Quote from jem:
Since we are in the last week... I want to prepare our post mortem of this thread which we will do next week.
This thread turned into a thread about the polls... mostly.
This is my view of issues to review next week, I welcome yours. This is obviously a bit biased to my view of things.
Some of the issues we pointed out about polls..
a. the polls were slanted using 7 to 20 points skews to the Ds.
-anything larger than the 2008 template was criminal.
- we argued 2008 template would never happen
- we thought a pollster was responsible if he or she was D plus 3 or 4 or less
- we suggested 2010 was a good template... as well.
- we held and still hold suspicions more Rs than D will turnout
- we predicted that most pollsters would change their ratios to more responsible ones by election day.
b. we also suggest that Nate Silver is a bit of a clown... we suggest to model polls correctly you should adjust the polls to a reasonable election turnout template and then make forecasts.
modeling slanted polls is a fools task. its like doing technical analysis on poorly constructed sentiment readings in the market.
you could get lucky because the results are binary, but a good result is not proof of skill.
c. finally we have been saying that with Romney earns 10 to 20 points more independents on election day... he will certainly win.
---