Bayesians for Biden

This article caught my eye:

Dave Dahlstrom has always voted red.

But an upclose experience with the coronavirus has made the U.S. Air Force vet vow to vote blue this time around. His wife, Cindy Dahlstrom, died of COVID-19 at age 72 on July 6.

Recent revelations by journalist Bob Woodward, documented with taped interviews, indicating that President Trump knew how lethal SARS-CoV-2 was and played it down “to avoid panic,” have made the grief of the bereaved all the keener.

Non-Bayesians seem to be sticking with the Donald:

Just how many people will need to die before minds change is a question Pontes ponders. After all, 200,000 American deaths has been insufficient to create a shared American reality. Only weeks ago, one of the loudest protestors in Shasta County, a businessman who had refused to take steps to prevent the spread of the disease, had watched his mother die of Covid. In that moment, a political opinion was challenged by a fact; one of them needed to be altered. The man called the coroner and demanded that the county change the cause of death.

And this also caught my eye:

“We’re going win the state of Minnesota because of her, they say,” Trump said of Omar. “She’s telling us how to run our country. How did you do where you came from? How is your country doing?”

The comments were the latest personal attacks Trump has levied against Omar, who arrived in the United States with her family after fleeing war-torn Somalia. She is a U.S. citizen.

A crowd at one of the president’s North Carolina rallies last year chanted “send her back” about the congresswoman as Trump looked on.

He has also accused her of being anti-Semitic because of her views on the conflict between the Israelis and Palestinians, and the president tweeted last year that the four “squad” members, all women of color, should “go back” to their countries. All four are U.S. citizens, and only Omar was born outside the U.S.

Scott Adams seems increasingly Bayesian.

In the same publication, this caught my eye:

Still, the debate doesn’t appear to have had a significant impact on the state of the presidential race. Only 2 percent of respondents said that it changed how they will vote, while 98 percent said it would not.

For the sake of argument, suppose that this 2% shifted from Trump to Biden (polls shows people thought Trump did poorly). Let’s say Biden goes from 3% up to 7% up. That sort of shift is not “significant” in a close race? Really? I’m not a political scientist, so maybe someone who knows more about politics can explain to me why losing several million of your supporters to the other guy is not significant in a closely fought race.

The FT has an amusing article showing how Trump is making Chinese exports great again.

PS. What does it say about the NYT that they thought this opinion piece editorial was worth printing?

Update: I wrote this post before the recent news:

Since joining his campaign in early 2015, Hicks was often at Trump’s side. She returned to his administration in spring, following an earlier spell as press secretary.

None of the entourage were seen wearing masks while getting off the plane. Hicks was spotted climbing into a staff van together with Bill Stepien, the president’s campaign manager, the New York Times reported. . . .

Speaking to Biden directly, Trump suggested his rival’s precautions were over the top. And ridiculous: “I don’t wear a mask like him. Every time you see him, he’s got a mask.

Not Bayesian.

Update#2: I noticed something weird. In response to the Trump/Covid story, the odds of the Dems taking the senate jumped by much more than the odds of the Dems taking the presidency. Are those markets very liquid?



21 Responses to “Bayesians for Biden”

  1. Gravatar of Michael Sandifer Michael Sandifer
    2. October 2020 at 09:05

    The race seems to have shifted in Biden’s favor this week, but there’s still tremendous uncertainty. Presumably, Trump’s diagnosis means his campaigning will be curtailed at the worst possible time, and Democrats were already vastly out fundraising and outspending Republicans all over the country.

    Has this week shifted your opinion about who’s likely to win?

  2. Gravatar of gt gt
    2. October 2020 at 09:23

    > For the sake of argument, suppose that this 2% shifted from Trump to Biden (polls shows people thought Trump did poorly)

    I agree that Trump did poorly, but my guess would be that of that 2%, something like 60% was shift to Biden and 40% was shift to Trump, meaning a total shift of only 0.4% of the population. Still a pretty large number, but unfortunately not nearly as much as if everyone shifted over.

  3. Gravatar of John Hall John Hall
    2. October 2020 at 09:26

    The Bayesians not-for-Trump camp encompasses both the Bayesians for Biden camp and the Bayesians for neither camp. I’m solidly in that last one.

  4. Gravatar of Marc Marc
    2. October 2020 at 09:29

    From the article with the 2% statistic:

    “The CNBC-Change Research poll surveyed 925 likely voters nationally from Sept. 29 to 30 and has a margin of sampling error of plus or minus 3.22 percentage points. Findings on the presidential debate are based on responses from 796 likely voters, with a margin of error of plus or minus 3.47 percentage points.”

