More Political Games: On Polls and the Two Parties

Unfortunately, this will be another post on the game theory in politics.
First up, the Huffington Post, a more liberal blog, decided to look at the methodology of the recent polls, and found the balance between liberal, conservatives and independents had changed.
While the Huffington Post took the angle that pollsters designed changed their methodology to make the race seem close. I'm doubting that analysis, and there's no way to confirm that bit of speculation at all, unless somebody within polling organizations says something stupid publicly.
Instead, this actually is another reason why YOU DO NOT BELIEVE THESE POLLS. All presidental polling has been historically innacurate. In this case, it'll be darned near impossible to tell the make up of Democrat, Republican and Independant voters in November, especially since it's only been a week after the last political convention.

Now for the bit of game theory that is completely misunderstood, third parties. It seems that Ron Paul is calling for voters not to vote for Obama or McCain, but any third party candidate. In reality, this isn't as smart as it seems.
This is taught in politics 101, that rarely and under special conditions that third parties even stand a chance, mostly when there's a political vacuum. Put it this way, there's no practical way for a third party to be in competition and not take away votes from someone else. The best analagy is if an Alaska crab fisherman set his cages down in the same location as another boat's; they only split the local crab population. Nether fisherman gets a good catch.
That isn't to say that third parties are usless in campaigns. They often can bring up topics that the two dominate parties avoid. But if you want to actually change how government is runned or how the parties act, it's better to change from within.