Trump could win the election in a nowcast by FiveThirtyEight

[sleestak]

@DonLogan

No.

Someone who turns “Men can be such pigs” into “All men are pigs” in order to be able to take offense at it is doing the same exact thing as someone who turns “Black Lives Matter” into “Black Lives Matter More.”

They’re stupid assholes. There aren’t enough resources to sit down with them every time they do it and give them a cookie and explain why they’re wrong. You just have to write them off and move on.

Or, if a sizable percentage of the population takes offense, instead of labeling them racists maybe, just maybe, you ought to check and see if your message is having the intended affect.

If I say something and a large chunk of my audience does not understand the point I am trying to make, I don’t call them all stupid racists and leave it at that. I check to ensure that I am being clear and then try another approach because, if the message I am sending is not being understood, then I am most certainly part of the problem.

Of course, it is just easier to scream racism and call them all stooopid. It is the lazy, intellectually cheap way to score points.

Slee

[RickJay]

@Jonathan_Chance

At Convention bounce? Surge now that the general election is underway? Sanity returning to the American electorate?

Silver himself says in their round table discussion that it’s bounce, and that there are not polls sufficiently fresh to measure the extent of the post-convention PR debacles.

His latest measured national polls both concluded August 1 (there is one that goes to Aug m2, but it doesn’t even have a grade on it and Silver doesn’t seem to assign a lot of trust to it) and really things have gotten way worse since then. 538 has NO genuinely recent state level polls for Florida, NC, Ohio, or Virginia, and the most recent one from Pennsylvania concluded July 29 so could at best measure convention bump.

[Jonathan_Chance]

@sleestak

Or, if a sizable percentage of the population takes offense, instead of labeling them racists maybe, just maybe, you ought to check and see if your message is having the intended affect.

If I say something and a large chunk of my audience does not understand the point I am trying to make, I don’t call them all stupid racists and leave it at that. I check to ensure that I am being clear and then try another approach because, if the message I am sending is not being understood, then I am most certainly part of the problem.

Of course, it is just easier to scream racism and call them all stooopid. It is the lazy, intellectually cheap way to score points.

Slee

Slee,

I’m just going to note you here instead of giving you a warning for failure to follow instructions. When I say take the hijack elsewhere I mean it.

[Dangerosa]

@RickJay

Silver himself says in their round table discussion that it’s bounce, and that there are not polls sufficiently fresh to measure the extent of the post-convention PR debacles.

His latest measured national polls both concluded August 1 (there is one that goes to Aug m2, but it doesn’t even have a grade on it and Silver doesn’t seem to assign a lot of trust to it) and really things have gotten way worse since then. 538 has NO genuinely recent state level polls for Florida, NC, Ohio, or Virginia, and the most recent one from Pennsylvania concluded July 29 so could at best measure convention bump.

But that PR debacle is likely to create an issue with being able to separate out bounce from debacle - and debacle from election milestone.

[Thing_Fish]

@ Carnal K:

Well, the question of real interest isn’t the trivial one of what the polls say right now, but how much and in what direction they will change between now and the election.

So my understanding of both these models is that they apply historical data to that question by looking how polls have typically moved in the past between (today’s date here) and the election. Silver says that his model becomes a couple percent more accurate by not treating all past data as equal, but controlling for variables such as economic indicators and the approval rating of the incumbent President. I have nowhere near enough knowledge to venture an opinion about whether he is right, though his track record is certainly excellent.

[Exapno_Mapcase]

Ari Fleischer pointed out that Trump could have had a boost over the weekend pointing out the weak economic numbers released on Friday. Instead he attacked Khan and kept doing so despite the apparent unanimous advice of his staff. Today we get reports that the staff is not merely in disarray but “suicidal.” (From a tweet quoted in a Washington Post article, IIRC.)

