Does recruiting ranking predict performance?

Donate To GopherHole
denguegopher

denguegopher

Make lying wrong again
Joined
Dec 10, 2008
Messages
3,087
Reaction score
1,665
Points
113
I just stumbled across this in a professional journal, thought some might be interested:

Looks like yes and no

Authors:

Trent J. Herda, University of Oklahoma
Eric D. Ryan, Oklahoma State University
Jason M. DeFreitas, University of Oklahoma
Pablo B. Costa, University of Oklahoma
Ashley A. Walter, University of Oklahoma
Katherine M. Hoge, University of Oklahoma
Joseph P. Weir, Des Moines University
Joel T. Cramer, University of Oklahoma



"Abstract
The purpose of the present study was to examine the relationships among National Collegiate Athletic Association (NCAA) Division I football teams' 2002 recruiting rankings from the Rivals (RIV) and Scouts (SCO) recruiting services and the Jeff Sagarin end-of-season performance ratings from 2002–2006. The RIV and SCO recruiting services included rankings for 100 common NCAA Division I football teams for the 2002 recruiting season. Each recruiting service included a total point system rating (TOTPTS) and average star rating (AVESTAR). The Jeff Sagarin NCAA football ratings system was chosen as an indicator of the teams' performance. Pearson product moment correlation coefficients (R) and the corresponding predictive indices (R2) were used to examine whether the 2002 RIV & SCO TOTPTS and RIV & SCO AVESTAR ratings could predict the Jeff Sagarin end-of-season ratings and total number of wins for each football team for the 2002 through 2006 seasons. In addition, R and R2 values were computed to examine whether the 2002 Jeff Sagarin end-of-season ratings and total number of wins could predict the following season's recruiting rankings (2003 RIV & SCO TOTPTS and RIV & SCO AVESTAR). The results indicated that RIV & SCO TOTPTS and AVESTAR predicted up to 45% of the variances in the end-of-season ratings and total wins. Thus, other factors (besides recruiting rankings) must be contributing to the end-of-season ratings for the 100 NCAA football teams included in this study. In addition, up to 51% of the variance in RIV & SCO AVESTAR and TOTPTS was predicted by the previous year's end-of-season ratings or total wins, which suggests that more successful seasons tend to yield better subsequent recruiting classes. "
 

Yes it does.

If you win, you have recruited well.

If you lose, you have recruited poorly.
 

If you could provide a name of the study or link that would be cool. I have a feeling I'm going to be deeply dissapointed in the logic of this, but it's always fun to see how stupid smart people are.
 

No way you can define success/recruiting rankings from this analysis only. Can't really make an educated statement about the accuracy of the recruiting system until you have a large sample size of these types of analyses (maybe 10), each from different, unrelated university studies. With that kind of sample size you can key in on any glaring inaccuracies from a particular study and find what related trends are between studies.

I bet it would be awesome to have some of those professors/instructors listed above for a few dozen college credits.
 

Eight authors on the study -- impressive! The eighth author probably was the person fetching them all coffee.
 


Word said:
Eight authors on the study -- impressive! The eighth author probably was the person fetching them all coffee.
Click to expand...

no *&^!#*&^!#*&^!#*&^!# huh!! I did this for the big ten, and simultaniously another guy did pretty much the same thing. It took me about two hours. so ten times that for all of the teams.

It's not rocket scince. I'm sure they had to use pretty words and stuff but really eight people??

The data was already compiled too. build a spread sheet, do some data entry. WTF.

I say two. One to do the work, another colleague to check the work.

I think I want their jobs.
 

Schnoodler said:
no *&^!#*&^!#*&^!#*&^!# huh!! I did this for the big ten, and simultaniously another guy did pretty much the same thing. It took me about two hours. so ten times that for all of the teams.

It's not rocket scince. I'm sure they had to use pretty words and stuff but really eight people??

The data was already compiled too. build a spread sheet, do some data entry. WTF.

I say two. One to do the work, another colleague to check the work.

I think I want their jobs.
Click to expand...

I agree, I've also done something like this for a number of seasons (seems to be common around here lately) just for my own interest - doesn't take that long. Plus - they aren't gonna get much from one year's recruiting ranking.
 

