(Cross-posted from Telescopic Turnip.)

“He—and he is almost always a he—is a venture capitalist who has analyzed the hospitalizations data! He is a growth hacker with a piercing view of race and measures of intelligence! He is an industry analyst with insight into viral spread! He is a lawyer exploding nuances of gender and sex!”

This is from Annie Lowrey’s classic essay about the Facts Man. The Facts Man is a specimen you will meet everywhere while roaming the corridors of the World Wide Web. He's looked at the raw data and concluded that the leading experts are sometimes wrong. Maybe you’ve even got your pandemic advice from one of them (you wouldn't do such a thing, would you?).

This is to be distinguished from Regular Crackpots, who gave a lot of thought to difficult problems in cosmology or quantum chromodynamics, and came up with their own alternative theory. They, too, came to the conclusion that the leading experts are wrong.

Why do so many people believe the Facts Man and subscribe to his substack, while the youtube channel of the Regular Crackpot channel only receives laughs and sneers?

Lutzs

In figure skating, you win by performing the most difficult tricks. The best figure skaters can do things like the quadruple lutz, a trick that’s outright impossible for most mortals like you and me. Only a handful of athletes in the world can perform it, and only for a short period in their lives.

Of course, you don’t need to do quadruples to enjoy figure skating, but you will not win the Olympics. Those who win the Olympics were raised in the best environment for it, had the best possible genetic background, and they still had to practice all their lives. If someone ever manages to pull out a quintuple lutz and does not completely screw up the artistic aspect, they will get the Gold.

Likewise, you will never make a groundbreaking discovery in quantum chromodynamics using freshman physics. You need to keep up to date with the current edge of knowledge, dig completely to the bottom of a particular problem, and derive some completely new maths just to make the smallest contribution. The best quantum physicists are the ones who can travel the farthest into quantum theory, and push the boundaries even further.

Panenkas

Football has no tricks. There are a few moves that require some technique, but a mediocre football player could probably reproduce most of Lionel Messi’s moves with enough training. With extra practice, even you – yes, you! – might be able to pull out a panenka. All of these were regularly performed by my schoolmates during recess (I know this because I could see them through the library’s window). This is why there are fiction films about Shaolin monks playing football: the film is entertaining because real football is not like that.

So how comes Messi keeps getting the Ballon d’Or, and you don’t?

Unlike figure skating, a great football player needs only a few basic techniques. But she needs to apply them consistently, relentlessly, keeping an eye on everything around, updating every second, improvising the perfect combination of these basic techniques at every moment.

Statistics are like football. The best statistics are done on clear, simple data and most of it can be done with a few simple formulas. But you need to apply them consistently, relentlessly, all day, everyday. You need to figure out how the data is structured, how to extract the answer to your questions, and what uncertainty remains. And that's only half of the way – like in football, the hardest part is reading the game. What's the social context? What are the incentives of the people involved? What are they trying to signal about themselves? What are you trying to signal about yourself? Why where you exposed to this particular piece of data in the first place? And you need to keep doing this permanently, for everything you read and everything you observe. If your facial expression when combing through data is not like Marco Verrati’s face on the picture above, you are doing it wrong.

Of course, there are fancy tricks in statistics, like all these mixed-effects hierarchical quadruple dirichlet kickflip regressions. This doesn't mean you are looking a good statistics, rather, it means that you are looking at bad data. The fancy regression techniques are just a way to tease some knowledge out of the wildest datasets. It's street football. But, being able to do stylish pannas for Youtube cameras is not enough to handle a live match with live teams and live supporters. It doesn't replace a profound understanding of the basics and a good intuition of how to apply them on the field.

Anyway, that’s my advice for how to think about The Facts. This, and fair-play. If you’re going to be a Facts Man, you should practice to be a good one and dribble them all. And if you’re already good at statistics, you can also reverse this advice and use it to become better at football.