Although there’s been rampant speculation that Trump will skip the debates, I’m sure he’ll be there. The nowcast responded to the enormous media coverage of the convention and of course quickly returned to the basic state of the race, where Clinton is soundly in the lead. The public does not pay attention to the daily gaffes and blunders the way we do. They will tune in for a debate. Getting his message across there is Trump’s best chance to even the race, just by sheer face time.

[DSeid]

@Thing_Fish

… though his track record is certainly excellent.

Is there any actual basis for saying that his track record of estimating the odds of November’s results at this point is so excellent?

Yes, aggregation of polls immediately before the election has been excellent at predicting outcomes, and it can be debated as to which aggregation system works better. But that is a very different thing.

FWIW here is Silver’s past, somewhat disparaging, view of economic fundamentals as a predictive tool.

[Voyager]

@Exapno_Mapcase

Ari Fleischer pointed out that Trump could have had a boost over the weekend pointing out the weak economic numbers released on Friday. Instead he attacked Khan and kept doing so despite the apparent unanimous advice of his staff. Today we get reports that the staff is not merely in disarray but “suicidal.” (From a tweet quoted in a Washington Post article, IIRC.)

Although there’s been rampant speculation that Trump will skip the debates, I’m sure he’ll be there. The nowcast responded to the enormous media coverage of the convention and of course quickly returned to the basic state of the race, where Clinton is soundly in the lead. The public does not pay attention to the daily gaffes and blunders the way we do. They will tune in for a debate. Getting his message across there is Trump’s best chance to even the race, just by sheer face time.

The Republicans bringing up the emails again and again are not likely to have any effect, since the impact on the perception of Clinton is already baked into the polls. But Trump insulting Gold Star parents is new. That is something that affects the perception of him in a new way, and probably is more important to his base than to the opposition.
It isn’t going to help his numbers, that’s for sure.

[Thing_Fish]

@DSeid

Is there any actual basis for saying that his track record of estimating the odds of November’s results at this point is so excellent?

Yes, aggregation of polls immediately before the election has been excellent at predicting outcomes, and it can be debated as to which aggregation system works better. But that is a very different thing.

FWIW here is Silver’s past, somewhat disparaging, view of economic fundamentals as a predictive tool.

Well, that’s a good point. AFAIK he only started doing this in 2008, so there’s way too small a sample size to say. My impression in following his work fairly closely is that he quite accurately predicts how each state will go immediately before the election (which as you point out isn’t really all that difficult to do), and that in the months before, his numbers tend to move in what seems like a gradual fashion; IOW, he starts with preliminary estimates and arrives at correct final results without having to make any big, sudden shifts. That makes me think that he knows what he’s doing, but I’m not sure if that’s good logic or not.

[asahi]

@Little_Nemo

Bullshit. Trump’s base isn’t made up of decent moderate people who are being falsely accused of racism or sexism. Trump’s base is made up of people who are racist and sexist.

I don’t disagree that there are a lot of racist and sexist people in his voting bloc, but the Trump phenomenon is a lot more complex than conscious bigotry. Progressives would be mistaken to oversimplify the rise of Trumpism. There are indeed moderate and decent people among Trump’s supporters, and to condescendingly suggest otherwise makes it more likely that this contingent will grow.

[DonLogan]

@sleestak

Or, if a sizable percentage of the population takes offense, instead of labeling them racists maybe, just maybe, you ought to check and see if your message is having the intended affect.

If I say something and a large chunk of my audience does not understand the point I am trying to make, I don’t call them all stupid racists and leave it at that. I check to ensure that I am being clear and then try another approach because, if the message I am sending is not being understood, then I am most certainly part of the problem.

Of course, it is just easier to scream racism and call them all stooopid. It is the lazy, intellectually cheap way to score points.

Slee

You must be joking. Is the phrase “Black Lives Matter,” having been spawned from police shootings of black people, really that difficult to understand? Is it?

Now who would possibly want to take offense at that and act like the phrase was “Black Lives Matter More Than White Lives”? Now who would it be that would do that? Who claims reverse-racism where it doesn’t exist? Racists.

