Here's another tool you might like to try and play with for modelling the general election outcome. Nate Silver you'll probably recall was the guru who got all of the swing states right in 2012. The BBC brought him over to the UK about 12 months ago, where he failed. His reputation in the US however remains in tact, albeit its increasingly damaged on this cycle due to his constant under estimating of Trump. So far as I can see this is down to two factors
1: His models are quantitatively cold and struggle to capture things like emotion. They assume rationality (which they have to) but something else that is much harder to capture is seemingly going on this cycle
2: The 'poll plus' model in particular has correlated endorsements as being historically the best predictor, and he over relies on this as a result. Trump has destroyed it basically. The problem I suspect is that a lot of endorsements occur when a result is becoming apparent and the endorsee wants to end up on the winning side. It might be questionable therefore as to whether they form opinion, or reflect it
Anyway, back to the tool
The historic (reset button) is the 2012 election. What you then need to do is slide the five factors according to how you think turnout and vote share will alter. As you do states change colour
I'm not totally convinced by it, but then to build an accessible model with complexity would almost be impossible. The three observations I'd make are;
1: The swings are uniform across the US, one suspects that in certain states it will be more aggressive than others, particularly in the rust belt, Florida, and hispanic states of NM, CO & NV
2: In a race that likely features a sexist male against America's first female candidate, it seems odd that gender isn't factored and assumed to be uniform across the socio/ economic strata
3: The lumping of Hispanic and Latino voters into one homogenous group could be errorneous. The Mexican voters of NV, NM and CO will trend democrat and behave differently to the Cuban voters of FL who trend republican. As someone pointed out to me yesterday, the Floridan Cubans are the most disproportionately influential group in the whole of the US. It's possibly this which is making me think Hillary has to put Julian Castro on the ticket after all
http://projects.fivethirtyeight.com/2016-swing-the-election/