National vs State : Is the following reasonable?

[Harry Hayfield]

A national poll comes out (say for the sake of argument McCain 48% Obama 48%). This represents a swing of 1.5% to the Democrats since the 2004 Presidential Election, therefore we apply a 1.5% national swing to the Dems since 2004.

That would give McCain 250 ECV's to Obama's 289 ECV's, but of those there woud be several that I would term "tossups" which for the purpose of the 2008 election, I deem any state which has a lead of less than 3% (the standard margin of error). In this example those states would be: WI, IA, NM, OH, NV, CO and FL.

In a state that is projected as a tossup, I then look at the last 5 polls in that state and then average the numbers. If those averages produce a lead for a candidate that is greater than the number of uncommitted, then that state goes to that candidate.

[True Democrat]

No.  This is not a European parliamentary election (though it doesn't really work there either).

[Queen Mum Inks.LWC]

No.  The Electoral College screws with math too much to make national polls worth anything unless somebody's leading by a pretty decent margin.

[Stranger in a strange land]

No.  The Electoral College screws with math too much to make national polls worth anything unless somebody's leading by a pretty decent margin.

pretty much yeah. Also the individual states are very different and often vote for very different reasons.

[RJ]

I don't know. Maybe it's not for the reasons he(Harry) says, but when the national polls were in a dead heat not long ago, the electoral math(including leaning states) favored Obama by a pretty decent margin.