|
ErIog posted:So is your contention that Nate' s just supposed to go, "Model's canceled! Sorry! It's not exactly the same as PEC so gently caress it!" you seem to be really mad about me making obvious statements. The fact that he's not a political pundit deserves repeating because people treat him as such. As does he himself to an extent. I don't really think he should do anything differently, except perhaps do even less interpreting, and maybe fire Harry Enten. And FWIW I read probably 95% of the 538 articles, listened to him on Freakonomics, listened to the 538 episode of the TWWW podcast, etc. But I just don't happen to fanboy for some internet guy with a stats BSc as hard as you do
|
# ? Oct 24, 2016 10:22 |
|
|
# ? Jun 12, 2024 09:14 |
|
Gyges posted:Really, instead of trying to find analogies like field goal misses and the like, we should be focusing on what the chance of Trump winning really is. It's the statistical chance that everything we know is wrong. It's not an event that rests on some place kicker holding the ball wrong, or the field being wet. It's more like being struck by lightning or being hit by a meteorite. Reason #1: Polling is an inexact science. Several polls averaging a 7 point Clinton lead with no poll showing a Trump lead does not mean a Clinton win is guaranteed. What it does mean is that clinton's electoral votes are estimated to be near 50 +/- a single-digit margin of error with more than 90% accuracy. 93% chance of a Clinton win sounds pretty accurate, although, based on the polls only, it should be much higher, like 98-99%, we'll get to that below. If you're having trouble finding this likely, think of this way: Does that mean that if we were to re-run the election 100 times, Trump would win 7 out of those times? No, in fact, given that the same nation with the same citizens is voting, the same candidate would win every time, because the votes would always be the same. But, what if we conducted 100 more scientifically unbiased polls? Maybe 7 out of those polls would show Trump winning. And only one of these 100 polls would be closest to the final outcome. So perhaps one of those 7 polls is the one that is closest to the real actual outcome. Reason #2: There are still two weeks left were anything can happen. Perhaps Iron Abuela will contract another more serious round of pneumonia, or some hacked email will actually contain something damaging for once. Perhaps Trump is able to find definite evidence that one of the 11 women who have come out against him was in fact 'motivated' by the Clinton campaign, lending a sudden nuclear injection of legitimacy to his 'rigged' claims. This is why the previous odds show 93% vs 7%. They take into account not only today's polls but also the uncertainties of the future. So yeah Trump is unlikely to win but it simply ain't over 'til it's over. Bisse fucked around with this message at 10:56 on Oct 24, 2016 |
# ? Oct 24, 2016 10:54 |
|
fishmech posted:There are literally programs and sites you can use to generate a "new" cryptocurrency in under a minute. What I love about that Trumpcoin effort, it's been through 3 thread incarnations due to flamewars, "fud", and at least one scam of the op and now it's in a moderated thread phase (ideologically free!), where there's been a compaction cycle to use Prester Jane's jargon to this: quote:He is only trying to get a new page added to this thread so that folks will not see the link to the speech today. and the same guy links this For some reason the bitcoin folks suspect a scam!
|
# ? Oct 24, 2016 10:57 |
Aaaaaand IDB poll is now showing dead heat too, with a 0.1% advantage to Hillary. Waiting for Rasmussen...
|
|
# ? Oct 24, 2016 11:00 |
|
canepazzo posted:Aaaaaand IDB poll is now showing dead heat too, with a 0.1% advantage to Hillary. This phenomenon where outlier polls magically start following the consensus is, for the record, called 'herding.' I actually doubt LAT/USC will herd because while it has some issues with sampling and weighting, they seem to have too much integrity and transparency to herd.
