4th attempt at post...Well, Doc, we're all out of "reply" buttons, which should be a signal to me to not write this, but, while I agree with what you are saying, my original objection to the way the CU study was being presented by Warlord (and many pundits around the airwaves) was as a model, so much better than any of the others because it had a perfect record since 1980. Quote from Warlord: " CU data analysis agrees with my assessment, and they've never been wrong. They even called the split-election in 2000 perfectly. CU has Romney at over 330." While it's true that "they've never been wrong". It's also true that they've NEVER BEEN RIGHT!...read on.
I'm pretty sure it is as you said "They could easily have mined the variables to fit the "wins" and thereby gamed the system,". I don't think the good professors were trying to deceive, as they ready admit their model has made only one prediction and that is for 2012. All the other, so-called, right answers were simply back-testing against known result to make sure the model was behaving as it should, since it was built to return the right results of those known outcomes.
Right now the CU model is 0 for 0, not 8 for 8. We'll find out tomorrow (hopefully) if it's 1 for 1 or 0 for 1.
Yahoo's new boards really aren't very good. I've tried this a few times, now and hasn't shown up...could they are filtering for Romney?. I'm not for either...I'm voting for Rosanne Barr!
The point is...the CU data has never been correct. THEY HAVE NO TRACK RECORD!!! THEY HAVEN'T PREDICTED ANYTHING EXCEPT 2012!!!
Look at their own papers...they created the model AFTER the 2008 election. How can it have been "not wrong since 1980" ??? It was created using the election data from prior years starting in 1980, so OF COURSE THEY GOT IT RIGHT!!!...but, sorry, that doesn't mean it can get anything right that hasn't already happened.
It's truly amazing to me how much attention this prediction has gotten. I don't get it.
Happy election day, to all...
In a "model" for prediction, the model can be "made" at any time(not necessarily prior to the events it is used to predict).
In this case with CU, there are certain parameters they use to say if X, Y and Z are present and A, B and C are not present, then the outcome of the race favors the incumbent. If X, Y and Z are not present and A, B and C are present then the election favors the opposition to the incumbent.
They then can "Look Back" with their model and apply these parameters to each election and "ask" the question whether the model was accurate in predicting the outcome of that given election when their "predictors" were applied.
Having done that, they can then say their model was accurate or not accurate for a given historical election. This is what they meant by.. "not wrong since 1980" , even though their model was created many years later.
It's not worthwhile for management to buy shares unless the price falls to 29. Obama is almost certain to win, so we might see a considerable correction after elections.
At least I'm hoping for one so I can buy AGNC for a discount.
In all things, seek the truth.
If you think O is going to win, it is probably because you saw that somewhere.
If you analyze polls and get neck-deep in details, you'll find there is a built-in assumption that there will be the same turnout as 2008. If you re-weight the polls based on how people identify themselves today, rather than in 2008 (four years makes a big difference) you will find quite the opposite.
I've spent about 100 hours analyzing the polls, maps and underlying data. As of right now I've got a very high probability of a Romney win. I'm currently, conservatively, scoring Romney at 291, with reasonable chances at Wisconsin and PA. CU data analysis agrees with my assessment, and they've never been wrong. They even called the split-election in 2000 perfectly. CU has Romney at over 330.
I'm not trying to be political... the impact is significant to REITs. Romney wants to fire the fed chief and end QE, while O has promised "more of the same".
When all others will marvel at "how wrong the polls were," I hope you will remember what I told you about actually analysing what you've been told rather than merely believing what you were told. Everyone, even myself, has bias. In all things, seek the truth.
Either I will be discredited, or almost everyone you've been listening to will be discredited.
Virginia will reveal the bias, and whomever wins Virginia early in the night will also take Florida, Ohio, Iowa, and Colorado. There's a slightly more leftward bias in NH, WI and PA. NC is already in-the-bag for Romney... I have no idea why anyone calls it a swing state. I'd say the same for NV for Obama... in the bag for O there.
Remember my post for four days ;) It's NOT long odds Romney will win. I have him at better than 50%, and more like 75% after re-weighting as people now identify, and accounting for undecideds historical trend, plus "voter enthusiasm" measures which sharply go against the current president.
The only poll that matters is the vote, but know that many polls are designed to influence public opinion rather than find the actual truth. With some scrutiny, however, it can still be found.