The Pundits curse - The more novel and the more declarative your interpretation, the more attention, but also the more likely you are to be wrong

Two pieces this morning which are something of a study in contrasts.  The first is I was wrong about Biden by Matthew Yglesias and the subheading is The worst feeling in the world.  The second is Blaming the media is what got Democrats into this mess by Nate Silver.  The subheading is Reality comes for everyone — and now it's come for Joe BIden.

Nate Silver has been sounding the statistical alarm on Biden's age/condition as far back as September 2023 but certainly since February I think it was.  And has taken immense professional heat for doing so.  The legacy mainstream media newsroom is something of a monolith.  There is some diversity in its various left leaning enthusiasms but it is near monolithic in its aversion to anything Classical Liberal and Empirical Rationalist.  I.e. anything to do with emergent order, populism, free markets, free people, etc.  

Silver has always been something of a fish out of water in that he is actually a reasonably rigorous empirical rationalist.  An understanding of statistics is weak at the public level and non-existent among basically innumerate journalists but he has successfully persisted in doing the best job possible despite all the criticisms.  

He is in the classic position of any statistician in that he calls a hundred races with a 70% chance of the candidate winning and then 30% of the time the candidate loses and all his journalist peers offer that as obvious evidence that he just basically wrong.  

Silver has, since the debate, offered several increasingly detailed and rigorous assessments of the presidential race, none of them favoring a Biden win.  He has been even handed in that he has been identifying what the model forecasts mean for the Democratic Party and the options it confronts.  None of them pretty.

Interestingly, the more challenging becomes the forecast, apparently the harsher becomes the criticism of Silver by his peers.  Each technical piece seems followed soon after by ever more explicit efforts at professional damage control along the lines of "I'm a Democrat who hates Trump.  Don't crucify me as the messenger of the statistical doom."

Blaming the media is what got Democrats into this mess by Nate Silver is the most recent and most explicit.  I have always thought of Silver as a technician.  I have assumed that he probably leaned left but that his first commitment was to reality as he understood it through his models.  And I still think that.

But it appears that he is taking a lot of hits for delivering empirical reporting which is disfavored by his peers and therefore he is having to become more and more explicit about his support for the Democratic Party.

Here is his ginger message.

But here’s the new reality: Trump is probably going to become president again. Because I still do believe in empiricism and probabilistic thinking, I want to be clear that this is by no means certain. In our model, Trump’s chances are “just” 70 percent — and the model makes two assumptions that may not map well to the real world. One is that it offers what is technically a conditional prediction — an assessment of the odds if Biden remains the Democratic nominee (and Trump remains the Republican one). That is a tenuous proposition: at betting markets, Biden is considered more likely than not to exit the race, and I believe those markets probably underestimate the chance that Biden will drop out.

But also — and less favorably for Democrats — the model makes a lot of implicit assumptions about Biden’s fitness for running a campaign that probably do not apply in this instance. There’s no parameter in the model for “guy who wants to be president until he’s 86 but could barely form complete sentences.” Instead, Biden’s condition at the debate, and in some other recent public appearances, would be highly worrying if you encountered it in an aging grandparent. At the very least, you’d encourage them to undergo neurological testing, something Biden — like a lot of stubborn people at his age — has refused to do. It’s time to confront reality: wishing you could have the old Biden back won’t be any more effective than wishing you were 17 years old again.

The balance of the essay consists of two arguments.  One is a plea that he and his model could be wrong.  The other is that the challenge for the legacy mainstream media is that is subject to epistemic closure.

Sure, the world is unpredictable, stochastic — hence the need for a probabilistic approach to assessing political and human affairs. But liberals, progressives, Democrats — whatever you wanted to call us — at least we were trying to get the right answer rather than succumb to the sort of postmodern relativism that Rove engaged in. We stood for Facts, Data, Empiricism: reality was on our side.

The years between roughly 2004 and 2020 could hardly have been more validating for this hypothesis. Rove’s Iraq War proved to be a disaster, and Barack Obama was elected in a landslide in 2008 in response — and then again in 2012, culminating in Rove’s meltdown on Fox News on election night. And although Donald Trump was nominated by the Republican Party and elected in 2016 — the clearest sign yet of the modern American conservative movement’s disconnection from reality — he got his comeuppance in 2020, when Joe Biden defeated him and Trump became the first incumbent to lose re-election since Bush Sr. in 1992.

This is as explicit as I have seen Silver be in terms of pledging fealty to the Democratic Party.  His aversion to Rove/Bush and admiration for Obama may be true but probably won't hold up in the long run.  My own interpretation is that Bush's War on Terror was necessary and tragic and suffered multiple tactical shortcomings but was strategically defensible.  In contrast, the adulation of Obama by the legacy mainstream media was always and in every way misplaced.  His centeredness, his poor grasp on policy and his lack of strategic vision meant that he was a tactical disaster for the Democratic Party and a strategic disaster for governmental and cultural institutions.  An assessment not yet in the ken of the pundits.

Silver has a number of themes which show up in Yglesias' essay as well.  Epistemic closure.  Fealty to Democratic Party niceties.  The role of the legacy mainstream media in suppressing news in pursuit of its partisan objectives.  

For an Empirical Rationalist, Silver seems to be increasingly desperate to protect his Democratic Party preferences to his Democratic Party colleagues.

Yglesias has always been much more a straight out Democratic Party activist.  He dabbles in factual reporting (and sometimes has some good material) and he seems to want the mantle of empirical rationalism but he is always first and foremost a party operative.  

 I was wrong about Biden by Matthew Yglesias is both intriguing and entertaining.  It is a clever essay attempting to cover why he was wrong about Biden and why Yglesias takes a diametrically opposite position to his position before the debate.  Essentially he argues that he was inside the bubble and misread the signals.  

Which is true enough.  But most Independents, Moderates, Classical Liberals, and Conservatives have always known that, and said so.  

The truly interesting, and entertaining thing in the essay, is that Yglesias' own contradiction.  Yglesias is apparently in the same dangerous position as Silver with regard to his Democratic tribe.  On the one hand he has to acknowledge to his reading public that he was markedly wrong on a critical issue (an issue on which his partisan opponents were right) and at the same time he has to reassure his Democratic brethren that he is still a member of the tribe.

His is the more challenging task though, in comparison with Silver.  It has always been possible to discern Silver's predicate beliefs but his technical work is pretty good.  Silver usually stands by his models whether they give him the answers he might prefer or not.  There is a clear basis to his empirical rationalism.

That is much less the case with Yglesias.  Yglesias does bandy around empirical facts but he is far more prone to being selective in those facts and definitely in their interpretation.

In his apology essay, Yglesias excoriates his error in missing the signals.  He acknowledges that he suffered epistemic closure (my words, not his).  That he and his legacy media pundit colleagues need to do better.  But he no sooner acknowledges his failure due to closure than he demonstrates he remains entirely committed to that closure.  

I do think that Trump has one upside for the GOP relative to Scott or DeSantis, namely that he has been willing to distance himself more from the anti-abortion movement. But if Nikki Haley were the nominee, she’d be crushing Biden right now and I think that’s kind of obvious. Am I going to write “Trump should step aside so the GOP can nominate Nikki Haley and crush Biden” as a take? Of course not. Because I’m a Democrat, and while I hate Trump, I also don’t want Haley to crush the Democrats.

If Trump had just retired quietly after January 6, 2021, I’d have breathed a huge sigh of relief for the future of American democracy.

But I’d still be a Democrat who cares about the interests of poor people and a woman’s right to choose and who thinks it’s a bad idea to enact budget-busting tax cuts for the rich. Indeed, I am still a Democrat today and I’ll vote for Biden over Trump any day of the week. Trump is a criminal and an insurrectionist, but I also just don’t agree with Republicans’ policy ideas. What I would like is for the Democrats to beat Trump.

I fully accept that Nikki Haley is perhaps the Republican most amenable to the broadest scope of Democratic pundits.  But there seems no possible world in which she could be a candidate opposing Biden and if she were that she would be crushing him.  She would have to gain more cross-over Democrats than the Republicans whom she would lose and I don't think there is any polling evidence or other argument that to sustain that view.

Then you have Yglesias' self immolation in order to seek grace from the Democratic mob.  My characterization of his confession:

I am a fervent Democrat

I hate Republicans

I hate Trump

I put my party's well-being above the well-being of my nation (I want my party win, not who I think is the best candidate)

I believe the myth that our democracy is at risk from Trump despite all the First Amendment transgressions and the lawfare of my candidate.

I don't want to cut taxes but I am fine with historic deficits.

I believe the myth that there was an insurrection despite all DoJ investigations having found otherwise.

I believe the felony charges which every independent lawyer expects to be overturned are actually true.

Its possible, but I don't think I am mischaracterizing him.  And if the above summary is true, then Yglesias is essentially saying

I made a mistake that you caught me making.  Sorry,  Not sorry.

Ugh.

Between the two essays though, there are a handful of things which come through and which I do think are pertinent and worth pondering.  The debate brought light to:

The epistemic closure of the legacy mainstream media.  This was already well established, it just became inescapably obvious.

Both Yglesias and Silver, in different ways and in different words, raise the important issue of Framing and Confirmation Bias.  That how we initially interpret the world inherently makes us prone to confirmation bias.  

The idea of preference cascades are floating around in the background.  As every major legacy mainstream media narrative collapses (Biden is still sharp, Russiagate,  Russian Disinformation, Hunter Biden Laptop, Covid-19 was the product of a wet market, the vaccines worked, the vaccines had no negative side-effects, AGW threat, etc.), there comes a point where either the reversal is buried or is embarrassingly transparent in a short time frame.  

The absolute need for humility in forecasting and the incentive structure against such humility.  That is most obvious in Yglesias' essay.  He admits to being wrong and says he is going to be more alert.  But it is pretty obvious that there is a conflict.  To be a pundit who gets attention, you have to be novel and declarative.  The more novel and the more declarative, the more attention but also the more likely you are to be wrong.  It is clear humility is needed but nobody, other than Silver, seems to be either acknowledging that or signing up for it.