2018 Race Forecasted Actual
US Senate GOP +1 GOP +2
Governorships DEMS +8 DEMS +6
US House DEMS +30 DEMS +37


In the Senate the forecast was entirely accurate with the exception of Florida, which was considered a tossup slightly favoring the Democrat. The election was indeed fairly close, but the slight favor ended up in the Republican's corner. Seats flipped by the GOP were Florida, Indiana, Missouri, and North Dakota. The Dems flipped Arizona and Nevada. Net gain for the GOP was 2 seats.

The Governors' forecast wasn't quite as good. The end result was close to expectations (net 8 mansion pick-ups forecast for the Dems, with net 6 realized) but the state-by-state accuracy left something to be desired. The forecast was wrong in Ohio, Iowa, and Florida. All three of those states had polling data favoring the Democrat, and all three went to the Republican. The toss-up in Kansas went to the Democrat.

The House forecast for the Dems to pick up 30 seats and take control of the chamber hit pretty close to the mark, with an actual pickup of ~37 seats. Splitting the tossup seats 50/50 had remarkable accuracy for the last few elections, given the incredible simplicity of the analysis. In any case, 37 seats fell well short of our "blue wave" what-if scenario of +49 Dem seats.