Every trade on this site is committed to a public GitHub repository, and the returns are calculated by open code from that record, not typed in by me. Losses stay up. You don’t have to trust me; you can check.

Strategy

High growth, high risk, unconstrained. Holdings may include ASX and US equities, ETFs, options including short-dated contracts, and crypto.

Most of the portfolio, around 75%, sits in a core of 4–6 high-growth companies I’ve valued in depth and plan to hold for years. The rest is a satellite for higher-risk ideas, including early-stage companies, options and crypto, where I size small and expect big swings both ways. Every position, core or satellite, has a published thesis and the conditions that would prove me wrong.

The portfolio started with A$10,000 and is measured against the S&P 500 in Australian dollars.

Position sizing and review rules

Before buying, I estimate how far a position could fall if my thesis is wrong, and I size it so that loss would cost no more than 5% of the portfolio. A stock whose bear case is −50% gets up to 10%; one that could fall 80% gets about 6%.

I review every position after each earnings report against the conditions in its note, and if one is met I sell, whatever the price. A 30% fall from cost doesn’t force a sale; it forces a written re-underwrite in the journal within a week that answers one question: would I buy it today at this price? I trim when the price is more than 20% above my fair value estimate, and I sell when my updated estimate falls below the price.

Performance methodology

Starting point. The portfolio starts with A$10,000 on 25 Sep 2026. The benchmark starts with the same A$10,000 on the same day.

Value. Each business day, the portfolio is worth the market value of every holding plus cash in every currency, converted to Australian dollars.

Benchmark. The S&P 500, measured through the iShares S&P 500 ETF (ASX: IVV) in Australian dollars, with distributions reinvested. It’s a real fund any Australian investor can buy, with its fees included. It’s unhedged, so like my US holdings it moves with the Australian dollar.

Returns. Returns are time-weighted and calculated daily, so adding or withdrawing money doesn’t distort them. The dollar chart compares my portfolio with what the same deposits would be worth in IVV.

Costs and tax. Returns are after brokerage, currency conversion costs and US withholding tax on dividends. They are before Australian tax.

Weekends. No valuation on weekends. Crypto never stops, so a weekend move turns up in Monday’s number.

Missing prices. When a market is shut and a day has no price, the last one carries forward, and the holding shows the date its price is really from.

The record. Every trade, price and calculation lives in a public GitHub repository, along with the data sources. Nothing is deleted, and every correction is logged with its reason.

Trading policy

Every trade is disclosed in the journal, with the date it was executed.

I set and apply my own rules on when I trade relative to publishing a note. Those rules are not published here, and nothing in this repository enforces or checks them.

The AI + Human process

AI helps with research: summarising filings, drafting a first pass at a thesis, screening comparable companies. Every valuation input, every judgement and every trading decision is mine. Each research note sets out where AI was used and where it was not.

Both AI and human analysis can be wrong, and this arrangement does not make either less likely.

Data sources

All prices are end-of-day closes. Nothing here is intraday.

Decision log

Every methodology and site decision is recorded, with its reason and its date, in section 19 of the project brief. Rows are appended and never removed, including the decisions that turned out to be wrong.