Election Simulators: Possibly Accurate OR Premature Joculation?

So I'm a little paranoid about presidential elections; 2000 and 2004 are my only previous electoral experiences - and both were dodgy. I've learned not to read too heavily in to polls (especially after learning in 2000 that the popular vote doesn't necessarily matter), and I've seen that the electoral votes aren't so easy to predict. Nonetheless, I like to know what the polls are saying and what the experts think, so I still keep myself informed on the current data.

I found an electoral simulator that takes regularly updated state-by-state poll results and turns these results into probabilities. The website explains it in detail. Swing states are the ones that could alter the outcome in favor of one candidate over the other, since each candidate is as likely as the other to win those electoral votes.

The site keeps track of the last 1000 simulations run by visitors to the website and comes up with rudimentary statistics (mean, median, mode) for the outcomes (McCain wins, Obama wins, Tie). I know it's not a prediction of the outcome, it's simply fundamental statistics that doesn't take everything into account when simulating probable outcomes - but these statistics are too interesting to ignore.

As of October 11 at 4:00 pm:
Election Winner:
Obama - 99.9%
McCain - 0.1%
Tie - 0.0%

Median Electoral Votes:
Obama - 334
McCain - 204

Average Electoral Votes:
Obama - 333
McCain - 205

Range of Electoral Votes:
Obama - 249-390
McCain - 148-289

Another interesting computation: based on the most recent undecided voter data, they compute the possible allocation of their votes between the 2 candidates, compare it to the existing "safe" electoral votes, and then determine the combination of states necessary for each candidate to reach 270 electoral votes. Assuming the current site statistics (swing states: Florida, Ohio, North Carolina, Virginia, Indiana, Missouri, Colorado, Nevada - totaling 111 electoral votes "up for grabs"), McCain has to get all of the remaining swing states' electoral votes to win. I chose to compare info from other sources with more recent data (changing the swing states around a a bit), which changed the math a little, proving further that although this data is interesting, it's not necessarily indicative of the actual election. Although the numbers are strongly in Obama's favor, it would be unwise to celebrate victory before it happens (ergo premature joculation... Cute, isn't it?).

Go play around and have fun on the website:
2008 Presidential Election Simulator