|
# ? Oct 24, 2016 11:20 |
|
stone cold posted:Did CelestialScribe ever end up toxxing for Trump? When was the last time CS even posted here? I'm on mobile so I can't search. Actually I can search and the freaks probated for 12 more days. Must have gotten a month on Oct 5, his last post here. Islam is the Lite Rock FM fucked around with this message at 11:34 on Oct 24, 2016 |
# ? Oct 24, 2016 11:31 |
|
Antti posted:This phenomenon where outlier polls magically start following the consensus is, for the record, called 'herding.' How can one distinguish herding from the polling realities swinging so much that baked-in bias cannot protect the favored side anymore?
|
# ? Oct 24, 2016 11:42 |
|
ErIog posted:You need to go study basic probability if you think the number of times you roll the dice has an impact on what the dice may show on a single roll. All of the models now point to a Hillary victory. The differences between them, and in their respective final election outcome odds, is in how much they model the possibility of their own model being incorrect. It's insufficient to say "the odds are the odds". You have to say "there is some uncertainty that the odds are what they say they are".
|
# ? Oct 24, 2016 11:48 |
|
WeAreTheRomans posted:It's not even that. It's that Nate has a perverse incentive to pursue his strangely conservative model because it promotes the horserace and drives site visits. Nate is also not a specially gifted or qualified statistician despite the myth that has grown up around him, and his attempts at punditry lack insight and objectivity. Most? He's 198 in 200 since 2008. That's a little more than most.
|
# ? Oct 24, 2016 11:55 |
|
Dick Milhous Rock! posted:Didnt Silver also have a weird economic feedback problem with his model? I seem to recall he weights economic data in his modeling because he believes it's more predictive. Silver has an economic component but only in the Polls Plus, and it doesn't really make much of a difference beyond the very early part of the campaign when polling is sparse. Sam Wang is from Princeton but he's a neuroscientist not a statistician and I have serious problems with his methods and conclusions. I don't know why people are holding him up on a pedestal particularly. Really the only professional statistician in the room is me, listen to me
|
# ? Oct 24, 2016 12:02 |
|
canepazzo posted:Aaaaaand IDB poll is now showing dead heat too, with a 0.1% advantage to Hillary. What's the IBD methodology? It has like 6-8 lower support for Hillary than other polls so some sort of 2004 turnout model?
|
# ? Oct 24, 2016 12:05 |
|
Neeksy posted:How can one distinguish herding from the polling realities swinging so much that baked-in bias cannot protect the favored side anymore? If it's systematic bias that gap shouldn't in general disappear. Like LAT might eventually show +5 C but only when everyone else is showing +15 or something. Herding is when ~magically~ everyone ends up with +15 C just before the election.
|
# ? Oct 24, 2016 12:07 |
|
Fangz posted:Silver has an economic component but only in the Polls Plus, and it doesn't really make much of a difference beyond the very early part of the campaign when polling is sparse. I wasn't going to wade into this but for a thread that sometimes sucks Taleb's dick they sure do hate 538 for not dismissing the chance of black swan events out of hand.
|
# ? Oct 24, 2016 12:09 |
Neeksy posted:How can one distinguish herding from the polling realities swinging so much that baked-in bias cannot protect the favored side anymore? Well in IBD/TIPP's case, they did it during the last two cycles so it seems pretty likely that they're doing it now, too. In 2008 they had a strong bias towards McCain and suddenly swung to match other pollsters in their very last poll before the election, and in 2012 they leaned heavily Romney before starting to swing in line with other pollsters the last week of October. Theris fucked around with this message at 12:24 on Oct 24, 2016 |
|
# ? Oct 24, 2016 12:21 |
|
538's model is fine. Polls+ will rise as the time in which a black swan event can happen shrinks. It will never go to 99% though as the simulations account for (unlikely, but still technically *possible*) catastrophic polling errors. Also, re the IBD poll, it overweights Republicans as a share of the sample compared to national party affiliation and has a record of herding closer to the election date. And it's still showing Clinton in the lead (just!).
