I believe I posted back in 2008 or 2012 that I have a website with presidential election predictions (prespredict.com), and it is now up and running for the 2016 election. It uses daily state poll data and calculates the probability that each candidate will win.
The methodology has been applied to the last three elections (2004, 2008, and 2012), and in all three past elections the calculation correctly predicted the winner. It also had a pretty good estimate of the number of electoral votes that each candidate ended up getting.
Of course, it’s still early and there are several caveats, but it is being updated daily, so you can check back periodically.
One issue that is nagging me a bit is that it is currently predicting Hillary’s chances of winning as extremely high (98.4%), while 538 has her at 87.8%. I believe my approach is a bit more simple than 538’s, but both seem to agree on the trajectory of the vote after the DNC, which is that Trump’s odds are rapidly diminishing.
For reference, the site also has the plots for the last three elections, so you can compare with this election. Previous elections seemed to show more swings and variation that this election.
Any comments?