Demo Trading and Sample Size

Making a demo real

• Set the balance to the amount you would actually deposit.
• Trade the exact rules, sizes and hours you would trade live.
• Journal every trade as if it were live.
• Treat a rule break on the demo as seriously as one with money; it is the habit you are testing.

Why small samples lie

Toss a fair coin ten times and getting seven heads is common. Take ten trades with a 50 percent strategy and seven losses is equally common, and it proves nothing about the strategy. The variance of a win rate falls with the square root of the number of trades: at ten trades the uncertainty is enormous, at a hundred it is about plus or minus 10 percentage points, at four hundred plus or minus 5. That is the arithmetic behind the advice to test for at least fifty to a hundred trades.

Streaks are normal

A trend-following strategy with a 35 percent win rate should expect a ten-loss streak in every hundred trades. A trader who did not know that abandons the strategy at loss six, one trade before it would have worked.

Deciding the sample in advance

Before the test, write down the number of trades or the period, and the results that would make you continue, adjust or stop. Deciding afterwards means deciding after seeing the results, which is how good strategies are abandoned and bad ones kept.

Key takeaways

• Run the demo with real balances, real rules and a real journal.
• Ten trades are noise; fifty to a hundred begin to mean something.
• Losing streaks of seven or more are normal at ordinary win rates.
• Fix the sample size and the decision rules before the test.