Methodology — MOSTLY NOTHING.

Last updated: October 6, 2026

If you clicked into Methodology voluntarily, we are probably going to get along 😂.

The actual investing work lives in Playbooks and Research. I think of this page as the anti-bullshit layer: enough bookkeeping that future me cannot look back three years later and somehow remember every decision as smarter than it really was.

The basic idea is simple. I keep contributions separate from returns, leave old decisions attached to their date, and try to make it obvious where a reported fact ends and one of my assumptions begins. Current portfolio data can move without quietly rewriting what I said at the time.

Pretty boring stuff. Also probably the part I will be happiest exists if this thing is still around in 2030.

Two scoreboards

There are two numbers I care about, and combining them would make me look better than I deserve.

Total invested assets

This is the wealth-building number: long-term investments across taxable brokerage, Roth IRA, 401(k), HSA, and any other long-term investment accounts that belong in the picture.

It answers one question:

How much capital have I actually built?

That number rises because I save more money and because the investments move. Both are good. They are not the same thing.

If invested assets go from $312,000 to $332,000 and I contributed $15,000, I did not make $20,000 investing. I mean, come on.

Normal Articles can carry a dated Portfolio Tally so that wealth-building path stays visible. The first baseline was roughly $312,000 on September 25, 2026. It is supposed to look old eventually. That is kind of the point.

Active portfolio performance

This is the scoreboard that does not let contributions—or the index sleeve—do the flexing for me.

The active Mostly Nothing portfolio is the individual-stock sleeve across the taxable brokerage + Roth IRA. FXAIX, FSKAX, FSMAX, FTIHX and money-market cash are excluded from the performance record.

The official active record starts Dec. 12, 2025, when I first bought individual stocks.

Now the question is clean:

Did the stocks I actively chose beat the boring alternative?

The sleeve has its own cash-flow boundary. Capital moved from an index fund into an individual stock counts as money entering the active sleeve even though it never left Fidelity. A stock-to-stock rotation stays internal. Dividends and reinvestments stay internal. If active capital moves back to the index sleeve, that is a withdrawal from the active sleeve.

TWR grades the stock-picking result without letting flow timing take credit. VOO is the primary benchmark over the same dates. The matched-flow comparison asks where the exact same active-sleeve dollars would have ended up in VOO.

The current numbers, formulas, dates and audit trail live on the Performance page. I keep them there on purpose. Two performance pages would eventually become two different versions of the truth.

Old me and current me are different records

A lot of Mostly Nothing makes more sense once this distinction clicks.

An Article or Research Note records what I knew, assumed and decided then.

The living pages answer what is true now:

Portfolio owns the current allocation.
Holdings owns current and previous real positions.
Research owns the current state of the broader company universe.
Performance owns the latest verified return evidence.

So an old Article can permanently say “~$312K invested on September 25, 2026” while the current portfolio shows something completely different.

Same thing with a stock. If I reject it at $100 and love it at $60 six months later, the $100 rejection stays. Maybe I was wrong. Maybe the evidence changed. Maybe the price finally got interesting. I want to be able to see the actual path instead of pretending the final view was always obvious.

The new view gets a new date.

What I change, and what I leave alone

I want the archive to preserve bad judgment. Bad arithmetic gets no such protection.

What changed?	What I do
A factual claim was wrong	Correct it
The math was wrong	Correct it
A source was stale or incomplete	Replace it with better evidence and note the correction when material
My valuation changed	Preserve the old view; publish/update the new dated view
The thesis changed	Preserve the old thesis; record what changed
The stock price moved	Update current surfaces where appropriate; do not rewrite old publications

If every old call somehow becomes obvious in hindsight, the archive is failing.

Contributions are not returns

This gets its own section because it is ridiculously easy to fool yourself with a rising account balance.

For Portfolio Tallies:

investment / market movement = change in total invested assets − net external contributions

That is a useful bridge between two wealth snapshots. It tells me how much of the change came from new money versus everything else.

It is not an official return calculation.

Official active-portfolio performance uses the methodology on the Performance page. The important boundary is the active sleeve, not the brokerage account. A transfer from FXAIX into MSFT is not a new household contribution, but it is new capital entering the stock-picking sleeve, so it counts as a performance flow there.

TWR grades the investment decisions. Money-weighted measures grade what the capital actually experienced given when it entered the sleeve.

Same dollars. Different question.

Investment view ≠ my action

This is probably the distinction I care about most when I write about an individual company.

A stock can look attractive and still get exactly $0 from me.

Maybe I already own enough. Maybe another holding has a better expected return. Maybe the correlated exposure is getting stupid. Maybe taxes change the math. Maybe keeping cash available is simply the better next-dollar decision.

When those answers diverge, I try to keep four things visible: the standalone investment view, the valuation/buy framework, a sensible sizing range, and what I am personally doing with my own portfolio.

“I am not buying it” is a portfolio decision.

“I do not think it is attractive here” is an investment conclusion.

Those are different sentences for a reason.

A zero-dollar action from me can coexist with a positive standalone investment view. Netflix and Broadcom are good current examples: the research can pass while my portfolio still says no position because the next-dollar comparison, overlap or concentration does not. Another portfolio with different holdings, taxes, risk tolerance and alternatives can rationally reach a different action from the same underlying work.

The actual investing work lives elsewhere

Methodology should not slowly become a third Playbook page.

The analytical machinery lives in:

How I Build a Bear / Base / Bull Model — how I turn business assumptions into scenario values and expected returns.
How I Score Capital Allocation — how I judge whether management is helping or leaking the value I am underwriting.

Company-specific Research applies those frameworks.

That separation matters because I do not want to invent a completely different process every time I fall in love with a stock.

Facts are not assumptions

For material company claims I usually start with SEC filings and earnings releases, then investor presentations and calls/company announcements. Reliable financial reporting fills in gaps when the primary sources do not answer the question cleanly.

When practical, the source sits next to the claim it supports.

Then there is the stuff I choose: growth rates, margins, terminal multiples, scenario probabilities, sizing, what counts as enough upside. Those are underwriting assumptions.

A company can report $500 million of revenue.

The company did not report that I should pay 15x earnings for it.

That line needs to stay painfully obvious. Otherwise a bunch of my assumptions start wearing a little fake moustache and pretending to be facts.

What stays private

The public record can include portfolio values and weights, holdings, transactions, cost basis, contributions, valuation assumptions, scenario ranges, rejected ideas, performance, and benchmark comparisons.

Account numbers, identifying information and unrelated personal-finance details do not make any of that analysis better, so they stay out of it.

I want enough transparency that somebody can audit the decisions. Doxxing myself so the methodology feels more serious would be a genuinely terrible trade 😂.

The future-me test is pretty simple

Years from now, I want to be able to open an old decision and answer three questions:

What did I know? What did I assume? What did I do?

If I can answer those three questions later without hindsight quietly rewriting the story, the methodology is doing its job.

For ownership conflicts, model limitations and the obvious “please do not outsource your financial decisions to a guy on the internet” part, see Disclosures.