Who would it be that would take the phrase “Men can be pigs sometimes” and act like it was “All men are pigs”? Now who would do that? Who claims reverse-sexism where it doesn’t exist? Sexists.

Who would it be who cannot comprehend the meaning of the original words? Idiots.

[tomndebb]

@DonLogan

You must be joking. Is the phrase “Black Lives Matter,” having been spawned from police shootings of black people, really that difficult to understand? Is it?

Now who would possibly want to take offense at that and act like the phrase was “Black Lives Matter More Than White Lives”? Now who would it be that would do that? Who claims reverse-racism where it doesn’t exist? Racists.

Who would it be that would take the phrase “Men can be pigs sometimes” and act like it was “All men are pigs”? Now who would do that? Who claims reverse-sexism where it doesn’t exist? Sexists.

Who would it be who cannot comprehend the meaning of the original words? Idiots.

**Knock it off. ** The BLM discussion is a hijack.

You are welcome to open a new thread to discuss it.

The next post in this thread arguing about BLM will receive a Warning.

**[ /Moderating ]
**

[Measure_for_Measure]

Back to the OP.
Nate and his crew joke about the Nowcast in this week’s politics chat. Recall that they present 3 forecast frameworks: they are polls-plus, polls-only, and the nowcast. The first one is their real forecast. The second tells you the effect of fundamentals on their forecast. The nowcast tells about the most recent polling.

Here’s what they say today: [INDENT][INDENT][INDENT]

clare.malone: I don’t think she got the bounces because of the Khans. OBAMA. Obama coming out there and basically taking shots at Trump I think added a lot.

harry: I think every little thing adds up. And we cannot disentangle what exactly is going on.

micah: Obama got a bounce too.

clare.malone: Right. So basically we should all check back in two weeks. And watch some OLYMPICS.

micah: Except, dear reader, make sure you check this site every day between now and then.

natesilver: Every HOUR.

clare.malone: NOW-CAST.

harry: DO IT!

clare.malone: But don’t tweet at me about it. That really stressed me out this weekend. I don’t know how the model is built, people. That’s all up in Nate’s mind-palace.

harry: Which looks like Trump’s apartment, fyi.

natesilver: The now-cast is a powerful drug. [/INDENT][/INDENT][/INDENT] The now-cast is basically a performance art joke about media representations of the horse-race. Ironically it’s better than most election reports, as at least it averages more than one poll. So the error bounds are smaller, though the underlying model, which presupposes that the election were held today, is somewhat silly. General elections in the US occur in November, at least for Presidents.

Polls-plus:

2016 Election Forecast | FiveThirtyEight

Note that polls plus is just a baseline, based on history. It doesn’t take into account qualitative information about the race. And boy does this cycle contain qualitative information.

How to think about 538’s 3-model frameworks:

A Brief Guide for Matthew Yglesias, Perplexed as He Is About the Metaphysical...

[Evil_Economist]

@Measure_for_Measure

The now-cast is basically a performance art joke about media representations of the horse-race. Ironically it’s better than most election reports, as at least it averages more than one poll.

It may be silly but it’s fun to follow. For example, right now Trump is down to 11%. Judging from outcomes, I believe that makes the RNC convention the worst convention in history.

[Dangerosa]

@Evil_Economist

It may be silly but it’s fun to follow. For example, right now Trump is down to 11%. Judging from outcomes, I believe that makes the RNC convention the worst convention in history.

Fun as long as you aren’t rooting for Trump to win. To be clear, I’m not, but there were a few days when looking at the nowcast was like looking at my 401k in late 2008 - just don’t open the envelope and hope it changes.

[JKellyMap]

@Dangerosa

Fun as long as you aren’t rooting for Trump to win. To be clear, I’m not, but there were a few days when looking at the nowcast was like looking at my 401k in late 2008 - just don’t open the envelope and hope it changes.