|
# ? Oct 24, 2016 12:24 |
|
Fangz posted:Silver has an economic component but only in the Polls Plus, and it doesn't really make much of a difference beyond the very early part of the campaign when polling is sparse. Do I have to start telling stories about PECOTA and the unintelligible mass of excel spreadsheets that are the reason why Nate Silver has a job casting entrails for the political junkies?
|
# ? Oct 24, 2016 12:30 |
|
Chinese Gordon posted:538's model is fine. Polls+ will rise as the time in which a black swan event can happen shrinks. It will never go to 99% though as the simulations account for (unlikely, but still technically *possible*) catastrophic polling errors. 538's polls plus will probably converge on to the Nowcast. In terms of the black swan event, what's left is pretty much not the threat of some late event, but instead a systematic bias in the polling method. (The stuff people said earlier about Trump supporters hanging up on pollsters would be an example of this, if folks want something to Arzy about. Similar, stuff like voter intimidation/suppression might well apply.) In 2012 the poll aggregate missed by 2% and in previous years it has missed by more. Of course in 2012 the miss was in favour of Obama, but the model is agnostic about that. And yeah, the IBD does party identification weighting, which is rather iffy. Especially when the description in their methodology is very vague: quote:We also use party weighting to best mirror our internal estimates based on our recent polls If someone wanted to put a finger on the scale that's where I would do it. rscott posted:Do I have to start telling stories about PECOTA and the unintelligible mass of excel spreadsheets that are the reason why Nate Silver has a job casting entrails for the political junkies? If you have an argument why this is relevant given his current methodology is fairly transparent and reasonable, and matches closely to betting market odds? Fangz fucked around with this message at 12:51 on Oct 24, 2016 |
# ? Oct 24, 2016 12:49 |
|
Pointing out that Nate Silver was a (not that great) sabermetrician is just as relevant as pointing out that Sam Wang is a neuroscientist
|
# ? Oct 24, 2016 12:56 |
|
Systematic polling error is possible, but I don't think widespread voter intimidation is realistic. In order to have a noticable effect, it would have to be incredibly heavy-handed and well-organised. I can buy the Trumpstaffel being idiotic thugs, but idiotic, *well organised* thugs? No.
|
# ? Oct 24, 2016 12:57 |
|
rscott posted:Pointing out that Nate Silver was a (not that great) sabermetrician is just as relevant as pointing out that Sam Wang is a neuroscientist Right, so like I said, both of them are amateurs. My point was more that 'but but but Wang is in Princeton!!!' is a ridiculous argument. Fangz fucked around with this message at 12:59 on Oct 24, 2016 |
# ? Oct 24, 2016 12:57 |
|
Chinese Gordon posted:Systematic polling error is possible, but I don't think widespread voter intimidation is realistic. In order to have a noticable effect, it would have to be incredibly heavy-handed and well-organised. I can buy the Trumpstaffel being idiotic thugs, but idiotic, *well organised* thugs? No. Its gonna be an ugly day when one of them walks in on someone pushing ballot buttons and pulls a gun when they see you voting the wrong way. If they actually load it and if you can overpower their decrepit rear end is probably likely against the Trumpstain, but I'm sure they'll be open carry fuckers looking to curtail any people not voting for their side.
|
# ? Oct 24, 2016 13:01 |
|
https://twitter.com/realdonaldtrump/status/790523240351498241 What the poo poo is this?
|
# ? Oct 24, 2016 13:02 |
|
I assume Hillary wants to take in more Syrian refugees? To which I say great.
|
# ? Oct 24, 2016 13:03 |
|
Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 WikiLeaks is now down to "leaking" publicly held policy positions.
|
# ? Oct 24, 2016 13:03 |
|
Crabtree posted:Its gonna be an ugly day when one of them walks in on someone pushing ballot buttons and pulls a gun when they see you voting the wrong way. If they actually load it and if you can overpower their decrepit rear end is probably likely against the Trumpstain, but I'm sure they'll be open carry fuckers looking to curtail any people not voting for their side. Isolated incidents like this I'm sure are possible, even likely. But like I said, in order for it to have a noticeable state/national effect it would have to be both very illegal and very well organised. There is no Trumpette SA equivalent (yet!!).