In the years 2005, 2006, 2007 and 2008 Cincinnati, an undefeated team, recruited 14 players with three or more stars. In that same time, Boise State, another undefeated team, recruited 21. TCU had 38 players with three or more stars. The Gophers? We had an astounding 54 players with three or more stars!!!! How have we done on the field with all this talent? We are regressing. Plain and simple.

Relying on "star" talent may not be a good idea. Just ask Iowa. We go after David Pittman, a 4-star recruit, and he is NOTHING. Some of our two star recruits who are (or were) in the NFL include Russell, VanDeSteeg, Payne, Eslinger, Montgomery & Spaeth. Decker, aother 2 star, will soon join them.

What has Davis done for the offensive line? What has Fisch done for the offense? Where is the improvement in coaching? Who is getting better?
 

It's very predictive. Your focus on exceptions, is just that... exceptions. It can't predict the outcomes between two similar teams. It can't control the upsets that do occur. It doesn't account for injuries, drop outs, transfers, coaching changes and turnovers.

It also doesn't factor in coaching ability. Which is an essential part of the equation. But lets not get crazy and say because coaching matters talent does not. A good coach gets better with better talent. A team hopes for both.

Recruiting analysis can't predict the national champion. All it can do is provide a baseline as to what a reasonable expectation is for a team to help with analysis. So far from what I've seen most teams in the Big 10 finish very close to their talent levels. But there is always an exception up and down. This year, illinois is down, and Iowa is up.

If a team continually outperformed their talent level you could say the coaching was superior. you could also say if they underperformed the coaching is inferior. as with most things you need to see a multi year pattern, or a couple years of great disparity to draw conlusions with such a short season.
 



Schnoodler said:
It's very predictive. Your focus on exceptions, is just that... exceptions. It can't predict the outcomes between two similar teams. It can't control the upsets that do occur. It doesn't account for injuries, drop outs, transfers, coaching changes and turnovers.

It also doesn't factor in coaching ability. Which is an essential part of the equation. But lets not get crazy and say because coaching matters talent does not. A good coach gets better with better talent. A team hopes for both.

Recruiting analysis can't predict the national champion. All it can do is provide a baseline as to what a reasonable expectation is for a team to help with analysis. So far from what I've seen most teams in the Big 10 finish very close to their talent levels. But there is always an exception up and down. This year, illinois is down, and Iowa is up.

If a team continually outperformed their talent level you could say the coaching was superior. you could also say if they underperformed the coaching is inferior. as with most things you need to see a multi year pattern, or a couple years of great disparity to draw conlusions with such a short season.
Click to expand...

Perfectly stated.
 

upchurch said:
In the years 2005, 2006, 2007 and 2008 Cincinnati, an undefeated team, recruited 14 players with three or more stars. In that same time, Boise State, another undefeated team, recruited 21. TCU had 38 players with three or more stars. The Gophers? We had an astounding 54 players with three or more stars!!!! How have we done on the field with all this talent? We are regressing. Plain and simple.

Relying on "star" talent may not be a good idea. Just ask Iowa. We go after David Pittman, a 4-star recruit, and he is NOTHING. Some of our two star recruits who are (or were) in the NFL include Russell, VanDeSteeg, Payne, Eslinger, Montgomery & Spaeth. Decker, aother 2 star, will soon join them.

What has Davis done for the offensive line? What has Fisch done for the offense? Where is the improvement in coaching? Who is getting better?
Click to expand...

Do you bitch in everyone of your posts?

I think your stat about TCU, Boise & Cincinnati is missing something too...combined games against teams in the BCS top 25 = 1 (Boise State beat Oregon at Boise State when Oregon was giving 4 OL their first start).
 

upchurch said:
In the years 2005, 2006, 2007 and 2008 Cincinnati, an undefeated team, recruited 14 players with three or more stars. In that same time, Boise State, another undefeated team, recruited 21. TCU had 38 players with three or more stars. The Gophers? We had an astounding 54 players with three or more stars!!!! How have we done on the field with all this talent? We are regressing. Plain and simple.

