richardhutnik on 03 October 2012
| binary solo said: http://elections.huffingtonpost.com/2012/romney-vs-obama-electoral-map "Our snapshot of where the presidential race stands is based on hundreds of state-wide and national opinion polls — filtered through a poll-tracking model — and updated throughout the day. " Suggests Obama is a shoe-in with 270 ECV already locked up in states strongly supporting Obama. Sure Huffpost is a bastion of leftwing socialist Obama loving bloggers. But it's collating hundreds of polls on an ongoing basis. So it's not just one polling organisation's results. I don't think that poll of polls can be so badly wrong that the real pictyure is Romney with a 5 point margin over Obama. I don't see Obama winning by anything like teh ECV count this map suggests, but I think things are in Obama's favour right now. Also the Iowa Electronic Markets are showing massive investment in Obama "shares" right now, and that's people putting actual money on the outcome. http://iemweb.biz.uiowa.edu/graphs/graph_Pres12_VS.cfm http://iemweb.biz.uiowa.edu/graphs/graph_Pres12_WTA.cfm This market correctly predicted (as nearly everyone in the world did) the 2008 result, but it was incorrect about the 2004 result. It was actually correct about the 2000 result (i.e. that Gore had more votes nationally than Bush), but of course number of votes nationally doesn't choose the president, even in years when the Supreme Court stays out of it. |
Margin of error is for a single poll. When you average together polls in the same time period, the margin of error should decrease. Assuming that the polls are decent, then if about all of them show someone leading, the odds are real strong they areleading. The prediction markets also do a good job predicting things.
