Markets can’t be predicted because they read your predictions. So what about a system that can’t? Part 5 of the No Crystal Ball series.
The Short Answer
The reason markets resist prediction is reflexivity: participants observe, adapt, and change the system being measured. The natural next question — the one that pulled me out of finance and into a laboratory — is whether there’s a system without that property. There is. Matter doesn’t read your research. A crystal doesn’t change its behavior because you published a paper about it. In non-reflexive systems, the observer problem vanishes, and honest prediction becomes possible again. That realization is why a markets person ended up doing physics.
The Sentence That Wouldn’t Leave Me Alone
Back in Part 3, I wrote a line almost in passing:
“The physicist can study electrons confident that the electrons aren’t reading physics journals. The trader cannot.”
I meant it as a contrast — a way to explain why markets are uniquely hard. But the sentence kept working on me after I wrote it. Because if the whole problem with prediction is that the system adapts to being predicted, then the sentence isn’t just a contrast. It’s a map. It points at exactly the kind of system where everything I’d concluded about the futility of prediction simply… doesn’t apply.
I’d spent years learning, the hard way, why you can’t forecast a reflexive system. The flip side of that lesson is a gift: you absolutely can forecast a non-reflexive one. And the universe is full of them. They’re called the physical sciences.
What Reflexivity Costs, and What Its Absence Buys
Let me lay the two worlds side by side, because the symmetry is the whole point.
A market is a reflexive system. You find an edge, you publish it, traders pile in, the edge evaporates (Goodhart’s Law). You forecast a rally, the forecast spreads, the rally front-runs itself, the timing breaks (reflexivity). The observer is a participant. Measurement changes the measured. Prediction is self-referential and collapses under its own weight. Everything in Parts 1 through 4 follows from this single property.
A crystal is not a reflexive system. It has a structure. That structure has properties. Those properties are fixed by the laws of physics, and the laws of physics do not care what you believe about them, do not read your papers, and do not adapt their behavior because a clever person figured out a pattern. You can publish your finding to the entire world and the crystal behaves exactly the same the next day. The edge doesn’t evaporate. The pattern doesn’t self-destruct. The thermometer keeps reading true.
In one world, knowledge is a depreciating asset — the moment you have it, the market starts taking it back. In the other, knowledge is a permanent asset. You learn how a material works, and that’s simply how it works, forever, for everyone, regardless of who knows it.
Coming from markets, that second property is almost vertiginous. An edge that doesn’t decay. A prediction that doesn’t poison itself. After years of watching every advantage erode the instant it was understood, finding a domain where understanding stays understood felt like stepping off a treadmill onto solid ground.
The Same Discipline, Pointed Somewhere New
Here’s what people get wrong about a move like this. They assume going from finance to physical science means throwing out the toolkit and starting over. It’s the opposite. The toolkit transfers almost perfectly — because the discipline was never really about money. It was about separating signal from noise in a complicated system.
- Signal versus noise. A market is mostly noise with a little signal buried in it. So is a raw scientific measurement. The skill of refusing to be fooled by a pattern that’s just luck — of demanding that an apparent edge survive an honest null test — is exactly the same skill, whether the data is price ticks or spectra.
- Regime detection. Markets have regimes; so does matter. Knowing which regime you’re in before you trust any model is the same move in both.
- Factor thinking. A market signal is a weighted combination of underlying factors. So is a material’s behavior. The instinct to decompose a complex outcome into the few variables that actually drive it doesn’t care what the outcome is made of.
- Intellectual honesty about what you don’t know. This is the one that matters most, and it’s the through-line of this entire series. The discipline of saying “the data supports this much and no more” is the same whether you’re briefing a portfolio committee or evaluating a physical prediction. Most people can’t do it in either domain. The few who can are dangerous in the best way.
I didn’t become a different kind of thinker. I pointed the same kind of thinking at a system that finally rewards prediction instead of punishing it.
Why This Isn’t a Career Change Story
I’m not telling you this because it’s autobiography. I’m telling you because it completes the argument the series has been building.
The whole “No Crystal Ball” thesis was: don’t predict reflexive systems — describe them. That’s correct, and it’s the right posture for anyone in markets. But it leaves an obvious door open, and intellectual honesty demands you walk through it: if the problem is reflexivity, then go find the systems that aren’t reflexive, and in those, predict your heart out. Description is the honest move in a market. Prediction is the honest move in a physics lab. The skill — reading structure, killing your own false positives, knowing the limit of your evidence — is identical. Only the substrate changes.
The crystal ball was never the problem. The problem was pointing it at people, who look back. Point it at matter, which doesn’t, and it turns out you can see quite a lot.
Key Takeaways
- Reflexivity is the root cause of why markets resist prediction — the system adapts to being observed.
- Non-reflexive systems don’t have that problem — matter doesn’t read your research and doesn’t change because you published.
- In markets, knowledge is a depreciating asset; in the physical sciences, it’s permanent — an edge that doesn’t erode.
- The discipline transfers intact — signal-vs-noise, regime detection, factor thinking, and honesty about the limits of evidence are the same skills, whatever the substrate.
- Description is the honest move in a reflexive system; prediction is the honest move in a non-reflexive one — same rigor, different domain.
What’s Next
In Part 6: “The Same Lens, Pointed Down,” I close the series.
If the same disciplined mind can read a market and read a material, what does it actually look like to run that toolkit at full power on a hard physical problem — and what does the markets world have to teach the science world about not fooling yourself? The lens that learned humility in the most humbling system ever built turns out to be exactly the lens you want for the problems that finally reward getting it right.
[Read Part 6 →]
Frequently Asked Questions
Why can physical systems be predicted when markets can’t?
Physical systems are non-reflexive — they don’t adapt their behavior based on being observed or studied. A material’s properties are fixed by physical law and don’t change because a pattern was discovered or published. Markets are reflexive: participants read the same analysis and change their behavior, which changes the system. Prediction works honestly in the first case and self-destructs in the second.
Do quantitative finance skills transfer to science?
The core skills transfer almost completely: separating signal from noise, detecting which regime a system is in, decomposing complex outcomes into driving factors, and — most importantly — intellectual honesty about the limits of your evidence. What changes is the substrate, not the discipline. A mind trained to not be fooled by noisy market data is well-suited to not being fooled by noisy scientific data.
What is the observer problem and why doesn’t it apply to physics?
The observer problem in markets is that participants are also observers — measuring and acting on the system changes it. In the physical sciences, the system being studied (a material, a reaction, a crystal) does not respond to being observed or to the existence of papers about it. The everyday physical world is non-reflexive, so prediction remains valid in a way it never is for reflexive human systems.
MarketCrystal provides AI-powered trend analysis for cryptocurrency markets. We don’t predict the future — we help you see the present clearly. [Learn more →]