We missed five states (assuming Clinton wins NH and MN): WI, PA, MI were all in the mid-90's. These were big misses. FL and NC were tight predictions that went for Trump. This is, of course, little solace to those who relied on these predictions. The headline prediction was 89 percent for Clinton to win and 11 percent for Trump to win; Trump won. Markets are anchored and buoyed by publicly available polling. And, publicly available polling made Wisconsin out-of-reach for Trump: +6.5 in both Pollster and RealClearPolitics. We will learn a lot over the coming days how polling did so badly. Polling is a mix of estimating the voter population and the sentiment of the voter population. I assume both were off: too few lower educated white voters and educated white women going more heavily for Trump.
If you are still interested, there will be more on this page about the evolution of polling which will take place over the next few years. I look forward to relying on my own data more and more moving forward.