Election models megathread

[Minnesota Mike]

1st update and Republicans are up to a 55% chance to control the Senate. What's changed in the last 24 hours?

[Minnesota Mike]

1st update and Republicans are up to a 55% chance to control the Senate. What's changed in the last 24 hours?

Dem internals came out showing them tenously leading in NV, which caused 538 to improve their % odds for Laxalt.

Seems like a big jump for one poll, especially one that had CCM up. They must be factoring in one hell of a house effect.

BTW just because I have to be that guy the Change Research polls are not internals, that would indicate they were paid for by the campaign or party, they were paid for by an outside Dem leaning group. Small but real difference.

[Minnesota Mike]

Which model do you prefer 538 Classic or Deluxe? Personally I like the Classic model, factoring in "experts" opinions kind of defeats the purpose of having an empirical model.

[Minnesota Mike]

There are a lot of highly questionable probabilities and predicted vote shares in individual races, but it seems like the model is erring roughly equally in both directions, so the national projections seem relatively reasonable.

I know any model will spit out some questionable results on some races but some here are real head scratchers. For example in MD-06 the Classic model has David Trone winning only 20% of the time and the average margin R+6.8%. Really? In a Biden +9.8% district with an incumbent who has been willing to self fund as much as it takes?

[Minnesota Mike]

You also forget Republicans ran bad candidates in WV and OH (and completely gave up on the latter). With Jenkins in WV and a competent Republican campaign in OH, they could definitely have won both races.

Have Republicans ever lost with a good candidate? It seems like every time a Republican loses the loss is blamed on a crappy candidate instead of on crappy ideas.

[Minnesota Mike]

Is there a possibility to ban Olowakandi at least from the Congressional & Governor Topics of Talk Elections until the 2022 Election is over!

I have not seen anything that violates the ToS.  If you find a user's signal-to-noise ratio too low, you can always Ignore him; the button is just up there on the right.

Although you can still see an ignored persons posts if someone else quotes them as I found out with this post.   I make liberal use of the ignore feature, I would guess about a quarter of posts are from people I ignore. I don't mute people who I disagree with, I just mute people who who add nothing to the conversation.  FWIW I think my record is 10 posts in a row in I thread from people I have on ignore.

[Minnesota Mike]

The 538 Deluxe model is down to Republicans winning the Senate 51% of the time. Classic model (which I prefer) has Democrats winning the Senate 61% of the time. A lot of coin flip seats but at least in the Senate not nearly as bleak for Democrats as the doomsayers would have you believe.

[Minnesota Mike]

The problem with models. I don't expect Democrats to win MN-01 or NE-01 but c'mon.

538 Deluxe

MN-01: Finstad 98%, projected pop vote 57.1 - 39.1.

NE-01: Flood 98%, projected pop vote 58.9 - 41.1.

[Minnesota Mike]

I realize Alaska's top 4 system does not fit smoothly into 538's model but shouldn't the Peltola at least be listed as a candidate instead of just Palin and Begich?

[Minnesota Mike]

Ladies and Gentlemen, FiveThirtyEight's Deluxe model has Republican chances for the House at 69% now. Lowest on record this cycle.

nice

Nice