For the first time, 538 are forecasting that the Republicans will pick up more than 50 seats in the mid terms next week.
That is only the mid point of their forecasts. One thing which makes them a good site is they give probabilities for different scenarios. They say there is still a 20% chance the Republicans will win fewer than 39 seats and the Democrats retain a majority.
However there is also a 30% chance the Republicans win 60 or more seats – beating the 54 seat pick-up of 1994.
15 Democrat-held seats are forecast to be at least 90% likely to go Republican, and 1 Republican-held seat to go Democrat for a net gain of 14.
A further 10 seats are forecast to be at least 80% likely to go Republican, and 1 the other way, for a cumulative net gain of 23.
At 70% you add on another another 14 seats for a cumulative gain of 37. And another 9 at over 60% for a cumulative gain of 46. So for the Democrats to keep the majority, at least seven seats deemed 60% or more likely to fall, have to be retained.
At 50% you add 8 more on to take the total to 54. That is not a 50% chance of winning 54 net seats but 54 net seats which have a 50% or greater chance of swapping hands.
If in the final week, it gets even better for the Republicans, how big could the landslide get? Those above 40% are seven seats taking it to 61, a further six above 30% to 67 and a further 9 over 20% to 76. The remaining seats are not thought to be contestable.
This is the first time 538 has forecast House as well as Senate races. It will be interested to see how well they do. I’ll take another look at their model on election eve.