|
# ? Oct 24, 2016 13:05 |
|
Chinese Gordon posted:Systematic polling error is possible, but I don't think widespread voter intimidation is realistic. In order to have a noticable effect, it would have to be incredibly heavy-handed and well-organised. I can buy the Trumpstaffel being idiotic thugs, but idiotic, *well organised* thugs? No. We're sorta in untested territory on this, though. Even the fear of Trump's 'poll watchers' can possibly have an impact, it's something that is very difficult to quantify.
|
# ? Oct 24, 2016 13:05 |
Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 He signs his tweets?!?
|
|
# ? Oct 24, 2016 13:04 |
|
Inferior Third Season posted:The problem is in determining what the odds actually are in a one-time event that can never be duplicated. After the election, assuming Hillary wins, we still won't know if 538's 87%ish chance of her winning was a better estimate than PEC's 99% chance. I, unlike other people posting here, understand that and am trying to point out that we know the chance of a Trump win is somewhere between 1% and 14%. Other people in this thread are trying to paint it like the 14% is unforgivable and we need the pitchforks and torches because they're rabid partisans who are loving stupid. So I agree with you, and I wish people would stop getting the loving vapors because 538 doesn't read 1% for Trump right now. WeAreTheRomans posted:I don't really think he should do anything differently, except perhaps do even less interpreting, and maybe fire Harry Enten. Enten has been even more anti-Trump than Silver has. What the gently caress are you smoking. You're an idiot. Go read PEC so you can stop dex_sda posted:Many people are treating him like he does, that's the whole point you dummy Who? He's done even less media this cycle than he did in 2012. He has his own podcasts and stuff now, but like, the only people who pay attention to him are people in this thread and people who follow ESPN. porfiria posted:A small point, but I disagree that 538 has had a "Hillary's gonna win" angle since August. I mean, he was writing stuff about Trump's "1 in 4" chances really recently. You're a loving idiot and I bolded the idiot part. That's 75% chance for Hillary to get anywhere from 50.1% to a complete blowout. Recently it's been like 15% for Trump to eke out a victory with perfect circumstances. Good job reading the headlines and having no reading comprehension you loving idiot. They've also written articles recently about how Trump's GOTV is going to be terrible. So please straight up brolic posted:538 is the most popular projections site. It isn't. So please shut the gently caress up. NYT and RCP are far ahead. You can pretend he's more popular than you want because people here want to feel like their dick's bigger because they called Nate Silver out, but that doesn't make it true. P.S. You're a stupid person. edit: Also Hillary's going to win, and I think 538 has terrible modeling, but the partisanship in this thread is out of control. (USER WAS PUT ON PROBATION FOR THIS POST)
|
# ? Oct 24, 2016 13:06 |
|
Fangz posted:We're sorta in untested territory on this, though. Even the fear of Trump's 'poll watchers' can possibly have an impact, it's something that is very difficult to quantify. I suspect that even if there is a Trumpstaffel Fear Effect it will be massively outweighed by the hugely superior Democratic GOTV operation.
|
# ? Oct 24, 2016 13:12 |
|
Chinese Gordon posted:538's model is fine. Polls+ will rise as the time in which a black swan event can happen shrinks. It will never go to 99% though as the simulations account for (unlikely, but still technically *possible*) catastrophic polling errors. Frank Luntz of all people took a dump on the IBD poll yesterday. It's a garbage poll for garbage people. Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 Probably the new Project Veritas video which is entirely wrong.
|
# ? Oct 24, 2016 13:15 |
|
I'm a little surprised by how bad Wikileaks/Russia is at propaganda. Even if their goal was simply to create a chaos and distrust, they are completely terrible at this. Anyways, with 15 days to go and in-person early voting starting in Florida, I have to imagine that if either side has any bombshells left to drop, we are going to see it in the next day or two.
