Paper trading & backtesting: keep the failures

Separate practice from evidence

Paper trading practises decisions and the order interface with simulated money. A backtest applies rules to historical data. Both can be useful, but neither reproduces every condition of live execution.

Use a written plan and record every matching case, including trades you would rather forget. Keep hypothetical fills, actual quoted prices and later interpretations separate.

Remove the view from the future

Decide when information was available. A rule cannot use the final high, low or closing indicator value before that bar has finished. Economic releases can be delayed and later revised.

Freeze the rule before evaluating an untouched period. Choosing the parameters, assets or dates after seeing the results can produce a flattering history. Missing delisted companies or inconvenient source gaps can distort it further.

Charge the test for trading

Include plausible commission, spreads, slippage and any applicable financing or borrowing costs. A simulator’s fill at a price does not establish that a real order of your size would have filled.

• Report losing runs, drawdowns and the number of observations alongside returns.
• Keep the original rule and the reason for every subsequent change.
• Judge process errors separately from outcomes; one profitable result proves little.