    Doesn’t this mean that 2% is not significant because it might well be .01% and based on this survey we just don’t know?

  5. Gravatar of ssumner ssumner
    2. October 2020 at 09:37

    Michael, It was certainly a bad week for Trump. Let’s see how this Covid-19 story plays out before I reconsider my prediction (which is pretty worthless in any case–the betting markets are more reliable.)

    gt, I don’t agree. Even if polls show X won a debate 60-40, the marginal effect is likely to be strongly skewed. It’s hard for me to imagine a Biden supporter watching Trump’s appalling performance and switching to Trump. Yes, it’s not 2% all one way, but I believe the net move was overwhelming toward Biden. That’s if the poll is accurate. Of course it’s probably slightly wrong, as a few people will claim to switch that did not really switch, so the net swing in the margin is obviously less than 4 points. On that we agree.

    John, I’m voting for Jorgensen.

  6. Gravatar of ssumner ssumner
    2. October 2020 at 09:46

    Marc, Yes, probably not statistically significant. But I still say the 2% move would be significant if true. Note that the term ‘significant’ has two meanings, statistically reliable and meaningful if true. So the wording of the article is still inaccurate.

    They should have said we don’t know if it had a significant effect on the campaign, not that it didn’t.

  7. Gravatar of stoneybatter stoneybatter
    2. October 2020 at 10:13

    Because words matter, that NY Times link is to an opinion piece, not an editorial. The latter are written by the NY Times editorial board, the former by contributing authors. This is a not-insignificant difference.

  8. Gravatar of Michael Rulle Michael Rulle
    2. October 2020 at 10:42

    yay—-of course Scott has implied this many times, but I am glad he finally came out speciically for Jorgensen. I knew there was no way he could vote with those nut jobs (:-)). He just finds Trump worse.

    The instant post debate repsonse was highly negative for Trump. In fact, I was about to turn it off within 30 minutes—I could not take it. But I am a bad watcher. I tend to focus on the guy I am rooting for–almost to the complete exclusion of the other guy. I think this is one reason Andrew Sullivan thought Trump was winning—he was focusing on Biden.

    Interestingly, when you read the debate it seems more normal—although it was a total “meme” debate. Trump had 2 or 3 substantive runs which were good. Biden, not so much—not answering the Supreme Court question was bad for ………who? Don’t know–I guess no one-. except the anti banana republic crowd which I am close to asking for my memebership card

    Trump getting COVID is bad novel material—yet here we are. If he even gets Boris Johnson sick (it was gone in 12 days but he had intensive care for 3) he has a higher chance of losing (I think) because ………”it makes him seem more wrong” (not sure really). If he gets better quick—what does that mean? Presidents get better care than old Black men in Queens?

    I know nothing—but I have not joined Scott yet in the Howard Hughes hotel—-I wish I did—I want to–maybe soon.

  9. Gravatar of Christian List Christian List
    2. October 2020 at 11:15

    I also don’t quite understand why Scott thinks that it’s an editorial. It is by Regina Ip, the founder and leader of the NPP, a pro-CCP party from Hong Kong. Maybe it was on the front page?

    Anyhow, her argumentation is perverse. The democracy movement came into being because the CCP ignored “one country, two systems”. The CCP wanted to achieve complete control over Hong Kong, which it has now basically achieved. Ip completely reverses the causality.

    Apart from that, her argumentation is frighteningly similar to Scott’s: Hong Kong is part of CCP China, so the CCP can do whatever they want. There must be no sanctions, and the West should not interfere.

    The CCP will be telling the same bogus story when they bring Taiwan home to the reich; and after that they will be drunk with victory, so it will be interesting to see what will happen then. History might give a few pointers.

  10. Gravatar of Danny Danny
    2. October 2020 at 12:27

    $140M bet in the Next President market
    $318K bet in the Senate Control market

    Do you really think Trump will still win or are you attempting to manage your own expectations so as not to be overly disappointed?

  11. Gravatar of ssumner ssumner
    2. October 2020 at 12:54

    Stoneybatter, Thanks, I corrected it.

    Christian, You said:

    “Apart from that, her argumentation is frighteningly similar to Scott’s:”

    Is there no limit to the extent to which you are willing to make a fool of yourself? She is as far from my views as it is possible to be.

    Danny, See my reply to Michael. And thanks for the info on trading volumes.

  12. Gravatar of Christian List Christian List
    2. October 2020 at 14:00


    yes, I may have summarized it in a cynical way. Her ideology is fundamentally different, that’s true. But the result remains the same: emphasis that Hong Kong is a part of China (nobody doubts that), rejection of sanctions, leaving all privileges of Hong Kong regarding the West in place, no termination of special trade treaties, no relevant interference from the outside. The only relevant difference in consequences I see so far is that you would support exiles and she would probably not. But that’s about it.

    Sure, you would also raise your index finger against China and say, “Na, na, na, stop doing that”. But that’s not a relevant difference since it contains no consequences and only consolidates the regressions.

  13. Gravatar of ssumner ssumner
    2. October 2020 at 14:03

    Christian, More lies. I never said I opposed sanctions in response to China’s action in HK. I said I opposed sanctions against the victim, not the villain.

    Please stop lying.

  14. Gravatar of Dale Doback Dale Doback
    2. October 2020 at 14:11

    That Scott Adams interview is fascinating to me. Trump’s racist dog whistle schtick was pretty obvious from the early stages, yet somehow his comments in the last debate were the final straw. I still thought Trump could shoot someone and he wouldn’t lose any votes.

  15. Gravatar of Mark Z Mark Z
    2. October 2020 at 14:50

    I’m not sure an anecdote about someone changing their opinion based on an anecdote constitutes Bayesianism (unless it’s a super meta joke), but in any case as of now Trump’s chances have fallen by about 10 percentage points while the senate is about where it’s been for awhile.

  16. Gravatar of Benjamin Cole Benjamin Cole
    2. October 2020 at 16:56

    The NYT Hong Kong op-ed might be worth publishing as a sanitized window into CCP views. So, now you know, and this is the kids glove version.

    On the other hand, perhaps craven capitulation to the CCP is the new norm, see Disney, NBA, Apple, Walmart, BlackRock etc. Maybe the NYT thinks the CCP is right on Hong Kong.

    Westerners can shift supply lines into India, Taiwan, Korea, Thailand and Japan. Makes sense to me. The world’s largest smartphone factory is in India already.

  17. Gravatar of ssumner ssumner
    2. October 2020 at 18:21

    Dale, Some people just don’t hear dog whistles. If Trump doesn’t use the N-word publicly (he uses it privately) then people don’t see the racism.

    Mark, Not at this site:

    Benjamin, You said:

    “Westerners can shift supply lines into India, Taiwan, Korea, Thailand and Japan. Makes sense to me.”

    Fortunately, important economic decisions are not made by you, they are made by people who have some understanding of business.

  18. Gravatar of BC BC
    2. October 2020 at 19:37

    “What does it say about the NYT that they thought this opinion piece editorial was worth printing?”

    To be fair, just because the NYT printed an opinion piece, that doesn’t mean that the NYT agrees with it. At the bottom, they explain, “The Times is committed to publishing a diversity of letters to the editor.” So, the NYT wants to print a range of views, as long as it’s not written by a Republican Senator or an American conservative, of course. Notice, they also link “diversity of letters” to another NYT letter that claims NYT’s letters to the editor “skew male”. It wouldn’t surprise me if authoritarian’s letters skew even more male. So, the NYT wants to make sure that pro-tyrant women are properly represented.

  19. Gravatar of ssumner ssumner
    3. October 2020 at 08:48

    BC, I like the way you start out “To be fair” and then trash the NYT much more harshly than I do. 🙂

  20. Gravatar of Benjamin Cole Benjamin Cole
    3. October 2020 at 18:17


    Reliance (India’s largest company) and Google are teaming up to produce a smartphone, using all Indian IP, at what Reliance’s Chairman says will be half the cost of rival smartphones. They say they are putting $4.5 billion into the project.

    But then, perhaps Reliance and Google do not understand business. Or Samsung, which already operates the world’s largest smartphone manufacturing plant in the world…in India.

    I think this is just starting. I sure hope so. Modi is no gem, but compared to the CCP he is okay.

    But in one regard, you are right. I don’t understand business. Craven capitulation to the CCP is not something I understand.

    Moreover, diversifying supply lines seems like smart policy regardless of immediate circumstance.

  21. Gravatar of ssumner ssumner
    5. October 2020 at 14:40

    Ben, You said:

    “Modi is no gem, but compared to the CCP he is okay.”

    Easy to say if you are not a Muslim living in India.

Leave a Reply