The Pythia Principle

Or, A Brief History of Being Wrong on the Internet

The Oracle would definitely be an obsessive vague tweeter.

At the risk of alienating the sportsphobes among you, allow me to open with a brief anecdote of my youth. One of my favorite things as a kid was when there were rain delays during Philadelphia Phillies games, they would replay old video yearbooks, essentially a summary of a previous season. Although this role is now easily taken by YouTube – and many of these are available on it – as a Phillies obsessed kid this was like manna.

One of my favorite stories I learned from them was the tale of Jim Rooker, a broadcaster for the Pittsburgh Pirates. During one game against the Phillies, the Pirates started particularly hot. The game became so lopsided that he stated “If we lose, I’ll walk back to Pittsburgh.” Not insignificant, those 300 miles through the heart of Pennsylvania. You can probably guess what happened next.

Sometimes your mouth writes a check your feet can’t cash. To Rooker’s credit – and we’ll come back to him later – he did spend twelve days walking to Pittsburgh, although after the season was over and for charity. But he did it, paying the price for being spectacularly and publicly wrong. To put this in perspective in time, this was about three years before Francis Fukuyama wrote a book called The End of History and the Last Man.

---

After my initial draft of this – about a year ago – Certified Friend of the Stack Klaus wrote an article on why we should trust experts. This article raised some complicated questions for me regarding experts. Because identifying an expert is quite difficult. Oftentimes, people are considered experts because of some form of credentials or experience. As my fellow sports fans know, the main way someone gains the position Mr. Rooker had – commenting on sporting events – is through gaining expertise by playing, as he did. You also know that hearing them talk often betrays the falsehood of thinking experience grants expertise. And yet, experts are important.

It's a simple truth that none of us can be experts in everything. More to the point, none of us can even be experts in all the things we care about. We just don’t have that kind of ability. So, what we do is farm out our thinking to others. This is a process called distributed cognition. We have far more cognitive tasks than we can handle, so we rely on others to help us make decisions and evaluations and do our predictions for us. Perhaps my favorite expert in the world is a man named Ethan Chlebowski. Much like Mr. Rooker, we’ll come back to him later. But he serves an important role in allowing me trust his expertise on cooking, allowing me more time to practice law and bloviate on the internet.

There’s nothing new about distributed cognition. People would farm out their decision about which movies to see to Siskel and Ebert.¹  Or outsource their opinion on the Vietnam War to Walter Cronkite. Or, to go slightly further back than the television age, let the Pope make their spiritual decisions for them. What’s new is that because of the flood of information the internet creates, we have a near infinite supply of experts. This should make our life easier. And yet, when people talk about the internet, they talk about its wrongness. But ironically, nothing has ever made it easier to avoid wrongness – particularly false experts – than the internet.  

---

I once wrote about how being right is overrated. In that piece, I used the example of Cassandra. Let’s instead discuss Pythia, more well known as the Oracle of Delphi. When Croesus consulted her on going to war against Persia, she replied “If you go to war, a great empire will be destroyed.” Even if you’re unfamiliar with the story, you can figure out how this ends.

In the decade and a half I taught college students how to take standardized tests, I often struggled to have them grasp that on such exams there is no such thing as being wrong for the right reasons. Partially, this is because we intuitively understand that being right or wrong is not divorced from how you got there. History applauds the Oracle for being cryptically correct about Croesus losing his empire. But maybe she was just wrong?

The worst example of being wrong is that of Nellie Connally, an amazing woman whose life story you should read, but is mostly famous for what happened on a sunny day in Dallas, November 22^nd, 1963. While her and her husband – the Governor of Texas – were riding in a car with President John F. Kennedy, she turned to him and uttered the words “Mr. President, you can’t say Dallas doesn’t love you.” You may be familiar with what happens next.

Imagine having said that. Imagine the strength to not have spent your life behind a vow of silence. No one has ever been more quickly and terribly wrong than her – except perhaps JFK himself who responded “No, you certainly can’t.” Yet, other than Lee Harvey Oswald and his CIA handlers no one could have possibly foreseen this sequence of events. There’s nothing shameful about being wrong under such circumstances.

But there’s another way to be wrong, one that should discredit you. The popular example of this is Neville Chamberlain, who declared after the Munich Conference that it was “peace in our time” shortly before the most terrible war our species has ever seen. I think it’s the incorrect world war to use as an example.²[1] More obviously wrong was how just about every European leader was wrong about World War I. And not just once. Over and over again, the political and military leaders were wrong. The boys would be home by Christmas. But this was wrong, fatally so.

This also did not require hindsight. It took a willful disregard of the previous half century of military history to believe that the new technologies would not produce a lengthy bloodbath. They weren’t wrong because they failed to foresee a Black Swan. They were wrong because they refused to recognize a White Swan. Many wise people at the time saw how disastrous it would be, but this was ignored, and for over a century we’ve been dealing with that wrongness. They weren’t just wrong, they were deservedly wrong.

Let’s go back to our friend Jim Rooker. There have been fifteen times between 1903 and 2022 when a Major League Baseball team took a 10-run lead in the top of the first, as the Pirates did. Fourteen of those teams won. His prediction was, like Mrs. Connally’s, perfectly reasonable. And yet they became famously wrong.

Experts can be wrong as frequently as fools can be right. How they get there is important. And this is our good news. The internet is forever. Never before in human history has it been this simple to find someone’s track record and reasoning. We should be living in the Golden Age of Accountability.

This leads us to the paradox that I’m naming after myself: Dan’s Paradox. We created a system that keeps a record of everything we say and yet being correct has not become important.

This is a baffling state of affairs, yet it is true. Find almost any prominent political commentator in America and Google “[their name]” AND “Iraq” OR “WMD.” For a funnier example, take Jim Cramer. Let’s be clear about two things. One, Jim Cramer is supremely entertaining and his job is incredibly difficult. Two, he’s a Philly guy and an Eagles fan so I’m not going to say anything bad about him. That said…

This was proven disastrously wrong within a week. And no one cares! And this isn’t silliness like Skip Bayless being repeatedly wrong and yet being paid millions for it. This is something that matters!

Politics, finance, sports, it doesn’t matter. The world is replete with people being heavily rewarded for being not just wrong but deservedly wrong. It is perhaps ironic that two of the movements most associated with predictions are led by a man who famously failed to forecast Donald Trump and one who was wholly taken by Sam Bankman Fried.³ If even the dudes who are all about predictions don’t suffer from being wrong, how do we expect anyone else will? What are we, as normal people, supposed to do when it comes to finding experts?

Let’s go back to Ethan Chlebowski. Do you know why when I named my personal favorite expert I named a YouTube chef? Because if a guy tells me how to pickle my own onions, and I do it, and they taste terrible, I’m going to know they taste terrible.⁴ The proof of the pudding, after all, is quite literally in the eating.

And this is where I find the solution to Dan’s Paradox. On something like “what will I eat for dinner” I care deeply about the rightness of my expert. But on more ephemeral matters, does it matter? Most likely, what I want is not correctness, but certainty. I don’t care about if the person I’m listening to has a track record proving expertise, I want someone with a halo of expertise to confirm what I already think. What we’re looking for aren’t experts with track records of excellent thinking. We’re looking for experts who do a good job or telling us what we want to hear.

I had a long running joke with a much smarter friend of mine that the power of tribalism and leadership cueing are so strong that if aliens showed up, within a couple weeks people’s opinions on them would match their partisan affiliations. We declared ourselves correct about that sometime in 2020.

You may have noticed over the last few years that a lot of people condemn how much wrongness and fake experts there are on the internet. I’ve defended the internet against similar accusations in the past. And I will continue to do so. The internet doesn’t create wrongness, it doesn’t even amplify it. The internet amplifies everything, rightness and wrongness. If anything, the internet has made it simpler to identify wrongness. We just don’t care.

It’s popular to blame technologies for their bad effects. And there are times I’ve done that here. But it’s often the case that the fault, dear Brutus, is not in our silicon but in ourselves. This is one of those. Look back at the examples I used about being wrong. I used examples from 1914, 1938, 1963, 1989, and 1993. Notice anything about those dates? Even the two I used from the Internet Era were about television and newspapers. We don’t just not care about whether experts are right, we’ve never cared. When we thought that snake oil salesman made some great points, we were fine with it. Being able to Google “is snake oil helpful?” would not have made much of a difference. Go look at that Athletic Greens container in your cabinet and tell me it would.

We need to take personal responsibility. There is no algorithm that will relieve us from our duties to think. If you care about experts being right, you have the greatest research tool ever created. Be skeptical and take advantage of it. I suggest all of us could be better about this and it would make a better world. But if you’re not willing to do that, perhaps just admit that none of us really care about rightness or wrongness. We just want to hear what we want to hear.  

1. 12-year-old me still appreciates their recommendation for Quiz Show.

2. And perhaps a writer more talented than I can opine on why this example, and not the one I will use, is preferred.

3. This is somewhat unfair to both. I don’t think Scott Alexander’s failure to realize SBF was a scam artist discredits him, but his post explaining why he was fooled discredits much of the intellectual framework behind the prediction market craze. And I think Nate Silver gets too much criticism for variance screwing him in 2016. But he didn’t turn down the millions he’s made when variance benefited him in 2008, so, live by the sword, die by the sword.

4. There’s a reason I embedded that video earlier, they tasted amazing and were incredibly easy. Pickle your own onions people.