Real money, real numbers, and far too little time elapsed.
Real accounts, real numbers, and a track record far too short to mean anything yet. Published anyway, because a number you only publish once it flatters you is not a record.
As of 21 August 2026, the Roth IRA balance has gone from $7,500 to roughly $9,000, and the taxable brokerage account sits around $12,000 — a few hundred above where it started.
Those are account balances, not returns. Balances move because of contributions as well as performance, and I am contributing, so do not read the first figure as a monthly return — it is not one, and anybody quoting it as one would be quoting it wrong.
The brokerage account was about $2,500 up at its peak. It is not there now. That gap is the most useful number on this page and it is the one a screenshot posted at the top would have hidden.
Why none of this is evidence
The accounts are about a month old. A month is not a track record, it is a data point, and a good month says almost nothing about whether the decisions behind it were sound.
Any return over a stretch this short is dominated by what the whole market happened to do. Separating skill from that needs years, not weeks, which is exactly why the backtesting framework re-runs everything on data the settings never saw.
If this had been down 20% instead, the correct conclusion would have been equally weak — and it would still be here.
What I am actually tracking
Return against the S&P 500 rather than against zero, because beating nothing is not an achievement.
The worst drawdown, since that is the number that decides whether a strategy is one you can actually stay in.
Every decision and the reason for it at the time, so that later I can tell the difference between being right and having been lucky.