She didn't overperform in red states per se, she overperformed in increasingly diverse or very liberal states. States where she bettered Obama's 2012 Margin - California, Virginia, Arizona, Texas, Georgia, DC, Massachusetts, Washington, Kansas (?) (and Idaho and Utah, but that's Mormon related). Unfortunately it was all too modest in comparison to the wave in the midwest.
The Trump Presidency | Biden Inaugurated
- Thread starter Raoul
- Start date
- Status
Farage is a cnut of the highest order. It's even more infuriating because he built his reputation bashing foreigners in the UK, but he can't see the hypocrisy in him meddling in the USA's business. Another self important, supercilious wanker.
That laugh makes me feel violent like every other time I see it.
Well, yeah. I get the feeling you've not got much of a basis in science. The scientific method behind the building of a model, especially a model relying on incomplete and flawed data which produces a probabilistic results, is not completely invalidated by getting it "wrong".
Wong's model was clearly a bad model, what with it giving Trump basically no chance even in Clinton's worst days. If I'm being cynical, I reckon he knew it was a bad model and was gambling on looking very smart when Clinton won by building a model which massively overrated the favourites chances. He'd be the man who knew with 99% certainty that Clinton was going to win, the new polling genius.
Whereas Silver's model didn't actually do that badly. 30% chance for Trump definitely gave him a chance to win and really given the polling there was no reasonable way to model it which made Trump a favourite. The polls are an imperfect way to judge the state of an election but as good as we have. All you can do is analyze what they're saying, undeniably Clinton was leading in the polls, and then try to judge the uncertainty accurately. Silver did that, gave trump a good chance, while Wong didn't. It won't be right every election.
I commented on election night that the feeling in the pit of my stomach was the same as I had when watching the film Funny Games. Both very unsettling experiences.
I have a full appreciation for stats, probabilities, and simulations. My overarching point is that literally every probabilistic model got it wrong, and were shown up by a qualitative punt from some random history professor.
Bet you can't wait for trump to announce Katrina Pierson and Ann Coulter as part of his cabinet
How on earth a probabilistic model like 538 'get it wrong'? That doesn't make any sense as it's impossible to know.
They gave Trump a 30% chance on the day of the election - that could well be spot on. Without being able to rewind time and let it play out over and over, it's impossible to know if Trump had a 70% chance, a 50% chance, a 30% chance etc.
What do you think the probability was? (That's rhetorical)
Bet you can't wait for trump to announce Katrina Pierson and Ann Coulter as part of his cabinet
That's hardly shocking. He seemed like a fish out of water at CNN. I'd imagine he will have some sort of assistant Chief of Staff type role at the White House. If Trump has any brains, he will kick Pierson to the curb, pretty much as he did for the final 3 months of the campaign.
The overall percentage could've been plausible, but they literally got every critical swing state wrong as well.
They gave probabilities for the swing states as well! They explicitly say it's a probabilistic model, not a prediction.
I hadn't even got round to considering this yet. Is he going to be as teflon as President as he was as a candidate? If yes, dear lord.
You say that like it's a nothing role.
Dude, seriously, you're being ridiculously optimistic in your reading of the landscape. I'm not sure if it's an ostrich thing or what, but it's ok to admit that things are utterly fecked. It's not like people don't already know it.
Bet you can't wait for trump to announce Katrina Pierson and Ann Coulter as part of his cabinet
No, I can't wait for Ann Coulter to turn on her beloved Donald when it becomes clear he can't do half of what he said he was going to do regarding immigration. As for Katrina, I've heard that the White House needs maids as many don't wish to work under the new occupants.
If Trump is reading this, I am available and willing to work with him.
I get that bit, but given Silver's previous 50/50 and 49/50 results in 08 and 12, he clearly got something horribly wrong. As in, his model was not able to incorporate concepts like silent Trump voters, early voting results, and demographics that were supposed to help Hillary, but wound up helping Trump (women and latinos). In Silver's pollster ratings, the only pollster that pretty much got most states bang on (The Trafalgar Group) was given a C rating, which speaks volumes. Nate simply didn't have the variables to pick up an assymetrical, Trump style candidate.
Last edited:
This is my biggest fear mate, it's not what Donald is capable of passing but what the blow back will be from those who voted for him when he can't deliver his promises.
I was thinking about something. Am I wrong in thinking that people made a mistake by looking at Latinos as a whole instead of taking into account that Trump mainly targetted mexicans, which means that for example Florida and its Cuban population got badly interpreted?
Chief of Staff is an important role. Deputy and Assistant jobs have to work under the CoS. If its Priebus then Lewandowski's influence will be severely curtailed.
Yeah, I can understand that, and that fear should be as real as the moron losing his temper with another world leader because they criticise him and it ending up being a massive global incident. Or like his completely moronic and stupid comments regarding not knowing what the USA gets from it's various strategic military bases and him making a total arse out of himself and putting somewhere like South Korea in imminent danger. Or him ordering things that will amount to serious war crimes resulting in him being tried for crimes against humanity or breaking international laws. All are very possible simply because he has already said them all at some point AND all show not only his lack of intelligence but also his complete lack of understanding how anything actually really works. If you add it all up he does come across as probably the most inept and woefully uninformed President of all time. I think only Sarah Palin and Michele Bachman come close to sounding so out of their depth as he does, and funnily (or scarily) enough, he will probably have both of those on his staff. Either way he upsets his supporters, or he upsets the people who didn't vote for him even more, or he upsets them all and other countries too. It's difficult to see an avenue for him where he actually doesn't upset anyone at all. The only question is going to be how much and at what cost?
This is all 20/20 hindsighting but I assume the Cubans in Florida didn't take the normalization of relations with Cuba too nicely.
