The Eight
Eight principles, in three groups, each with a test that says plainly when it's been broken.
The structure borrows from the Noble Eightfold Path: three groups covering how you see, how you act, and how you hold. The shape is borrowed, and the words are my own, written for investing.
The thesis
Markets manufacture craving and aversion. The record shows how I handle mine.
Buddhist psychology is, functionally, a technology for managing craving and aversion, and markets manufacture both on a daily schedule. Greed and fear are states that arise in you, and there's a long body of practice for noticing them without obeying them.
The portfolio is the Core: US-listed stocks tilted toward cheap, profitable, conservatively run companies, the kind other people often find unattractive or frightening. Over long periods, holding companies like these has tended to pay. Economists call that extra return the value, profitability, and investment premia. Whether I still hold them when it gets ugly is what the record's for.
The catch is that those premia exist because owning these companies sometimes hurts. Cheap, unloved companies can fall hardest in bad markets, and the strategy can trail the market for years.
Why every principle has a failure test
Every principle below ends with broken when: a specific, checkable condition.
Without those, principles are meaningless. With them, the framework can be audited: every broken principle is published in that month's entry, and every breach of the written rules goes in the Breach Count, which is never reset. A principle that can't be broken can't be followed either.
Wisdom: how you see
Before any money moves.
1. Clear View
Know what you actually own.
The Core. Know every holding's weight, sector, and why the rules bought it. Know the factor loadings the record actually shows, which can drift from the ones the screen aims for.
Broken when: at a monthly date, I can't say what I hold, my largest sector exposure, or how the latest loadings compare with the targets.
2. Honest Intention
Name the real reason before you change anything.
The Core. It's for money with a horizon of decades. Nothing else belongs here. Every rule change names its purpose before it's made, and the purpose can't be a recent result or a feeling about particular companies.
Broken when: I proposed a rule change and the honest reason was a recent result, or how specific holdings felt.
Conduct: how you act
The moves themselves.
3. Plain Speech
If you can't explain it in thirty seconds, leave it out.
The Core. The strategy fits in thirty seconds: buy cheap, profitable, conservatively run companies every month, and hold them when others won't. A proposed rule that needs a diagram to explain stays out. I write down every definition and stick to it. Redefining "value" because your version stopped working is the most common quiet cheat in this field. Any result I show in public has VT beside it, and the gap stays visible when I'm behind.
Broken when: a rule went in that I couldn't explain in thirty seconds, or a result was shown without its comparison.
I've leaned on this one plenty already. It ruled out several ideas that were defensible on their own terms but failed the thirty-second test.
4. Pre-Committed Action
Write the rule first, so the moment never decides.
The Core. I write and date every rule before the money it governs moves. That includes what gets bought, how much, when it gets sold, and when the rules themselves can change. Contributions follow a written rule: at least 30% of gross income, on the 1st of each month. Nothing gets decided because of a mood or a headline.
Broken when: I made a decision the written rules didn't call for. That's a breach, and it goes in the Breach Count.
Rules that respond to conditions are fine, as long as the response is written in advance, and discretion at the moment isn't. If my rule were "I'll work out what's best each month," the Breach Count couldn't be measured, since there'd be nothing to deviate from.
5. Sound Livelihood
Your work funds the portfolio. The portfolio doesn't fund your life.
The Core. In the first decade, the contribution rate matters more than any allocation decision, because the portfolio's appreciation is small next to what goes in each month from income. Savings are also the one input that keeps working in the months where nothing else does. I take money out for just two things: once toward a home, and for a medical emergency.
Broken when: a contribution was cut because of spending, or money came out of the Core for any other reason.
Discipline: how you hold
The part most people fail.
6. Steady Effort
Follow the same rules in every market, especially the bad ones.
The Core. Every month I put in 30% or more of my gross income, on the same day. The months that test this are the ones where the Core has fallen and every headline says to wait. Holding back then is market timing, even when it feels like caution.
Broken when: I cut, delayed, or held back a contribution because of how the market felt.
7. Mindful Attention
Watch yourself, not the market.
The Core. It isn't checked between monthly dates, because nothing I'd see would change what I do anyway. What I should be watching is the pull I feel to check, sell, or buy more. If a temptation gets as far as a concrete step, like opening my brokerage account or looking up a price, I write it in the temptation log and don't act on it.
Broken when: a temptation got as far as a concrete step and I didn't log it.
Checking counts as a temptation, not a breach. The temptation log is published every month, with whether each temptation became a breach. A temptation logged and resisted is the clearest evidence that the rules are being held rather than simply never tested.
8. Unbroken Horizon
Judge the strategy in decades, not months.
The Core. Give it decades before you judge it. Factor premia are paid on average and can go missing for ten years at a stretch, so a good year proves nothing and neither does a bad one. Each monthly entry reports what happened without treating it as a verdict.
Broken when: I judged the strategy on a month or a year of results, in either direction, in private or on camera.
The Breach Count
Three records are kept, separately, and all are published every month:
- A breach is a chosen action, or deliberate inaction, that the written rules didn't call for. Each one adds to the Breach Count, and nothing ever takes it back down.
- An Error is an unintended deviation: the wrong lot sold, a missed transfer. Counted separately, so the Breach Count measures behavior rather than competence.
- A temptation is an impulse to act outside the rules that reached a concrete step, whether or not it went further. Checking the portfolio between monthly dates is a temptation.
The Breach Count is the headline figure because it's the only meaningful number from month one, it can't be gamed in a flattering direction, and it measures what this project is actually about. A low Breach Count alongside poor returns is an honest and interesting result.
Each breach is recorded with the date, the principle, what was done, and what the rule said. No justification is recorded. The reasoning always felt good at the time, and preserving it would turn the log into a defense rather than a record. None is ever removed, including deviations that turned out well.
Not breaches: a rule turning out to be wrong (that's a revision, made in January), a month with no activity, a loss, a sale deferred by the one-year hold, a purchase rounded to a whole share, a permitted withdrawal, and checking. An honest mistake counts as an Error and stays out of the Breach Count. And a breach is never undone by another deviation. The portfolio returns to the rules at the next monthly date.
"Couldn't a bot run this?"
The rules do pick the stocks: anyone running them on the same data would buy the same companies. But a bot has no temptation to break its rules, so following them proves nothing: its Breach Count is zero by default. My Breach Count is a record of a person choosing, month after month, not to override.
A meditation instruction is a rule too (sit, notice the breath, return when the mind wanders), and nobody calls a meditator a bot for following it. Picking rules is easy; sticking with them is the hard part.