Every strategy has to beat random buying. Most don't.
A strategy that made money in the past isn't proof of anything. Most stocks went up in 2025, so almost any rule looked smart. The real question is whether the rule did better than buying the same stocks on random days.
The rule
Written down before any test. When to buy, where the stop is, when to sell. No changing it after seeing the result.
The backtest
Thousands of past trades in a rising market (2025-26) and a falling one (2022). Real broker fees, and only prices known on the day.
Against random
Every trade is compared with buying the same stock on random days. It must beat that by at least 0.5 points a trade, over 50+ trades.
20 trades, live
Rules that pass trade live with paper money. Each trade is posted when it opens, before anyone knows how it ends.
Why random buying is the bar
Greed says a strategy works because it made money. Fear says nothing works. Both skip the real test. If buying at random on the same stocks did just as well, the rule added nothing, and you paid fees for a feeling.
What happens without a backtest
You find out with your own money. One losing trade feels like bad luck, so you try again. By the time the pattern is clear, the losses are real. A backtest shows those losses first, on paper.
What a backtest can't do
- It can't promise the future. Markets change, which is why every strategy that passes still has to prove itself live.
- It can be fooled by luck. Test enough rules and one will look great by chance, so the gate is strict and we count every try.
- It can hide costs. Ours include real broker fees on every trade, and we check that a result survives double fees.
Strategy Graveyard
Strategies that sounded smart and lost to random buying get their own episode instead of being quietly dropped. The failures are what make the passes believable.
Fees
About 1.6% a round trip on a small account, based on a real broker's US-stock fees and currency exchange. Bigger accounts pay less, so results here are on the strict side.