) I did not want to go any earlier than 2005 for Rivals since their ratings get a little wonky. I also did not want to use any Sagarin rankings earlier than 2007 since most of the impact players from that year would have been from the 2002-2004 recruiting classes.
6) I used Google docs since it was the only thing I had on hand and it was sort of a spur of the moment thing.
Correlation:
I did a quick correlation number. For those not used to it, 1 and -1 are perfect correlation. 0 is no correlation. For data like this, >0.8 or <-0.8 is good correlation, and between -0.5 and 0.5 is weak.
Correlation# = .65
So, not good correlation, but not irrelevant either. In other words, not the only factor in success, but certainly an important one. I know, duh! Yet it's nice to put it in data form.
The most underperforming team is Michigan and the most over performing team is Boise State. Rivals rankings and Sagarin rankings matched up perfectly for Rutgers.
When I am bored enough I will try and fill it out some more. I believe the correlation will actually get better, but I won't go into all of the reasons.
Pretty lame, I know, but next time you have the ever so tiresome debate here is at least a little bit of data.
Last edited:
