President Trump

JimHow · Post by **JimHow** » Sun Oct 18, 2020 5:37 pm

Two weeks to go. As CV has said repeatedly: If we turn out the vote, we will win.

JimHow · Post by **JimHow** » Sun Oct 18, 2020 5:44 pm

I've been reading about the Thai king with a $40 billion fortune who lives in Germany....
Just shaking my head at what is going on in this world.

JimHow · Post by **JimHow** » Sun Oct 18, 2020 10:40 pm

Good article on why they like Orange Head:

https://www.nytimes.com/2020/10/16/opin ... e=Homepage

Comte Flaneur · Post by **Comte Flaneur** » Tue Oct 20, 2020 7:14 pm

I have a question for stefan - and perhaps David too - about the 538 methodology

This from a critic of the methodology

“Something that I have seen frequently in the media are the forecasts by 538 and similar ones (which cite Biden at 87% chance, Trump with 13% chance). However, they are not robust and very mistaken.

538 forecasts the elections by first constructing a model and then running loads of Monte Carlo simulations on the model, yielding an X times that one candidate wins and (1-X) times the other wins. So the 13% chance 538 associates with Trump is merely 13% of the MC runs spitting out Trump as the winner.

Its not a real life estimate of the election - its just an MC simulation run of your model that is used as a “forecast” by 538. And yet most the media runs around with such estimates as if being true forecasts (which they are not).

Its quite straightforward to see the main problem in such an approach - the elections are essentially a binary game with two possible choice. Given high uncertainty as to the result, the probabilities of the outcome for both are ~0.5. (High uncertainty implies the election can go either ways, so intuitively the probabilities hover approx. to 0.5). And broadly speaking the electorate is split 50-50 to right and left. Meaning that Trump has a far higher probability than 0.13 that 538 is forecasting. Last major landslide was Reagan’s second term victory in 1984 where he won 58% of the votes. Since then no candidate has won by more than 54% of the votes. (George H.W Bush’s win in 1988 was close - 53.4%, but no other candidate has come close).

Adding atop these historical trends the massive uncertainty in these elections, the probabilities should be around 0.5, or hover around in a range relatively close to them. So essentially implying that betting odds which are far closer to 0.5 are much much more accurate in depicting reality than 538 forecasts.

Ends

I would be interested in your thoughts Bill.

Blanquito · Post by **Blanquito** » Tue Oct 20, 2020 7:55 pm

Whomever wrote this seems to be making a sophomoric conflation of margin of victory and probability of victory.

Blanquito · Post by **Blanquito** » Tue Oct 20, 2020 8:03 pm

All 538 is doing is calculating the odds of victory in the election based on the polls at a given date as those odds applied in prior elections, therefore assuming that the same relationships between polls and election outcomes that applied in past elections apply today. The Monte Carlo simulation allows these predictions to be probabilistic rather than deterministic.

Blanquito · Post by **Blanquito** » Tue Oct 20, 2020 8:23 pm

That post claims that the electorate is evenly divided, which if true would result in a roughly 50-50 chance of victory for both candidates (ignoring the realities of the Electoral College), but that’s why we poll! And the polls show that the electorate is not evenly divided, not even close. And given the date and Biden’s lead in the polls, 538’s models find that Biden has an 87% chance of victory, based on probabilities derived from polls and past elections.

I like this table from Nate Silver as a handy summary to translate the popular vote royals into odds of victory:
Chance of a Biden Electoral college win if he wins the popular vote by X points:
0-1 points: just 6%!
1-2 points: 22%
2-3 points: 46%
3-4 points: 74%
4-5 points: 89%
5-6 points: 98%
6-7 points: 99%

stefan · Post by **stefan** » Tue Oct 20, 2020 8:49 pm

Ian, as far as I know, the only serious science involves a specific point in time. A poll shows Biden gets 57% of the vote and Trump gets 43%, with an error of 6.9% in the direction of Biden with confidence 95%, say. One interpretation of this statement is that had the poll involved the entire voting population, there is a 95% probability that Biden would have won. Even that is a misstatement, analogous to the following. Suppose you have a bin that contains 1000 balls, some red and some blue. You draw at random 100 balls, 57 of which are blue and 43 of which are red. The teacher asks, "what is the probability that there are more blue balls than red balls?" I am too lazy to compute the answer that the teacher expects, but the real answer is "either one or zero". Still, in such a situation, even without doing the computation, at even odds I would be willing to bet a lot of money that a full counting will show that there are more blue balls. (I like it that you, unlike the teacher, asks for "betting odds".)

I agree with you about the 538 forecasts. If someone offered me 87-13 odds that Biden will win the election, I would take tRump in a flash while hoping that I would lose the bet. 68-32 or thereabouts is what you can get currently.

JCNorthway · Post by **JCNorthway** » Wed Oct 21, 2020 1:21 am

i recall probability theory from my MBA days. But 35 year later, it makes me dizzy.

JimHow · Post by **JimHow** » Wed Oct 21, 2020 1:39 am

This race is eerily similar to 2016.
Trump had some shit go down with about a month left in each race that seemed to finish him off.
But he slowly bounced back. And he avoided mistakes in the last two weeks.
Here’s a story from two weeks before the 2016 election, with Hillary up by 5 points. The CBS poll had her up by 9 points.
She had been up by like 8 points just a week earlier.
Joe is up by a little more, but the same dynamics are in play.
And it is funny to see the smug media saying things like “no candidate this far behind at this point has ever won,” etc.
That may have been the case, but there has never been a candidate like this guy.

https://www.cbsnews.com/news/election-2 ... ction-day/

JimHow · Post by **JimHow** » Wed Oct 21, 2020 1:47 am

Also, Comey hit with about 10 days to go.
And Teump, Fox and the Republicans are throwing Hunter out there hoping it sticks.
Meantime, Joe is taking 4 days off from the campaign with less than two weeks to go, not answering the charges.
Good grief.

Chateau Vin · Post by **Chateau Vin** » Wed Oct 21, 2020 4:55 am

Joe is taking time off to prepare for the debate. What is he? A high school student to skip something to prepare for a test?

Man, it's all about turnout for the dems, and he is back into hiding, instead of campaigning in PA, WI and MI...

DavidG · Post by **DavidG** » Wed Oct 21, 2020 4:58 am

I think you are all correct. What are the odds of that?

I’m encouraged by Biden's lead, the smaller numbers of undecideds than at this point in 2016, and Biden's outspending Trump.

I’m worried by the same stuff Jim is. Gotta be out there in public. I think the Hunter Biden stuff is a flop so far but who knows what crap they’ll pull next. And I really worry about Trump rigging the election by suppressing the vote and/or by disqualifying huge numbers of legitimate ballots. This is the first election I have worried about that. We know that when Trump protests mightily about the other side doing something unethical or illegal, that’s the very thing he’s up to. He’s been howling a lot about rigging the election. And unlike 2016, this time he’s got the DOJ and all the power of the Presidency at his command to assist him.

Chateau Vin · Post by **Chateau Vin** » Wed Oct 21, 2020 5:21 am

DavidG wrote:.
.
.
. when Trump protests mightily about the other side doing something unethical or illegal, that’s the very thing he’s up to.
.
.
..

I agree with your statement, and that is the thing I have been banging my head all along with my friends.

When he is inept, he calls others inept. When he is corrupt, he calls others corrupt before anyone calls it out. He calls others liars when he lies. He calls others manipulators, while he manipulates things underneath. The list goes on...It's like a preemptive act for him to hide his misdeeds that are underway or about to embark on.

Even the whole notion of 'counter puncher' is geared to give the impression that he is being attacked without reason...This country's dire predicament of current state of politics was primarily created and being fueled by Faux News, Facebook and the Social media...So much so for free speech...

Comte Flaneur · Post by **Comte Flaneur** » Wed Oct 21, 2020 8:00 am

stefan wrote:Ian, as far as I know, the only serious science involves a specific point in time. A poll shows Biden gets 57% of the vote and Trump gets 43%, with an error of 6.9% in the direction of Biden with confidence 95%, say. One interpretation of this statement is that had the poll involved the entire voting population, there is a 95% probability that Biden would have won. Even that is a misstatement, analogous to the following. Suppose you have a bin that contains 1000 balls, some red and some blue. You draw at random 100 balls, 57 of which are blue and 43 of which are red. The teacher asks, "what is the probability that there are more blue balls than red balls?" I am too lazy to compute the answer that the teacher expects, but the real answer is "either one or zero". Still, in such a situation, even without doing the computation, at even odds I would be willing to bet a lot of money that a full counting will show that there are more blue balls. (I like it that you, unlike the teacher, asks for "betting odds".)

I agree with you about the 538 forecasts. If someone offered me 87-13 odds that Biden will win the election, I would take tRump in a flash while hoping that I would lose the bet. 68-32 or thereabouts is what you can get currently.

Thanks Bill

Racer Chris · Post by **Racer Chris** » Wed Oct 21, 2020 11:59 am

JimHow wrote:Also, Comey hit with about 10 days to go.
And Teump, Fox and the Republicans are throwing Hunter out there hoping it sticks.
Meantime, Joe is taking 4 days off from the campaign with less than two weeks to go, not answering the charges.
Good grief.

I'm sure Chris Wray won't do anything to help Trump right now.

Rather than get in the way, Joe is letting Trump sink his own ship. Everything Trump's tried so far has only caused voters to move away from him.
In 2016, Trump had the media on his side - the shiny object phenomenon. In 2020, the media is onto his scheme, so instead they are working to expose his corruption at the critical time.

Joe has already said all he needs to on the latest attempt to scandalize his family. Fox is pretty much alone with trying to gain traction on the Hunter story. This will hurt their brand going forward.
Even Bill Barr isn't stupid enough to try charging Joe with anything.

Hm$(still) · Post by **Hm$(still)** » Wed Oct 21, 2020 12:53 pm

It’s wonderful having a mathematician in the group. Thanks Bill.

Howard

Chateau Vin · Post by **Chateau Vin** » Wed Oct 21, 2020 3:36 pm

Great article, and here's what's wrong with some of the people...Writing in a dead person? Great going, Governor....Sounds self-serving with a holier than thou attitude...

https://www.usatoday.com/story/opinion/ ... 712993001/

Racer Chris · Post by **Racer Chris** » Wed Oct 21, 2020 4:43 pm

Former chairman of the RNC, Michael Steele, ought to have a word with the MD governor.
He's in the latest Lincoln Project ad endorsing Biden.

Racer Chris · Post by **Racer Chris** » Fri Oct 23, 2020 1:40 pm

Well, that wasn't as painful as the last time. The new rules, the new moderator, and Trump managing to stay on the rails mostly, gave it some semblance of a debate at least. However only Biden offered anything of substance, while Trump spent his time lying and sowing chaos.
Biden wasn't at his best but didn't make any big mistakes, and Trump's blows never landed with any force. Trump likewise didn't make any big mistakes, and was almost civil, but was on defense more than he was on offense.

JimHow · Post by **JimHow** » Fri Oct 23, 2020 1:49 pm

All true, but did it give those who want to vote for Orange but are reluctant to an opening to do so?
The polls are narrowing and they are going to get tighter between now and Election Day.
Remember, anything under a 5% Biden national popular vote victory and Joe's chances of winning the electoral college deteriorate quickly.
I'm still not liking the numbers, they are much too Hillary-like, in fact, Joe is doing worse among blacks and white men, I believe.
Pennsylvania continues to give me heartburn.

Jay Winton · Post by **Jay Winton** » Fri Oct 23, 2020 1:51 pm

Yes, Biden was good enough and, of course, made far many valid points than the trumpster. Trump's base will support him but I don't think his performance will bring many undecided voters to him. Covid is spiralling seemingly out of control in many red states and hospitals are being overwhelmed. Biden was correct to keep hammering away at the president's miserable management of the epidemic. The final countdown.

JimHow · Post by **JimHow** » Fri Oct 23, 2020 1:58 pm

Looks like it's all going to come down to Pennsylvania and Amy Coney Barrett.
I've been impressed with the Pennsylvania attorney general and the PA courts.
And the Republican head of the Pennsylvania state legislature wrote an op-Ed saying they're not going to interfere with the election, which I have to believe is going to be the case, it would be a complete fiasco if they tried.
The most sobering thing that has happened, of course, is that Thomas/Alito/Gorsuch/Kavanaugh have telegraphed that they will interfere with the election.
And that will leave it all to Amy.
The politicization of the courts is probably the biggest thing that is going to lead to our demise.

Chateau Vin · Post by **Chateau Vin** » Fri Oct 23, 2020 9:03 pm

Can’t believe the SCOTUS has become the ruling class of the country... what a fall from grace for US democracy...

Blanquito · Post by **Blanquito** » Sat Oct 24, 2020 12:46 am

Love the logic, simplicity, and (of course) the takeaway in this Tweet by Nate Silver today:
“One way to think about Trump's chances: if the polls move toward Trump in the closing days rather than Biden (50/50 chance) *and* there's a polling error in Trump's favor (50/50 chance) then he's 50/50 to win. That gets you to his 1 in 8 odds in our current forecast.”

Blanquito · Post by **Blanquito** » Sat Oct 24, 2020 1:00 am

So even if we are sure the polls will tighten (which I’m not, not with COVID case skyrocketing and 1/3 of the likely electorate having already voted), that still only gives him a 25% chance of winning given that the polls would have to be wrong again in his favor (by no means a given — the polls underestimated Obama in 2012) and even then he’d still have to win a close election... nothing guaranteed of course, but without a genuine ‘Comey Letter‘ moment to shake things up in the last 10 days, I’m liking Biden’s chances.

JimHow · Post by **JimHow** » Sat Oct 24, 2020 1:17 am

Tell me what’s going to happen in Pennsylvania, both on and after Election Day, and I will tell you whether I agree with you or not.

hautbrionlover · Post by **hautbrionlover** » Sat Oct 24, 2020 4:12 am

Pennsylvania is key.
Unless, of course, Biden manages to win Florida, but what are the chances of him winning FL if he can’t win PA?
And NC, AZ and IA are all up for grabs (and possibly GA, but I wouldn’t bet on it).

Blanquito · Post by **Blanquito** » Sat Oct 24, 2020 3:25 pm

Pennsylvania is absolutely ground zero and it is uncomfortably tight. But even in 2016, Hillary wins PA if not for the Comey Letter. Hillary wins PA if not for so many late deciders picking Trump*, Hillary wins PA if the African American turnout was even decent, Hillary wins PA if she bothered to campaign there.

Looking at polls in October for Pennsylvania from pollsters with a B+ rating or better, here's what we see:
Oct. 23 -- Muhlenberg College -- Biden +7
Oct. 21 -- Fox News -- Biden +5
Oct. 21 -- Quinnipiac University -- Biden +8
Oct. 21 -- Suffolk University -- Biden +6
Oct. 7 -- Quinnipiac University -- Biden +13
Oct. 6 -- Emerson College -- Biden +3
Oct. 6 -- Monmouth College -- Biden +11
Oct. 3 -- Siena College/The New York Times Upshot -- Biden +7
Average = Biden +7.5

Whatever one thinks of 538's models, these polls translate into a 87% chance of a Biden victory in their model.

The thing that has me worried about PA isn't the voters -- it does seem like Biden has a robust enough lead to absorb a 2016-like polling error and it is clear this is going to be a a very high turnout election pretty much across the country -- it's all of the mail-in ballots not getting counted for days or weeks after election night. If the election is close in rest of the country and it all comes down to PA and it's uncounted mail-in ballots, then we are looking at the 2000 recount nightmare in Florida all over again.

*https://www.washingtonpost.com/news/the ... ald-trump/:
"In fact, if you look at the four closest states where Clinton lost, exit polls show late-deciding voters in each of them went strongly for Trump in the final days. In Florida and Pennsylvania, late-deciders favored Trump by 17 points. In Michigan, they went for Trump by 11 points. In Wisconsin, they broke for Trump by a whopping 29 points, 59-30... And these weren't small groups of voters. The number of undecided and third-party-supporting voters who were still free agents in the final week was as many as 1 in 8 voters nationally -- an uncharacteristically high number for the eve of an election. (As Nate Silver noted, it was just 3 percent in 2012.)"

