Great work here by Doug, highlighting potential regression (positive too) from a select group of RBs. I'd keep this one handy in your bookmarks for reference throughout the summer.

Look for similar analysis for the WR & TEs next week.

Started by Mike FF Today, Jun 12 2018 06:55 PM

14 replies to this topic

Posted 12 June 2018 - 06:55 PM

Great work here by Doug, highlighting potential regression (positive too) from a select group of RBs. I'd keep this one handy in your bookmarks for reference throughout the summer.

Look for similar analysis for the WR & TEs next week.

Follow me to the best fantasy draft program in the business.

_____________

Like FFToday on Facebook | Follow FFToday on Twitter

Posted 19 June 2018 - 09:41 AM

And here we are with the WRs and TEs as promised. Good work here by Doug.

Some interesting data on JuJu. I think most are expecting a significant boost in his production but the consensus is likely running a little hot right now.

Doug and have both projected about the same: 78-1014-7 & 75-1045-7.

_____________

Like FFToday on Facebook | Follow FFToday on Twitter

Posted 19 June 2018 - 09:07 PM

Perhaps I'm missing the point? There is no such thing as "positive touchdown regression." Scoring more TDs than the previous year is a progression.

STEELERS

Posted 20 June 2018 - 04:26 AM

Well it depends. For a data set, the term arithmetic mean, or sometimes "average" are used to refer to a central value. In this case simply the sum of the values divided by the number of values. This is for a basic discrete set of numbers.Perhaps I'm missing the point? There is no such thing as "positive touchdown regression." Scoring more TDs than the previous year is a progression.

So in a very basic view of the data, in this case a positive regression would be a "positive" (whether + or -) value compared to whatever you are measuring statistically from the previous cycle of measurement back towards the mean given that the previous measurement was actually "negative" or below the mean.

So if a player has a mean of 1000 yards rushing over 8 years, and last year he rushed for 800 yards, then 900 would be a positve regression. Now, that statement can be picked apart in many ways but I think that is the gist of what is being referred to here as "positive regression". No?

Posted 20 June 2018 - 06:37 AM

gist

“When I played pro football, I never set out to hurt anyone deliberately - unless it was, you know, important, like a league game or something” ~ D. Butkus

Posted 20 June 2018 - 08:47 AM

Good stuff from Doug. I scanned them real quick and saw a couple I may disagree with, but I'll have to sit and actually read the reasoning (when I get the chance), as it very well could change mind. Thanks for the info.

Posted 20 June 2018 - 01:29 PM

I thought the same thing and kind of still do. However, you've got to pick your battles in this life and this isn't one I'm willing to fight with the fantasy industry ... at least not yet.

Posted 20 June 2018 - 02:09 PM

It's a positive regression towards the mean. If used correctly in the fantasy stat analysis, that is valid terminology.

I thought the same thing and kind of still do. However, you've got to pick your battles in this life and this isn't one I'm willing to fight with the fantasy industry ... at least not yet.

Posted 20 June 2018 - 09:12 PM

It's a positive regression towards the mean. If used correctly in the fantasy stat analysis, that is valid terminology.

I understand. That is ultimately why I used it. It's just a bit of an oxymoron and it seems like there should be a better phrase to use.

Posted 23 June 2018 - 09:39 AM

Interesting analysis as always. Doug thank you.

I think it is often easier to pick negative regression vs positive. All the players IDd for + regression are over 30.

I believe relative skill position efficiency over the age of 28 is on a downward plane so this will be interesting at least as to the wr and rob positions. I think theres less predictability at the TE position.

I am starting to incorporate an age cap in my drafts. An older player has to fall a bit for me to draft that player

I think it is often easier to pick negative regression vs positive. All the players IDd for + regression are over 30.

I believe relative skill position efficiency over the age of 28 is on a downward plane so this will be interesting at least as to the wr and rob positions. I think theres less predictability at the TE position.

I am starting to incorporate an age cap in my drafts. An older player has to fall a bit for me to draft that player

Posted 24 June 2018 - 02:18 AM

Interesting analysis as always. Doug thank you.

I think it is often easier to pick negative regression vs positive. All the players IDd for + regression are over 30.

I believe relative skill position efficiency over the age of 28 is on a downward plane so this will be interesting at least as to the wr and rob positions. I think theres less predictability at the TE position.

I am starting to incorporate an age cap in my drafts. An older player has to fall a bit for me to draft that player

Thanks DrG. You seem to be as big of fan as my work as anyone on here. (Of which there are many ...)

I'm not following one of your comments about all of my positive regression candidates being over 30 though. I think less than half of my WRs/TEs are. I can only see on RB that is.

As for the age cap, I'm doing something similar. I'm not sure I have an official age yet for each position, but I have a general range. I'd say 28 is often the last "elite year" most RBs have.

Posted 25 June 2018 - 06:51 PM

I always consider FFtoday as my fantasy home. Thoughtful and intelligent analysis like this is the reason.

Posted 25 June 2018 - 09:10 PM

Well it depends. For a data set, the term arithmetic mean, or sometimes "average" are used to refer to a central value. In this case simply the sum of the values divided by the number of values. This is for a basic discrete set of numbers.

So in a very basic view of the data, in this case a positive regression would be a "positive" (whether + or -) value compared to whatever you are measuring statistically from the previous cycle of measurement back towards the mean given that the previous measurement was actually "negative" or below the mean.

So if a player has a mean of 1000 yards rushing over 8 years, and last year he rushed for 800 yards, then 900 would be a positve regression. Now, that statement can be picked apart in many ways but I think that is the gist of what is being referred to here as "positive regression". No?

Perhaps the point you are missing is that the mean is not over a eight year period, but rather just compared to last season's numbers. It is based on a one year cycle, not a cumulative view of data.

It's an oxymoron as said.

STEELERS

Posted 25 June 2018 - 10:17 PM

There has to be a mean for there to be positive regression. That's the definition of it in a way. Plus, you have to compare it to last year's stats in this case as that is what is being compared. Last year's stat vs. that which is projected this year based upon how that player performed in that area over his NFL career. Are you saying the player in question out performed his yearly average last year for stat in question?Perhaps the point you are missing is that the mean is not over a eight year period, but rather just compared to last season's numbers. It is based on a one year cycle, not a cumulative view of data.

It's an oxymoron as said.

Posted 29 June 2018 - 08:00 AM

I should have clarified .... rbs yes

Generally I think there is some wr fade after 28 but then there are guys like roddy white, Brandon Marshall etc. I think rbs are most productive under age 26

Generally I think there is some wr fade after 28 but then there are guys like roddy white, Brandon Marshall etc. I think rbs are most productive under age 26