Relying on "star" talent may not be a good idea. Just ask Iowa. We go after David Pittman, a 4-star recruit, and he is NOTHING. Some of our two star recruits who are (or were) in the NFL include Russell, VanDeSteeg, Payne, Eslinger, Montgomery & Spaeth. Decker, aother 2 star, will soon join them.

What has Davis done for the offensive line? What has Fisch done for the offense? Where is the improvement in coaching? Who is getting better?
Click to expand...

You've made this exact same post twice now, verbatim, with still a third being very similar aside from slight differences in the information presented.

How about this - the next time you're going to cut and paste this uninformed idiocy, give me a call, and I'll come over to your house, kick you in the nuts, and we'll call it a day? It'll save us all a lot of time and suffering.

Jesus.
 

dpodoll68 said:
How about this - the next time you're going to cut and paste this uninformed idiocy, give me a call, and I'll come over to your house, kick you in the nuts, and we'll call it a day? It'll save us all a lot of time and suffering.

Jesus.
Click to expand...

I don't think it'll save THAT much time. Unless you guys live close to eachother
 



grunkiejr said:
Do you bitch in everyone of your posts?

I think your stat about TCU, Boise & Cincinnati is missing something too...combined games against teams in the BCS top 25 = 1 (Boise State beat Oregon at Boise State when Oregon was giving 4 OL their first start).
Click to expand...


Sure if you are looking at this year alone but if you go back to 2005 which the poster was referencing then these programs have done very well against BCS level programs in relation to their level of recruits, TCU alone played several BCS programs in the regular season(Oklahoma, Texas, Texas Tech, Stanford, Baylor) and then add in their bowl game opponents. The fact is that many of these programs would love to play more top caliber BCS opponents but have a hard time getting teams to want to take the risk and play them because if they lose it looks bad. I know for a fact that this has been the case with TCU.
 


full manuscript

frozengopher said:
No way you can define success/recruiting rankings from this analysis only. Can't really make an educated statement about the accuracy of the recruiting system until you have a large sample size of these types of analyses (maybe 10), each from different, unrelated university studies. With that kind of sample size you can key in on any glaring inaccuracies from a particular study and find what related trends are between studies.

I bet it would be awesome to have some of those professors/instructors listed above for a few dozen college credits.
Click to expand...

I emailed the author and he sent me the link ; hit "download" :http://www.bepress.com/jqas/vol5/iss4/4/
 

I think it would be cool if they pursued answers to the remaining 55%. Some ideas:
- Number of home games
- Size of home stadium
- Average distance traveled to road games
- Legacy (maybe winning % over last 10 years)
- Average Sagarin ranking of opponent (strong schedule vs. weak schedule)
...

Of course some of these are ebedded in the Sagarin ranking but it would be cool if they could explain a larger portion of the variance.
 

What an ill concieved study. It is interesting to see the impact of one recruiting class as it matures but hardly a solid way to analyze the correlation between recruiting and future success. If they wanted to do it properly they would have measured the team strength based on the recruiting scores. And then compared that to actual results. From there they could have played with a weighting scale to see which weights most accurately predicted future reslults.

It seems like they did the least they could do, and consequently produced the least usable data.

Look at our freshman class, they have hardly touched the field. What possible impact could they have on our results this year? Meanwhile, what if that class was the only good one and the rest surrounding them were average or poor, again there would be little impact from the one class as they mature eventually becoming the seniors. Just a strange approach.

But they did show that winning breeds good recruiting. So that was handy info. the power of momentum.

I'll look at it again later, there is also some other very interesting things going on in the study I don't think they caught or were interested in.
 

denguegopher said:
I emailed the author and he sent me the link ; hit "download" :http://www.bepress.com/jqas/vol5/iss4/4/
Click to expand...

Thanks for the link, dengue. I was going through it and had terrifying flashbacks of Z scores, coefficients of determination, variance, and confidence intervals. I had quantitative stats less than a year ago, but have forgot pretty much everything.
 

frozengopher said:
Thanks for the link, dengue. I was going through it and had terrifying flashbacks of Z scores, coefficients of determination, variance, and confidence intervals. I had quantitative stats less than a year ago, but have forgot pretty much everything.
Click to expand...

It does hurt the brain. I'm still in pain and have avoided going back for a second look.
 

You must log in or register to reply here.



Top Bottom