1. Define the observation rule before seeing outcomes
Decide what counts as a signal, which timestamp/entry is used, how long a pending order remains valid and whether a signal can be excluded. Rules created after a loss introduce selection bias.
2. Record the full trade lifecycle
For every signal record source timestamp, instrument, direction, entry/zone, stop, targets, later edits, activation, partial closes, cancellation and final close. Preserve losing and break-even outcomes as carefully as winners.
3. Choose a consistent multi-target method
Options include equal position slices, a fixed fraction at TP1 with the rest to TP2/TP3, or one standardized target for analysis. State the method. Counting TP1, TP2 and TP3 as three full-size wins from one trade inflates results.
4. Separate pips/points from money return
Pips do not account for stop size, position size or instrument value. XAUUSD and indices make cross-provider pip comparisons especially weak. R-multiple—profit/loss divided by initial risk—can be more comparable if the stop is known.
5. Report more than win rate
Useful statistics include number of trades, wins/losses/breakeven, average win, average loss, expectancy, maximum losing streak, drawdown, profit factor or total R, and the percentage of signals that were actually actionable at the received price.
6. Add execution assumptions
Record spread, commission if relevant, slippage and whether the assumed entry was realistically available after the signal timestamp. A provider result based on a perfect source entry may not reproduce for a delayed follower.
7. Keep open trades separate
Do not force unresolved positions into wins or losses at an arbitrary date. Report them as open and publish the measurement period. If a provider changes historical posts, preserve the original observation trail where legally permitted.
8. Publish limitations next to the result
State whether the record is provider-published, independently observed, demo/live, broker-specific or reconstructed. Transparent limitations increase usefulness; hiding them makes a precise-looking number less trustworthy.