Stanford opens as 3.5 pt favorite over Colorado... - 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: Stanford opens as 3.5 pt favorite over Colorado... (/thread-12873.html)

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - BostonCard - 10-17-2016

In both AP and USA today rankings, Colorado is tops among the "others receiving votes" category, which would place it as the 26th best team.  Stanford also received some votes, but fewer (equivalent to a 32/33 ranking).  If someone had told me that the Stanford/Colorado team would feature a ranked team, I wouldn't have been surprised, but for that ranked team to almost be Colorado is very impressive.

BC

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - stupac2 - 10-17-2016

(10-17-2016, 12:25 PM)DC 86 link Wrote:The problem with Sagarin's (and others') predictive models is primarily that they are purely statistical with no adjustments for factors that may have distorted previous results (e.g. lucky plays, officiating errors, personnel changes) or will impact future results (e.g. injuries, weather conditions).

Well, sure, but really the problem is that you can't perfectly predict the world, and college football is a subset of the world. There are systems that do a better or lesser job, but there's not really any way to talk sensibly about the metagame without relying on analytics like these.

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - DC 86 - 10-17-2016

(10-17-2016, 01:12 PM)stupac2 link Wrote:[quote author=DC 86 link=topic=15920.msg176277#msg176277 date=1476732357]The problem with Sagarin's (and others') predictive models is primarily that they are purely statistical with no adjustments for factors that may have distorted previous results (e.g. lucky plays, officiating errors, personnel changes) or will impact future results (e.g. injuries, weather conditions).

Well, sure, but really the problem is that you can't perfectly predict the world, and college football is a subset of the world. There are systems that do a better or lesser job, but there's not really any way to talk sensibly about the metagame without relying on analytics like these.
[/quote]

Absolutely, statistical analysis is the starting point for any retrodictive or predictive effort. My point is that betting line movements are much more reliable than purely statistical models, in the same way that prediction markets have generally proven to be more accurate than even the most sophisticated political polls.

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - fullmetal - 10-17-2016

I like our chances against Colorado if we can get people healthy again.  Meeks has to play.  Shawngren have to keep evolving their playcalling--Notre Shame was a great indicator of good things to come.

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - CrazedZooChimp - 10-17-2016

(10-17-2016, 01:54 PM)DC 86 link Wrote:[quote author=stupac2 link=topic=15920.msg176285#msg176285 date=1476735148]
[quote author=DC 86 link=topic=15920.msg176277#msg176277 date=1476732357]The problem with Sagarin's (and others') predictive models is primarily that they are purely statistical with no adjustments for factors that may have distorted previous results (e.g. lucky plays, officiating errors, personnel changes) or will impact future results (e.g. injuries, weather conditions).

Well, sure, but really the problem is that you can't perfectly predict the world, and college football is a subset of the world. There are systems that do a better or lesser job, but there's not really any way to talk sensibly about the metagame without relying on analytics like these.
[/quote]

Absolutely, statistical analysis is the starting point for any retrodictive or predictive effort. My point is that betting line movements are much more reliable than purely statistical models, in the same way that prediction markets have generally proven to be more accurate than even the most sophisticated political polls.
[/quote]

Well, as I said, in this case the betting line is almost exactly identical to Sagarin's predictor line, so maybe his method is working here...

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - BostonCard - 10-17-2016

(10-17-2016, 12:25 PM)DC 86 link Wrote:[quote author=CTcard link=topic=15920.msg176255#msg176255 date=1476726793]
[quote author=CrazedZooChimp link=topic=15920.msg176253#msg176253 date=1476726398]

I'm actually surprised the "recent" model has us ranked so high, seeing as we're 1-2 in our past three games with an ugly score differential.

Indeed, sometimes the ratings Sagarin posts are bizarre enough that I suspect actual calculating errors.
Of course since he doesn't publish exactly what he does, no one can check.
[/quote]

Sagarin's highly flawed models have been reverse engineered by various groups over the years. The "Predictor" model basically consists of weighted linear regression using almost-uncapped score margins. Sagarin's season ending ratings have retrodictive value, but any objective analysis (measuring deviation from the final score) concludes that the consensus opening and closing betting lines are far stronger predictors. The problem with Sagarin's (and others') predictive models is primarily that they are purely statistical with no adjustments for factors that may have distorted previous results (e.g. lucky plays, officiating errors, personnel changes) or will impact future results (e.g. injuries, weather conditions).
[/quote]

"All models are wrong. Some are useful."

-- George P. Box

Of course betting lines will be better than any computer system, because if computer models were better, you would have a chance at arbitrage. In that sense, the efficient market hypothesis applies to betting markets. The betting models incorporate information from the computer ones, plus all the subjective matters you mention.  There is no way to systematically apply an adjustment based on external factors to a computer model, so any such application would be ad hoc (say, deducting a couple points because of McCaffrey's injury), but how do you determine the magnitude of the deduction?

[s]Sagarin is a tool[/s]. The Sagarin rating system is a tool.  It can offer some insight about games.  As long as you understand its intrinsic limitations, it is still useful, which is about all you want from something like that.

Beats the hell out of Jon Wilner's twisted insights in his rankings.

BC

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - JJJ - 10-17-2016

Speaking of Wilner...?

http://www.mercurynews.com/2016/10/17/stanford-football-report-card-grading-the-win-over-notre-dame/

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - BostonCard - 10-17-2016

I have no quibble with the final grade; a "C"  is harsh but fair given that we played against a meh team in a yucky game where neither team played well but Notre Dame clearly played worse.  I do think that the B- for the defense was unjustified, given that the unit was outscored by Notre Dame 10-9, meaning that all the offense had to do to win the game was kick a field goal.  Add to that the fact that we still had one DB out, and I thought the unit deserved better than a B-.  And although he is quick to note that ND's defense is bad (and thus our offensive output should be taken in that context), he neglects to mention that ND's offense has been pretty decent (at least until the last two weeks).

BC

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - washingtonismoney - 10-17-2016

Right. And last week's performance was hurricane-influenced.

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - 82lsju - 10-17-2016

is this a hint that CMAC will be back on Saturday??

---

Re: Stanford opens as 3.5 pt favorite over Colorado... - stupac2 - 10-18-2016

(10-17-2016, 08:44 PM)eric link Wrote:is this a hint that CMAC will be back on Saturday??

Didn't we think there was a hint in the WSU hype video that Meeks would be back for that game? I wouldn't read much into it.

---