Politics Indiana: Trump in a romp, Sanders ahead of Clinton, CNN reports Cruz dropping out, Kasich too

Kasich would help him majorly in the North East. I could see a Trump/Kasich ticket winning a couple states, like New Hampshire and Pennsylvania.

 

Denny bringing the heat.

#barfogotburned

 

Thre would be a huge market for insurance that covers elective procedures. Everything else is a common IMHO.

No vote Bernie. Bernie Wan Kenobi, he's our only hope!

 

538/Nate Silver got it wrong because of an over-reliance on the "party decides" political theory of primaries, that even with no "smoke-filled back rooms," the party still wields enough power, via control over party apparatus and donors as well as the influence of endorsements, to steer the nomination to the guy they want. There's been a lot of evidence for this in the past, but Trump blew through that, largely because of how much free media he generated. You don't really need party apparatus or donors if you generate ~$2 billion of media time, at no cost to you. As for endorsements, well...in probably the least establishment-friendly primary of all-time, that proved to work in Trump's favor too.

538 has two projection models, Polls Only and Polls Plus. Polls Only just goes by the polling data, averaged and weighted for recency and house bias. Polls Plus adds in external factors, things like endorsements and organizational strength...essentially things that mostly (but not only) track the establishment's thumb on the scales.

During the primary, their Polls Only projection model was pretty much always dead on. Their Polls Plus model tended to underrate Trump for the reasons given above, though it actually rarely got states wrong--when it was wrong, it was mostly wrong on how wide the polling spread would be, underrating his delegate gains. Even then, though, their model weren't off by a lot (I watched their models pretty consistently, since this primary has been fascinating)--the 538 writers were just skeptical of their own results because Trump was so unlike past candidates and it seemed somewhat unthinkable, for empirical reasons based in past races, that the party wouldn't unify and drive Trump out.

There is no "party decides" theory working against Trump in the general election, though, and no particular reason to doubt the polling data. Doubting the polling data, after all, was what got 538 and Nate Silver, the people you're railing against, in trouble.

 

What do you disagree with? I simply pointed out why they got it wrong...I didn't try to suggest they didn't get it wrong.

Knowing why something was gotten wrong is useful, especially when you say "THEY GOT TRUMP WRONG CLEARLY EVERYONE'S GETTING TRUMP WRONG AGAIN."

I mean, by your logic, Trump will rule America as a king for the rest of eternity. Constitutional limits? Human life spans? Please...that's "expert talk" and experts got Trump wrong.

 

Right...so he got it wrong. However, that was in defiance of the polling numbers. His site had the polling right, and his site's models generally correctly called the states and even the vote shares (in the case of Polls Only). So the moral of the story is, trust the data. Instead, you're making the same mistake Nate Silver did, in ignoring it.

To wit:

Just as Nate Silver assumed that Marco Rubio was going to erase Trump's lead in the polls. Assuming someone is going to erase polling leads tends to lead to being wrong a lot. It's surprising that you've decided to follow Silver's approach in discounting polling leads! It almost sounded like you were criticizing him before, but it turns out you wanted to emulate him.

 

Yeah, but that reasoning is based on an incorrect premise. The polls have pegged Trump pretty perfectly all along. If Trump had been trailing in the polls all along, you'd have a point.

Your reasoning is really, "The polls don't say what I want now, so I'll pretend they've been wrong all along."