Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[smirk]

I think his model stinks.  And accuracy?  Accuracy is easy.

Hells, accuracy is mostly all I want. And Wang is accurate*. He's accurate on a fairly simple poll median model, and I don't see where Silver's addition of fundamentals is adding anything much more than horse-race volatility. The more assumptions you layer on top of a model, the more room for error you introduce and at some point if you're not careful you cross the line into punditry. Not saying that Silver is there yet, but I do get the impression that he's slowly edging closer (see Trump's Six Stages of Doom). I appreciate that adding the correct fundamentals to a model would help, I'm just not confident Silver has reliable ones. (Full disclosure: I still read FiveThirtyEight and wouldn't have much of a problem relying on it if I had to; I just tend to like PEC's model better. "Silver lost me" was a little hyperbolic.)

Also, curious: are you referring to a particular consensus published somewhere that Sam is outside of? I know there's him and Nate, and Drew Linzer over at Votamatic, and one or two others who have been running data-driven election forecasting models since 2004 or so, but otherwise it's still a relatively new field. I'd be interested in more reading on the subject.

_{*In presidential years, anyway. His 2014 wasn't as good.}

[mainiac]

Also, curious: are you referring to a particular consensus published somewhere that Sam is outside of?

It's just a general trend I have noticed over the years but you can see it yourself if you take a look at the NYTimes comparison on this page: http://www.nytimes.com/interactive/2016/upshot/presidential-polls-forecast.html

Sam Wang's predictions are more confident then NYT (cohn), 538, Cook, Roth and Sabato, the other aggregators and experts.  This comes from the expection that polls will be mean reverting which I think is a fundamentally fallacious assumption.  We have seen elections where the polls did not revert to the candidate who lead for a long time: 1992, 2000, 2004 spring to mind.

I think that there is a case to be made that the 538 model takes things too far by treating polls as basically a random walk biased back towards "fundamentals" but the folks over there are well aware of that.

In fact I would say that it is Sam Wang, not 538 who is adding a layer of assumption into his model and setting him up for a "nobody could have predicted" error.

[Max™]

2016: The year the GM banned exploding dice once and for all.

[mainiac]

Not so much exploding dice as possibly weighted dice.

[Max™]

That would imply Trump had some sort of deliberate advantages, rather than being the only one crazy enough to try to roll a 1% skill check, crit succeed, and get several more critical successes afterwards.

[smjjames]

http://www.politico.com/story/2016/10/gary-johnson-north-korea-229221 Eh? No idea why he'd object to mentioning the leader of NK. Or maybe he objected to the question.

[Reelya]

Johnson actually argued that the President knowing WTF was going on in the world would be dangerous for the US military:

These foreign policy answers have raised serious questions about Johnson's knowledge of basic international facts. The Libertarian nominee though has tried to wave those questions off. In an interview with MSNBC's Andrea Mitchell on Tuesday, the former New Mexico governor argued that a deep foreign policy expertise is a pathway to starting wars.

"You know what? The fact that somebody can dot the I’s and cross the T's on a foreign leader’s geographic location then allows them to put our military in harm’s way," Johnson said.

Basically he's arguing here that if you pretend the rest of the world doesn't exist, America will be safer.

[Sergarr]

Johnson actually argued that the President knowing WTF was going on in the world would be dangerous for the US military:

These foreign policy answers have raised serious questions about Johnson's knowledge of basic international facts. The Libertarian nominee though has tried to wave those questions off. In an interview with MSNBC's Andrea Mitchell on Tuesday, the former New Mexico governor argued that a deep foreign policy expertise is a pathway to starting wars.

"You know what? The fact that somebody can dot the I’s and cross the T's on a foreign leader’s geographic location then allows them to put our military in harm’s way," Johnson said.

Basically he's arguing here that if you pretend the rest of the world doesn't exist, America will be safer.

It's typical libertarianism, actually. They have always had to pretend that the rest of the world, along with its entire history, doesn't exist, in order to justify their inane "government that can fit in a shed" societies.

[Frumple]

I suppose you could make the argument the U.S. is safer if johnson ignores that the rest of the world exists?

[mainiac]

Imagine how safe we would be if Trump didn't know about the rest of the world either.

[scriver]

So yeah. US is literally an Oligarchy.

And the Canada is literally a monarchy.

It is also literally a theocracy!

[smjjames]

Johnson actually argued that the President knowing WTF was going on in the world would be dangerous for the US military:

These foreign policy answers have raised serious questions about Johnson's knowledge of basic international facts. The Libertarian nominee though has tried to wave those questions off. In an interview with MSNBC's Andrea Mitchell on Tuesday, the former New Mexico governor argued that a deep foreign policy expertise is a pathway to starting wars.

"You know what? The fact that somebody can dot the I’s and cross the T's on a foreign leader’s geographic location then allows them to put our military in harm’s way," Johnson said.

Basically he's arguing here that if you pretend the rest of the world doesn't exist, America will be safer.

It's typical libertarianism, actually. They have always had to pretend that the rest of the world, along with its entire history, doesn't exist, in order to justify their inane "government that can fit in a shed" societies.

I've been thinking that they he seems to be a proponent of the 'head in the sand' strategy, which isn't even a real strategy.

[Karnewarrior]

I heard the People's Republic of Ostrich is repealing the military reforms regarding sand-heading due to inefficiency issues when trying not to get eaten by Lions.

[misko27]

I've been thinking that they he seems to be a proponent of the 'head in the sand' strategy, which isn't even a real strategy.

It is now.

[Rolepgeek]

(Usual caveat: confidence in win probability aside, it only matters if folks actually go out and vote. So go vote.)

From what I understand it kinda...doesn't? Because the electoral college is a thing and can instantly cast aside the popular vote like a sack of bricks? Like what happened with George W?

So yeah. US is literally an Oligarchy.

Actually to be frank I'm fairly impressed with how rarely the electoral vote and the popular vote don't line up. Further, it's not a matter of casting aside, it's a matter of the structure of the system. They aren't obligated to vote for the popular vote in the country (otherwise it would just be a matter of who got more people period and then 100% of the electors vote for them), they're obligated to vote for who they said they would vote for, once voted upon themselves. It's entirely possible for three states to vote 60% in favor of one candidate and 40% in favor of the other, and for the other two to vote 100% in favor of the other, and for the first to still win, despite only having 30% of the popular vote. But none of them 'cast aside' the popular vote. It's just multi-tiered first-past-the-post. It's why I find gerrymandering so upsetting.