The model is a combination of individual statewide polling averages and a demographic model based on historical voting patterns adjusted to current nationwide polling. Uncertainty is assumed to double when projecting every eight weeks into the future, and the model currently assumes the probability of a major third party candidacy is zero. Individual state probabilities (including congressional districts for NE and ME) are then included in a stochastic model (N = 10,000) for the Electoral College.
The model will be updated once a week, because that’s about as quickly as the country can register any event, and a further subdivision would better represent fluctuations in polling more than a measurable shift in the thinking of the electorate.
If the election were held tomorrow, Hillary Clinton would have a 99.2% probably of winning, and would likely crush Donald Trump in the Electoral College. Of course the election is not tomorrow, yet the model still projects Clinton to have a 93% chance of winning a majority of Electoral Votes this November. As discussed earlier, Trump performs strongly in the Rust Belt and Midwest, but is considerably behind Romney’s level of support in the Deep South and western states. Far behind in the once reliably Republican Virginia.
This likely overstates Clinton’s lead. Republicans are still getting used to the bitter taste of Trump, and while some may never, most will eventually come around to voting for the Donald. As they do, the race will predictably tighten. Even late in the race it is likely that Trump will overperform expectations during the debates, since the expectation is that he’ll brag about his penis again. He won’t. He’ll demonstrate a below average understanding of the issues, and this will earn him considerable praise.
So here we are, mid-May with a long summer and brutal fall ahead. All I can say is that we’ll be there with you.
The Seneca Cup
Every presidential election the Mugwumps submit their individual predictions for the electoral map a year before the election takes place. Whoever is closer to the actual outcome wins the Seneca Cup. Each Electoral Vote predicted correctly earns an Electoral Point (double points for third party candidates). Additionally, correctly predicting entire geographical regions earns bonus points equivalent to the number of states that comprise the region.
Here are current standings based on the outcome predicted at present by the election model: