22 Comments
User's avatar
Tad Thurston's avatar

Here are a couple predictions from astrophysics that blow my mind:

In the '40s and '50s, assuming the Universe was something like 10-15 billion years old and that General Relativity worked, Gamow and his students predicted a leftover heat of 5-10K. The real number is 3K, but that was a crazy awesome prediction for the leftover heat of the Big Bang that hadn't been discovered yet.

The other one that doesn't get enough press is the prediction of what elements formed in the first minutes of the Big Bang -- 75% H (by mass), 25% He, 1 atom of deuterium per 100,000, and 1-2 atoms of lithium per 10 billion (along with others). These are specific predictions of what we should see in primordial gas clouds covering 12 orders of magnitude, and it's *what we actually see* with small errors. That it's anywhere in the ballpark for all these elements simultaneously is just astonishing.

Andrew Smith's avatar

These are great! I don’t know if you can call it a prediction or not, but 2.00231930436… as the value for g is really, really impressive.

Tad Thurston's avatar

Oh, absolutely, that's a great one. Not many predictions are within a ten-trillionth of the measured value!

Andrew Smith's avatar

I mean, it's *pretty* close.

David Perlmutter's avatar

"Recency bias" plays a major role in many predictions not working out, because none of us can truly predict or manipulate future events. There are occasions when an individual or a product does live to the prediction of its enduring potential, but that has nothing to do with an early prediction of same.

Which is why the majority of these predictions are merely hype.

Andrew Smith's avatar

What about the tiny minority? Got any favorites from history?

David Perlmutter's avatar

Bobby Orr and Wayne Gretzky both became the pro hockey legends everyone at the start of their careers said they were going to be.

Andrew Smith's avatar

This makes me think of Babe Ruth's supposed "called shot" - he pointed to the bleachers and then hit a home run on the next pitch, supposedly.

David Perlmutter's avatar

Yes. He’d be playing long enough by that time to actually be able to do that.

Andrew Smith's avatar

On the other hand, nobody knows if he really did, apparently!

David Perlmutter's avatar

The Babe was an unpredictable guy- a lot of the time you never knew if he was going to show up for a game drunk or sober.

Bill C. White's avatar

The most precise physical prediction in history is the electron anomalous magnetic moment (\(a_{e}\)). Theoretical predictions from Quantum Electrodynamics match experimental results to 11 significant digits (about 0.16 parts per trillion), making it the most stringently tested theory in science.

Andrew Smith's avatar

Physics always feels like cheating in situations like this.

I mean, you’re right, of course!

Bill C. White's avatar

Hahaha!

Bill C. White's avatar

I worked with classification my entire scientific career. A confusion matrix is helpful to see different errors.

https://www.wikiwand.com/en/Confusion_matrix

Oh and "skin the game" is the true indicator of a serious prediction. See Nassim Taleb.

https://www.wikiwand.com/en/Nassim_Nicholas_Taleb

Andrew Smith's avatar

I agree. If you don’t have any skin, you’re speculating. Nothing wrong with that - but there needs to be a distinction drawn.

Andrew Smith's avatar

Yep, that’s a good one. How about Michael Burry?

mc's avatar

Also good but very early

Andrew Smith's avatar

Great point. Maybe Einhorn? Was he a bit later? Either way, I like the finance angle here.

Andrew Sniderman 🕷️'s avatar

I predict every social platform will build native or integrate existing prediction markets

https://support.substack.com/hc/en-us/articles/28879761546260-How-do-I-embed-Polymarket-odds-on-Substack

Andrew Smith's avatar

They are everywhere, aren't they?