I don't trust polls that need to adjust the samples.I trust a poll that calls 1000 registered voters and the poll results are the answers that the 1000 voters gave
http://electoralmap.net/analysis.php Did Intrade correctly predict the 2008 Presidential Election? The answer is - without a doubt. The Intrade prediction market was forecasting a sizeable Obama victory for most of the election season, and correctly predicted the results of nearly every state in the Union. Shown below is a graph which uses the state-by-state betting data from the Intrade prediction market. The graph shows the total number of electoral votes Barack Obama was predicted to win from late May until election day. States had to be trading above 55 on the prediction market to count toward the total number of forecasted electoral votes. Barack Obama was predicted to win over 270 Electoral votes (the winning threshold) for most of the election season, except for a few weeks during the late summer. Once the sub-prime lending mess began to unravel on Wall Street, Intrade predicted an overwhelming victory for Senator Obama. Shown below is the same chart for Senator McCain. At no time during the entire election season was John McCain forecasted to win 270 or more electoral votes. Immediately after naming Alaska governor Sarah Palin as his running mate McCain saw a brief surge in his forecasted EV count, which quickly dissipated. Senator McCain consistently found himself in the position of needing to win every single swing state to be able to secure a winning 270 electoral votes. There was only one point in mid-September when McCain was actually forecasted to win more electoral votes than Obama. Other than that, Obama enjoyed a sizable lead for most of the election season. Shown below is a graph of Obama's lead over McCain during the election season.
http://en.wikipedia.org/wiki/Intrade Political contracts Intrade has offered since 2002 the widest range of markets for political events, such as "George W. Bush to win 2004 US Presidential Election". In the 2004 presidential election, the market favorite won the electoral vote in every state. This occurred when, even as late as election day, many pollsters and analysts were predicting a John Kerry victory. In Florida, a number of polls put Kerry ahead in that state, or said the race was too close to call. The betting markets, however, correctly and consistently showed Bush would win Florida comfortably
Your method gives you a very accurate picture of how registered voters feel about the candidate, but is a poor predictor of elections. The reason is, depending upon the election, 30-50% of registered voters don't vote -- and those who don't vote are not proportionally split among the parties. In other words, the non-voters are disproportionate to their percentage of registration. To ignore this fact is to ignore a basic law of mathematics by assuming statistical neutrality where none exists.
In a market like intrade, the incumbent will be ahead until there's a lot of evidence otherwise. The reason for this is that many "early bettors" play the general statistical odds that a sitting president has about a 70% chance of being re-elected. We'll probably be looking at a one or two point spread come October.
On top of that the market is so thin on the bid ask when I look, it is hardly worth placing a bet. So it is easily rigged for the PR value.
New CNN poll released today- Obama +9 http://politicalticker.blogs.cnn.co...r-gap-and-likeability-keep-obama-over-romney/ (CNN) - President Barack Obama holds a nine-point lead over Republican challenger Mitt Romney thanks in part to the perception that the president is more likeable and more in touch with the problems facing women and middle class Americans, according to a new national poll. A CNN/ORC International poll released Monday also indicates a large gender gap that benefits Obama, but the public is divided on which candidate can best jump-start the economy. According to the poll, 52% of registered voters say if the presidential election were held today, they would vote for the president, with 43% saying they would cast a ballot for Romney, the former Massachusetts governor who is making his second bid for the White House. The poll was conducted Friday through Sunday, a few days after former Sen. Rick Santorum of Pennsylvania suspended his bid for the GOP nomination. Even though former House Speaker Newt Gingrich and Rep. Ron Paul of Texas continue their long-shot bids for the nomination, Romney is now generally considered the presumptive nominee. The survey indicates women voters back Obama over Romney by 16 points (55%-39%), virtually unchanged from an 18-point advantage among women for the president in CNN polling last month. The poll was conducted two days after Democratic strategist and CNN contributor Hilary Rosen created a controversy by saying that Ann Romney "never worked a day in her life." "That remark may have little long-term effect on women voters," says CNN Polling Director Keating Holland. "By a two-to-one margin, the women surveyed saw President Obama as more in touch with the problems facing American women today." Half of those questioned say that Obama is more likely to stand up for what he believes, with only 29% saying that about Romney. Nearly half say that Romney is more likely to change his position on the issues for political reasons, just 39% saying the same thing about the president. Obama has double-digit leads over Romney on likeability, honesty, confidence, values, leadership and almost every other characteristic tested, with one important exception. "Obama and Romney are essentially tied on who is more likely to get the economy moving again, and that may provide Romney an opening to chip away at Obama's current overall lead," says Holland. According to the poll, Obama holds a 48%-43% margin over Romney among crucial independent voters. The survey also indicates s generation gap, with all age groups, except those 65 and older, backing Obama. And the poll points to an income divide, with the president holding a 20-point lead over Romney among those earning less than $50,000 per year, while those making more than that figure divided between the two candidates. Obama's likeability and strong performance on personal characteristics helps explain why three-quarters of his supporters questioned say their vote will be a vote for Obama, not a vote against his opponent. By contrast, more than six in ten Romney supporters say their choice will be mostly be a vote against Obama. "That's a significantly higher level of anti-incumbent voting than polls found in previous years. In 2004, for example, 55% of Democratic nominee John Kerry's supporters said their choice of Kerry was really a vote against President George W. Bush. The question for 2012 is whether Romney has to provide his supporters with more reasons to vote for him in order to win, or whether a negative anti-Obama message is enough, given that historically high level of anti-incumbent voting," adds Holland. Two-thirds say they have made up their minds, with just 29% saying they could change their minds between now and November. As for handicapping the election, right now Americans don't think Romney has a good chance of winning the White House. Only 35% said Romney will be victorious in November. "That is certainly not a prediction of what will happen, of course, but it is worth noting that in the last four presidential elections, the public was able to correctly pick the winner in polls taken in the spring or early summer," adds Holland.
Well, maybe. How do you adjust for likely voters this far out? I doubt they'd tell you their exact methodology, but I'm sure whatever algorithm they use has to almost certainly have a smaller margin of error as you get closer to Election Day. So, a pollster using "likely" may not necessarily be more accurate now. After Labor Day, yeah.
when did they close the betting? By the end of the game I have gotten every super bowl bet correct as well. When I was trading full time I made most of my living because I traded great in the afternoon... I used T/A to make very good bets about where we would close.