Tell me about it. All ten of my fingernails are just now starting to recover from being bitten to the quick (no joke!).

[septimus]

Over a period of just six days, chance for GOP White House as shown at Predictwise and Betfair has fallen from 33% to 23%. Whether due to his insults against the Khan family, his fallings-out with GOP leaders including his own hand-picked V.P. candidate, or just general derangement, this collapse was strong and sudden.

I feel like America may have escaped a grave danger. But don’t relax! Be sure to vote. Encourage fellow Americans to retain their sanity.

If the GOP collapse leads to a Democratic majority in the House of Reps, ring all the bells, and declare a great day of celebration!

[asahi]

@septimus

Over a period of just six days, chance for GOP White House as shown at Predictwise and Betfair has fallen from 33% to 23%. Whether due to his insults against the Khan family, his fallings-out with GOP leaders including his own hand-picked V.P. candidate, or just general derangement, this collapse was strong and sudden.

I feel like America may have escaped a grave danger. But don’t relax! Be sure to vote. Encourage fellow Americans to retain their sanity.

If the GOP collapse leads to a Democratic majority in the House of Reps, ring all the bells, and declare a great day of celebration!

There’s time for a recovery, but not much. I think Newt Gingrich is right: he has about two weeks to get his shit together or it’s over. Trump needs the republican machine, both at the national and state levels, to do the grinding work of winning elections. Trump has been not only making an ass of himself but he’s now having a visible impact on the party’s chances of keeping the House and Senate. The GOP is afraid of Trump’s rabid cult-like following, but if Trump’s decision to implicitly support Paul Ryan’s primary challenge results in a shocking defeat or even a nail-biting win, I suspect you will see the party completely abandon him…which could kick off a civil war among conservatives that gets brought out into the open and gets really, really nasty.

[BigAppleBucky]

I don’t know why folks are discussing a Democratic majority in the House. It just will not happen. Sure, the Democratic House candidates will probably garner five or even ten million more votes nationwide, the the GOP will only lose a few seats.

The Senate, where no there is no gerrymandering to help the GOP, is definitely in play. The Democrats can take it. Not predicting they will, but it’s certainly possible.

538 did offer this on the Senate:

Trump May Start Dragging GOP Senate Candidates Down With Him

Republicans in most Senate battlegrounds are running ahead of Trump in their states.

Trump may drag down some sitting GOP Senators.

[Pleonast]

@BigAppleBucky

I don’t know why folks are discussing a Democratic majority in the House. It just will not happen. Sure, the Democratic House candidates will probably garner five or even ten million more votes nationwide, the the GOP will only lose a few seats.

Gerrymandering works by giving the controlling party slim majorities in many districts and concentrating opposition voters in a few districts. So if everyone votes the way they have in the past, the controlling party gets a disproportionate number of seats. But, if those slim majorities slip away, you get a tidal wave going the other direction.

I worked out the math for the case of Ohio:

@Pleonast

I think many people underestimate the potential impact of gerrymandering. Take Ohio, for example. The voting indices by district, in order from pro-Republican (+) to pro-Democrat (-) are: +14, +9, +9, +8, +8, +6, +6, +5, +5, +5, +3, +2, -12, -14, -15, -30. The baseline is 12 R and 4 D in the House, if the presidential race has no down-ticket impact. If there’s a 3.5-point effect in favor of the Democrats, it’ll go 10 R, 6 D. A 5.5-point effect, results in 7 R, 9 D. And 6.5 goes to 5 R, 11 D. In contrast, even a 10-point shift in favor of the Republicans does not result in the Democrats losing any seats.

Gerrymandering really tilts the results non-linearly.

Of course, what are the odds of a 5 or 6 point shift in favor of the Democrats? I’m not sure, but when something like 30% of the Republicans are not happy with their nominee, it wouldn’t take “something extraordinary”. More like “against the odds, but possible”. I’ll be very interested in Nate Silver’s analyses.