Paper testing is for finding system behaviour

Paper mode is sometimes treated as a rehearsal for profit. A more useful purpose is to observe system behaviour. It can show whether a signal arrives, whether a rule blocks an action, how quantities are rounded, which fields are logged and how a recovery path behaves when data is late or unavailable.

That distinction matters because simulated fills are not live fills. Paper mode cannot reproduce every spread, queue position, gap, partial fill or broker rejection. It is still valuable for software and process questions, as long as the limits of the simulation are recorded rather than forgotten.

Define the test before starting

Write down the change you are testing, the observation window, the fields you will review and the condition for keeping or discarding the change. Keep the rest of the workflow stable where possible. If several rules change at once, a later result cannot tell you which change caused the behaviour.

Useful review fields include signal timestamp, data source and observation time, intended quantity, broker constraints, configured limits, simulated fill, cancellation reason and any manual intervention. An empty field is a finding; do not replace it with a plausible value.

Treat the handoff as a separate test

Moving from paper to live execution is not a single switch. Recheck account identity, permissions, symbol mapping, currency, minimum size and risk limits. Use credentials without withdrawal permissions where the broker supports that choice. Confirm how the system pauses if a connection or required quote is unavailable.

Paper testing cannot establish that a strategy will make money. It can make the operational path more legible and expose assumptions while the cost of being wrong is lower. The final decision to use capital remains yours.