polls are NOT reliable. considering you could run the exact same poll say 20 different times with different groups of people and it could come out different each time. i sure am not going to go by some stupid poll to make my own decisions in life.

Good advice. Although polling has become far more accurate in recent years, there are always upsets that seem to undermine traditional faith in political research methods. I certainly don't put a lot of faith in polls. They are a good indicator of trends, or shifts in public perception, nothing more. But there is plenty of data that can be garnered from it nevertheless. I will illustrate how you can use the polls as a predictive device.

In terms of the current Presidential election, my general rule is that the underdog usually gets about 80% of the undecided vote going into Election Day. Those are the voters that generally decide tight races. 6% of the electorate is currently undecided right now, and John McCain (the current underdog) is polling in the low to mid 40's nationally. If 80% of that 6% breaks for McCain, you have a fairly tight race nationally. I expect McCain to get about 48% of the popular vote. Of course, the Electoral College is what decides elections.

If Obama is polling above 49% in any given state, you can put it in his column. If McCain is polling in the high 40's in any given state, he has a better than even shot at winning that state. States that are tied in the polls and where Obama is polling at lower than 47% will likely go for McCain. Indiana and Missouri are two states that come to mind here. Given these parameters, I believe that Obama will win by at least 310 electoral votes. It all depends on turnout.

That takes us to another question; we really don't know who will be turning out this election. Pollsters are generally using modeling based on turnout in the last presidential election, but that really doesn't tell us anything. I expect that turnout might be as high as 150 million voters, that's up from about 120 million in 2004! That's a huge jump in new voters! I believe that there is a chance that the polls might be underestimating support for Barack Obama. Here in Florida where I live, the Democrats have registered over 600,000 new black voters this year. That's a parameter that the polls are not taking into account. Many of those new voters are turning out for Obama. We know that in FL, Democrats are turning out in droves for early voting, by a 2-1 margin over the Republicans according to some early estimates. They have excitement on their side, their core voters are going to the polls, along with people that usually don't vote. And this is the same story all over the country.

This in short is a different election; different dynamics are in play, so I wouldn't put too much stock on what the polls are saying. Now it could be tighter than we think, but it might be that Obama is poised to win a lot of those marginal states. The classical rules of polling DO NOT apply here! I wouldn't put too much faith in the polls.