What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Regression to the Mean stats (by position) (1 Viewer)


Do you guys know of any good articles that compare Regression toward the mean across different fantasy positions? I'm trying to get some projections worked out for this year, and I'm wondering how much I should regress last season's performances toward the mean, as a whole.

I found this awesome article at ESPN that does this for IDP's (http://espn.go.com/fantasy/football/story/_/page/nfldk2k13idphistory/jj-watt-hard-time-following-amazing-2012-season), but wondering if you knew of any that compared regression for QBs vs. RBs vs. WRs vs. TE's. Thanks!

Last edited by a moderator:
Chase has some good RTTM articles on this site. Not sure if it matches what you are looking for but he's a smart guy. There's a thread or two around here also.

Thanks for the link MathNinja that was an interesting read. I wonder if ESPN pays Joe for using his term VBD? It amuses me how commonplace these types of ideas have become. :)

If you want to do RTTM analysis you can do so by position as was done in the article you linked. I think it is also useful to look at team data as well and would look to do that 1st before moving on to positions.

For example I think it is useful to know that league wide rushing yards/carry in 2011 and 2012 was 4.3 in 2010 this was 4.2 so in general all teams should be moving towards that number. Now this data could be misleading however due to QB runs and other WR runs. So separating the RB from the QB and WR is a good thing to do I think for this purpose, using whatever time frame/data set that you decide on. That would be a better way to project RB ypc than just the overall team ypc I think.

However the way teams do things are different than each other, so you also want to look at the teams averages I think before then moving to the individual projection.

Chase has said recently that YPC is not a useful predictive statistic. While that is likely true YPC still is a very relevant variable in any rushing projection. So this type of analysis while perhaps not predictive, is still something you want to know.

I really like your interest in these concepts and I look forward to reading observations you may arrive at through it, as well as discussion about the topic in general. This is something I would like to get better at.


Users who are viewing this thread