[Husker] Average 4-year Recruiting Rankings and WL%

Fri Feb 1 12:19:35 CST 2013

Good questions.

2.  geography - I'm working on something like that now.  Lots of WL data
back to the earliest days of CFB, but not so much for recruiting.  Standby
for something on this soon.

1.  The correlation runs from High Rank to recruiting average, but not the
other way around.  I'm guessing this is because it's a self-selecting
sample.

I need good data for atheletes.  Particularly recruiting information for
each athlete, as well as data on who is actually still on each teams
roster.  If I could refine that information I'm confident that it would
show some kind of valid statistical relationship.  Can anyone supply
this/point me in the right direction?

On Fri, Feb 1, 2013 at 5:25 AM, Scott Stewart <fourtwophd at gmail.com> wrote:

> OMG, now I am obsessing about this. LOL.
>
> Question 1: Which way does the correlation run? Does success improve
> recruiting or recruiting breed success?
>
> Question 2: How does geography factor into recruiting? My impression has
> been that there are "pockets" of highly rated recruits (California, Texas,
> Florida come to mind) with a sprinkling randomly distributed throughout the
> country. My second impression is that recruits generally stay relatively
> local (I guess that could be defined as a 1,000 mile radius and/or attend a
> school in conference that represents their area). Again there are those
> that don't fit this trend.
>
> This is truely contagious and evil. LOL.
>
>
> On Thu, Jan 31, 2013 at 10:05 PM, Scott Stewart <fourtwophd at gmail.com>wrote:
>
>> Excuse the last post. I fired off to quick. I tried to pull it back, but
>> don't know how.
>>
>> I will give you that there is a correlation between recruiting and
>> performance. However, I do think you are overanalyzing this correlation and
>> in overstepping the data.
>>
>> My first concern with the data was in your original post. It was that the
>> best correlation between recruiting and performance (ranking) occurred at
>> six years past the recruitment class. As you noted that is an odd result.
>> Most of the recruits, if granted a red-shirt year would be gone in the
>> sixth year (I am assuming that year 1 would be the recruits first year out
>> of high school, not their Senior year in high school). To me that suggests
>> there is a third factor influencing the data. Does good recruiting begat
>> better recruits? or is something else at work.
>>
>> I was less concerned when that data did not hold for CFB as a whole.
>> Suggesting there is something going on in the NU data. I would suspect it
>> has something to do with coaching changes, or lack of them, during periods
>> of time.
>>
>>
>> On Thu, Jan 31, 2013 at 9:42 PM, Scott Stewart <fourtwophd at gmail.com>wrote:
>>
>>> I am sorry maybe I misunderstood your post.
>>>
>>> When you said, "My next project is to reject/fail this hypothesis:
>>> recruiting *affects* a great deal at the most elite levels"...I thought
>>> you were saying that better recruiting ratings lead to higher BCS rankings
>>> in elite football programs vs. non-elite programs. So to evaluate that
>>> recruiting affects, or causes, better rankings in these elite programs
>>> versus non-elite programs are you planning on employing some form of
>>> structural equation modeling design. And if so, what factors are you
>>> including with recruiting?
>>>
>>> Scott
>>>
>>>
>>> On Thu, Jan 31, 2013 at 9:11 PM, Paul Dalen <quesohusker at gmail.com>wrote:
>>>
>>>>  Do you think I don't know the difference between correlation and
>>>> causation?
>>>>
>>>> Paul Dalen
>>>>
>>>> On Jan 31, 2013, at 8:03 PM, Scott Stewart <fourtwophd at gmail.com>
>>>> wrote:
>>>>
>>>> Paul,
>>>>
>>>> Be careful when using correlations to imply "effect" or causation.
>>>> Correlation does not imply causations. There is a large correlation between
>>>> number of penguin populations and temperature and (my favorite) number of
>>>> pirates in the Caribbean and global warming though no one would believe
>>>> that penguins or pirates make the climate cold.
>>>>
>>>> It is equally likely that success is related to high recruiting
>>>> classes, or that a third factor (climate, coaching, conference) is a
>>>> co-variate.
>>>>
>>>> Scott
>>>>
>>>>
>>>> On Thu, Jan 31, 2013 at 1:32 PM, Paul Dalen <quesohusker at gmail.com>wrote:
>>>>
>>>>> I forgot to mention that for the 8 years looked at, when comparing the
>>>>> 4-year recruiting rank averages, 6 of 8 times the higher ranked team
>>>>> of the
>>>>> two won the National Championship.  The exceptions were Texas in 2005
>>>>> (an
>>>>> extremely close game and generally accepted upset) and Alabama in 2009
>>>>> (remember Colt McCoy went down minutes into the game).
>>>>>
>>>>> My next project is to reject/fail to reject this hypothesis:
>>>>>  Recruiting
>>>>> affects a great deal at the most elite levels of college football, but
>>>>> that
>>>>> effect diminishes significantly outside the Top-5 or Top-10.
>>>>>
>>>>>
>>>>>
>>>>> On Thu, Jan 31, 2013 at 12:25 PM, Paul Dalen <quesohusker at gmail.com>
>>>>> wrote:
>>>>>
>>>>> > Looking at teams that have played in the BCS Championship Game, they
>>>>> have
>>>>> > the following 4-year recruiting averages:
>>>>> >
>>>>> > ................................Top
>>>>> >
>>>>> 100..............................5-stars....................4-stars........................3-stars......................4-year
>>>>> > Recruiting Rank Average
>>>>> > BCS 1&2....................5.3 per year.....................3.3 per
>>>>> > year...............10.0 per year...............7.6 per
>>>>> > year........................9.8
>>>>> > AP Top 25..................2.3 per year.....................1.4 per
>>>>> > year................5.6 per year................9.7 per
>>>>> > year.........................31.7
>>>>> >
>>>>> > BCS CG participants average twice the number of top-100s, 5-stars,
>>>>> and
>>>>> > 4-stars as the rest of the AP Final Top-25.
>>>>> >
>>>>> > It's well established that exceptional recruiting does not predict
>>>>> final
>>>>> > record or ranking, but final ranking, particularly for the highest
>>>>> ranked
>>>>> > teams, predicts a great deal about sustained recruiting.
>>>>> >
>>>>> > Paul
>>>>> >
>>>>> >
>>>>> >
>>>>> > On Tue, Jan 29, 2013 at 4:36 PM, Skylar Dodds <sklarbodds at cox.net>
>>>>> wrote:
>>>>> >
>>>>> >> Lastly, and probably the most convincing start of them all for
>>>>> >> recruiting.  Not one team... Not one has won a BCS title without
>>>>> consistent
>>>>> >> top 15 recruiting leading up to it.  And most average in the top
>>>>> 10.  That
>>>>> >> is NOT an accident.  Yes, some teams can vastly outperform their
>>>>> ranking,
>>>>> >> but there's zero question those are exceptions or outliers.  Boise
>>>>> State
>>>>> >> had been an example for a while, Kansas State this year, etc.  But
>>>>> make no
>>>>> >> mistake, there is a strong correlation between recruiting and top 25
>>>>> >> finishes, and especially if you want a championship.
>>>>> >>
>>>>> >>
>>>>> >
>>>>> _______________________________________________
>>>>> husker site list
>>>>> husker at tssi.com
>>>>> http://romaine.tssi.com/mailman/listinfo/husker
>>>>>
>>>>
>>>>
>>>
>>
>