You literally don't know what you're talking about on this one I'm afraid. I followed the various models throughout and 538's ended up being by far the best match to reality in the end. Their methodology turned out to be sound.
They gave Trump a 30% chance - which means he would win around 1 election for every 2 Clinton wins - and he won. That's completely consistent with model. Using the same polls, some of the alternative modellers had Clinton with a 85% or even 99% chance, which turned out to be ludicrous. Nate's model looked at the fact there were 12% undecided voters (compared to 3% undecided in 2012) and introduced lots of corresponding uncertainty into their model to account for it.
In the final weeks of the election, most of 538's editorial coverage was on how Trump had a really strong chance of winning and highlighting that the rest of the media seemed to be crazily working on the basis that Clinton had it sewn up. Since the election all the other modellers have been complimenting 538 on getting it right... now you're here trying to claim they got it wrong.
No wonder you were ridiculously over confident going in to the election if you didn't think Trump had a real good chance of winning, not far behind Clinton's. 538 was saying this was going to be an extremely close election - almost everyone else said Clinton had it more or less in the bag.
And it seems to me anyone claiming with any certainty that Trump was going to win was equally as wrong as those claiming Clinton would win. It was extremely close, so anyone claiming much certainty about anything was way off.
I still don't understand why you think the odds were wrong just because the slight underdog won? That happens all the time with probabilities. It's literally the whole point of them.
Well Mexicans do make up a majority of Latinos in the US. But what the likes of Silver and others weren't able to pick up is that Mexican Americans didn't vote as a homogenous block. More than expected actually voted for Trump, which was completely against the prevailing narrative that they would crush him for calling Mexicans rapists and wanting to deport and build a wall etc. Also, Trump was helped by a decent Cuban turnout in South Florida. When you combine all that with Hillary's problems, it was enough for him to take FL.
Yeah, I heard about that and it probably played a part in it, in fact there is probably a lot of factors that we kind of overlooked.
Given that Trump won against all expectations, the best you can say for Silver is that his forecast, by giving her a 70% chance of winning, got it least wrong. Obviously Wang's ridiculous forecast obscures the fact that Silver wasn't able to get his percentages right in Michigan, Wisconsin, and Pennsylvania. If you have followed previous elections where Silver was involved, you'd know that this one was abysmal for him, which is not coincidentally why he's been explaining himself for the past 72 hours.
They have always been Republican leaning, so this wasn't much of a surprise. The Cuban generations are split about reestablishing relations with Castro. The older Mariel boatlift generation are against, their kids are either indifferent or for.
I think one of the biggest questions at the moment should be what is going to happen to the Trump business? I remember seeing an interview with Donald Jr where he got quite flustered, but in it, it was absolutely clear he didn't understand the problems of the kids running the business and the conflict of interests that would arise from that. I think ultimately this could be something that could potentially bring an end to the Presidency because I can't see him being able to keep clear of his business.
Its going to be a huge fight. The media are going to be pushing the blind trust thing, whereas Trump is going to push that his kids run the business (as if that wouldn't be a galactic conflict of interest).
It doesn't seem like a good point though. Saying they're "wrong" when they produce a probabilistic answer seems sort of perverse for one. Sure, maybe it would be splitting hairs to say Wong wasn't wrong but I think he was basically media whoring (Most probably what your history professor is doing). Silver gave it roughly 1 in 3, so to call that wrong seems to be just completely misunderstanding what probability is.
Then you have the fact that every model predicted Clinton to win because the polls showed Clinton ahead. The polls were wrong, undecideds could have had a big effect, and they off in the same direction in a key area, similar voter turnout predictions being wrong in the midwest.
The professor is very clearly chancing his arm to get some media attention and he's even managed to get a book. Retrofitting a model to a small set of past results, the last 7 elections, with 13 true/false statements has no predictive power. That doesn't show anyone up especially as there are only 2 reasonable options. Anyone in this thread could do what he did in an afternoon: you make a Venn diagram of the winners and losers of those elections as well as the candidate you want to predict to win, populate it with facts and make a list from the section which only contains the winner for your model criteria. Notice the part where you can pick who you want your model to choose as the winner hence it having no predictive power. Similar to how there is always some bellwether count which has chosen the last 10/20 winners, then you find out after the election they picked the loser this time.
There was no way to for any poll aggregation model to show Trump as favourite given that almost all the polls had him trailing. That's just a fact. Silver isn't a pollster, he builds a model to interpret the polls and underlying uncertainties.
My hunch is that Trump actually wasn't the favourite. If you could wind the clock back a week and run it 100 times, I think Clinton wins slightly more often than Trump.
Silver has in the past been known to actually pick states (49/50 and 50/50 in the past two cycles), so there was an expectation that despite his more recent probabilistic model, that he would get it right as he did in the past.
Most polls got it wrong. Although if you search for Trafalgar Group, you will see that they got most of their polls bang on or within the margin of error.
http://www.realclearpolitics.com/epolls/latest_polls/president/
Nope, he's never made anything but probabilistic models as far as I'm aware. Here's the 2012 model, which gave Obama a 90% chance: http://fivethirtyeight.blogs.nytimes.com/fivethirtyeights-2012-forecast/.
Yup, exactly right. Donald Jr just could not see how it would be a conflict of interest and actually completely illegal. I really can't see how this WONT be a yuuuuge problem for him.
Given that Donald "I know the tax code better than anyone" Trump doesn't know what a blind trust is I think it's quite obvious he'll have his kids running it and to hell with the conflicts.
http://www.nytimes.com/2008/11/10/business/media/10silver.html
"Mr. Silver had predicted the popular vote within one percentage point, predicted 49 of 50 states’ results correctly, and predicted all of the resolved Senate races correctly...."
seems like each election one pollster gets it just right.
- Status