Senate

35 of the 100 seats are up in the Senate. There would be 33, but MN and MS have a special election. The 33 others are all Class One Seats.

I used 270towin.com to generate these predictions, and I have four tiers of these: High Confidence, Competitive, Very Close, and Full Predictions.

There are four tiers of each color: Dark (Safe, High Confidence and Margins), Medium (Likely, Decent Confidence and Solid Margins), Light (Leans, Some Confidence and Competitive Margins), and Low Opacity (Tilt, Very Low Confidence and Margins).

Editable maps will be provided for all maps except for full predictions.

Final Predictions

Very High Confidence:

Competitive:

Very Close:

Full Predictions:

Results:

Color Key

0% - 4% = Low Opacity

4% - 8% = Light

8% - 12.5% = Medium

>12.5% = Solid

Out of all the predictions, this is the one I got right on the nose. My initial model from August 28, which was done from gut feeling got all 35 states right, and this one only missed Arizona and Indiana. Surprisingly, this was less numeric of an approach compared to my House predictions, but not less than the governors. Texas was a bit closer than expected, but not much closer than the 2016 partisanship would suggest.