Blanquito · Post by **Blanquito** » Sat Oct 24, 2020 3:35 pm

Personally, I am going to be watching FL and NC more closely than anywhere on election night. I understand the skepticism the Biden can win these states, but he has small but significantly bigger leads in both states than Hillary did this close to the election -- Hillary had a 55% chance of winning both FL and NC the day before the election (and just 5 days before, the polls and models actually favored Trump ever so slightly), while Biden currently has a 70% chance in FL and 66% in NC.

And more to the point, Trump realistically has not shot without winning both NC and FL, and both states will count all of their votes on election night as state laws lets them count mail in ballots as they come in (not having to wait until Election Day to start counting like the law requires in PA).

stefan · Post by **stefan** » Sat Oct 24, 2020 3:39 pm

Even Texas is in play. From Newsweek:

"The latest Morning Consult poll, released Thursday, showed the Democrat with a 1 percentage point lead over the Republican among 3,000 likely voters."

Unfortunately, Biden hurt himself badly in the debate with his comment about transitioning from oil. He could have, and should have, spoken longer and more sensitively on the topic. It reminded me of what Hillary said in 2016 about coal mining. Even semi-reasonable people accept that we must address climate change, but too often the left and even moderate wings of the democratic party ignore the fall out for the many people that depend on fossil fuel production for the livelihoods.

Comte Flaneur · Post by **Comte Flaneur** » Sat Oct 24, 2020 7:06 pm

Yes that was a significant gaffe and mistake. It will please the basket weavers in California but they weren’t exactly going to go with the Donald anyway - and it could put PA back in play and put Ohio and Texas out of reach.

JimHow · Post by **JimHow** » Sat Oct 24, 2020 8:22 pm

Man does this feel like four years ago.
Trump disciplined and on the attack, the Democrat on the defensive with ten days to go.
It just feels like a Trump win to me.

DavidG · Post by **DavidG** » Sat Oct 24, 2020 8:34 pm

Spot on Stefan regarding Biden's answer on transitioning away from oil. That includes fracking which is a big deal in PA as well. Biden should have spoken more calmly and at greater length on that topic, making it clear that there has to be a gradual shift to renewables, and starting with yadda yadda... create millions of good energy jobs yadda yadda...

Jim, it sure as shit feels like 4 years ago. God I hope I’m wrong. I’ll go re-read Patrick's posts and take a few deep breaths.

Chateau Vin · Post by **Chateau Vin** » Sat Oct 24, 2020 9:06 pm

I personally think Biden’s fracking response is not that big of deal... Also, I think AZ is more in play than NC or FL.

The horse has left the barn, and people have made up their minds. Now, it’s all about turnout, baby...I am more interested in the voter turnout and participation, especially in PA, MI and WI, come Nov 3....

jal · Post by **jal** » Sat Oct 24, 2020 9:13 pm

Dozens of trucks with Trump flags in Sarasota today, honking and slowing down traffic.
Voted this past week in person. Apparently 4.7mm voted early in Florida already.
I'll be glad when this is all over.

Racer Chris · Post by **Racer Chris** » Sun Oct 25, 2020 12:10 pm

Pence chief of staff tests positive for Covid
The pandemic is Trump's albatross.

JimHow · Post by **JimHow** » Sun Oct 25, 2020 1:14 pm

I'm enjoying watching Barack going medieval on Orange Head.

DavidG · Post by **DavidG** » Sun Oct 25, 2020 2:03 pm

JimHow wrote:I'm enjoying watching Barack going medieval on Orange Head.

Agree. What took him so long? Wish he’d started this a month ago.

Antoine · Post by **Antoine** » Sun Oct 25, 2020 8:58 pm

So clearly Bordeaux wine enthusiasts support Biden but are shit-scared Trump will win again... Well, they were less scared 4 years ago so some of them did not bother to vote... this time, the odds are the same but people know what the then unknown quantity would deliver...

My guess as a foreigner: Americans aren't that stupid... Trump to be beaten by a landslide. I bet a bottle of 2014 of Ducru.