Hacker News Reader

Comment by untog | Hacker News Reader

You could just as easily counter and ask why anyone would expect a Republican victory when Democrats did so well in the 2018 midterms.

Previous result are only predictive to a point.

View on HN

The party that is "behind" always does well in the mid terms, this is a very well documentes phenomenon. In fact, I think the Dems did much worse in 2018 than expected

That’s not true. The 538 model was very close to the actual outcome in 2018.

By 'expected' I don't mean the models where wrong, I mean the country is on fire and the Democratic party could/should have run away with the election. Instead they lost seats in the senate and did not gain back all the ground they lost in the state governments.

For comparison, the Republicans did very well in the 2014 mid term, convincingly taking over both houses of congress.

The democratic party had every advantage for a convincing win in 2018 and they instead had a very poor showing. The party is not good at elections and I don't see a reason why we expect this election to be different.

"Democrats won the House with the largest midterms margin of all time" [1]

Dems did well in 2018, the issue was that the 1/3 of senators that happened to be up that year were in very bad states for dems geographically.

[1] https://www.axios.com/2018-midterm-elections-democrats-won-h...

The senate stayed red because there were basically zero unsafe red seats open and lots of unsafe blue seats open. If, by coincidence, zero republicans were up for reelection in the senate it would be impossible for the dems to gain seats even in the biggest wave possible. Any analysis which ignores this issue is just foolish.

They did well in the House, but lost seats in the Senate. Midterm elections are an imperfect signal, like you're suggesting, but it's not true that the democrats won a landslide or something in 2018.

The Senate map for a given year can tilt towards one party or another; 2018 was a structurally bad year for the Senate Dems.

https://fivethirtyeight.com/features/democrats-horrible-2018...

Right, this is a good data point in favor of the argument that it's unwise to extrapolate too much from midterm races for the general.

No. It is a good datapoint that one should consider the seats in play when making claims about the senate. The map this time is much less unfavorable for the Democrats.

Reply on news.ycombinator.com