|
# ? Oct 24, 2016 13:15 |
|
https://twitter.com/kylegriffin1/status/790522535976763393
|
# ? Oct 24, 2016 13:19 |
|
Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 So what does he do? He screams ISIS because that kinda worked last time he had to change the story. Even his idiot base knows the deflection of a defeated child, and while they lack any sense of nuance or critical thinking, they'll abandon him because they only vote for winners. The final cratering is coming, and it's going to be beautiful.
|
# ? Oct 24, 2016 13:19 |
|
If Clinton really was pushing to take in as many refugees as possible I'd vote for her even harder.
|
# ? Oct 24, 2016 13:21 |
|
Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 Way to say all Syrian refugees are terrorists you orange dick bag. On another note, it disheartens me that apparently Ivanka's brand is doing really well despite being the Cheeto's surrogate. This really proves to me that if you put a pretty face in front of a fascist and bigoted platform it will sell well. It pisses me off to no end that Ivanka gets such a huge pass on all this poo poo. Agrajag fucked around with this message at 13:26 on Oct 24, 2016 |
# ? Oct 24, 2016 13:22 |
|
Mr Hootington posted:https://twitter.com/realdonaldtrump/status/790523240351498241 Its another stupid as hell O'Keefe video featuring Huma Abedin. It was mentioned a while ago, but basically: Tom Guycot posted:Just catching up on the thread this morning, but if you want an idea on how this works, watch the undercover video of "Huma admitting Hillary wants to let in all the Syrians!!". They actually leave in the person doing the filming going up to her, saying her hi's, hello's, how much she loves Hillary and Huma's speach etc, but then saying "you have to promise, promise me that Hillary will bring them all over, promise me". Huma of course in a crowded room of people greeting and talking to her just gives a 'yes of course Hillary is very committed to helping refugees yada yada'.
|
# ? Oct 24, 2016 13:25 |
|
Agrajag posted:Way to say all Syrian refugees are terrorists you orange dick bag. It's not a new stance by the gop.
|
# ? Oct 24, 2016 13:24 |
|
ErIog posted:Enten has been even more anti-Trump than Silver has. What the gently caress are you smoking. You're an idiot. Go read PEC so you can stop Dude, please step away from your posting devices for a while and chill out. How can you be so vexed about polling aggregation that you're multi-quoting like 10 people (who don't even disagree with you) and calling them loving idiots? They should fire Enten because he is a goony idiotfucker, not because I care about his political leanings (which are pretty obvious, and I am aware of). I am not arzying in any way, and I am not unhappy about polls, polling, or poll aggregation. ErIog posted:edit: Also Hillary's going to win, and I think 538 has terrible modeling. I agree. You seem to be arguing against positions that nobody is taking here.
|
# ? Oct 24, 2016 13:32 |
|
Agrajag posted:Way to say all Syrian refugees are terrorists you orange dick bag. The Clinton Camp (rightly IMHO) doesn't really want to attack family members of the opposing candidate. That'd add unwelcome credibility to Trump attacking Hillary for Bill's stuff and potentially put Chelsea on the hit-list. To be honest, Ivanka has a better excuse for supporting Trump than everyone else.
|
# ? Oct 24, 2016 13:35 |
|
|
# ? Jun 12, 2024 09:14 |
|
this seems timely http://www.nytimes.com/2016/10/24/opinion/my-wifes-killer-was-not-an-illegal-immigrant.html quote:On Nov. 1, 2006, I found my wife, Adrienne Shelly, dead in her West Village office. Adrienne, an actor and filmmaker, had been brutally murdered by a 19-year-old undocumented Ecuadorean construction worker; he later said they were having an argument and, fearing she would report him and have him deported, he killed her and staged her death so it would appear to be a suicide. Our daughter was just 2 years old at the time.
|
# ? Oct 24, 2016 13:35 |