Statistical projection time - Printable Version

+- The CardBoard (https://thecardboard.org/board)
+-- Forum: C-House! (https://thecardboard.org/board/forum-4.html)
+--- Forum: The CardBoard (https://thecardboard.org/board/forum-5.html)
+--- Thread: Statistical projection time (/thread-9244.html)

---

Statistical projection time - washingtonismoney - 05-21-2015

The two Football Outsiders projections have weighed in -- and they look pretty solid for us: Bill Connelly's S&P puts us at #11, and FEI puts us at #13.


Here's the opposition ranks:

Opponent	S&P	FEI
Northwestern	#62	#56
UCF	#60	#45
SC	#13	#13
OSU	#70	#72
Arizona	#35	#31
UCLA	#7	#6
Washington	#55	#53
WSU	#66	#79
Colorado	#79	#94
Oregon	#4	#3
59ers	#51	#71
N.D.	#16	#8

sources:
http://www.sbnation.com/college-football/2015/5/20/8625261/2015-college-football-rankings-projections
http://www.bcftoys.com/2015-fei

---

Re: Statistical projection time - CTcard - 05-21-2015

(05-21-2015, 09:01 AM)washingtonismoney link Wrote:The two Football Outsiders projections have weighed in -- and they look pretty solid for us: Bill Connelly's S&P puts us at #11, and FEI puts us at #13.

It is a little odd to run a projection of an after-the-fact descriptive performance stat. But, at least Connelly qualitatively tells us what he is doing - as far as I can tell Fremeau does not.

From Connelly there are a couple of interesting bits.

His two most important categories are the weighted performance over the past 5 years and "the likely changes associated with each team's returning starter figures". Since the latter is apparently not just number of returning starters, but changes from past performance, it is at least partially the same thing as the 5 year weighted past performance criterion.

I was rather amused at this juxtaposition:
"recruiting rankings are relatively solid predictive stats"
followed by,
"I blend them together based on what has produced the best results in the past. That means ... recruiting carrying the least [weight]."
That's not quite a contradiction, but close.
[url=http://www.footballstudyhall.com/2014/2/5/5382140/recruiting-matters-why-the-sites-get-the-rankings-right?_ga=1.134040525.234824359.1405719794][/url]

---

Re: Statistical projection time - Nan3cy - 05-21-2015

Isn't it a way of saying "I pay attention to recruiting rankings, but I do so as little as possible"?

---

Re: Statistical projection time - Hulk01 - 05-21-2015

Squares with my sense that we face four tough opponents, one risky one (Arizona), and seven ordinary ones.

But can anyone confidently predict UCLA and Oregon until tney see what their new quarterbacks can--and cannot-do.  If Rosen is Hundley-like, UCLA is very good. If he's not, they do well only because the Pac 12 has only three other very good teams.

Oregon has lost not just Mariota, but three excellent offensive linemen--Fisher (2d round), Grasu (ist team all con) and Stevens (second team).  A top five finish seems an audacious bet.

---

Re: Statistical projection time - fullmetal - 05-21-2015

I'm all for roast duck.  Delicious.

---

Re: Statistical projection time - 82lsju - 05-21-2015

(05-21-2015, 01:11 PM)fullmetal link Wrote:I'm all for roast duck.  Delicious.

---

Re: Statistical projection time - pcgoode - 05-21-2015

No team is perfect, but of the four difficult opponents we face, I would say only one of them has a proven quarterback - USC. 

---