0
18
<h2 style="margin:26px 0 12px 0; line-height:1.3;">The chart price is not the trading result</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">A strategy can identify a valid pattern and still fail economically because the expected move is too small relative to execution friction. Spread, commission, slippage, latency and rejected orders belong inside the strategy model.</p><h2 style="margin:26px 0 12px 0; line-height:1.3;">Measure cost as a share of the edge</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">A two-point cost has a different meaning for a target of ten points than for a target of two hundred. Compare average expected trade value with realistic total cost. Small, frequent trades are especially vulnerable.</p><h2 style="margin:26px 0 12px 0; line-height:1.3;">Test variable conditions</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">Spread changes through the day and can expand around news or illiquid periods. Backtests should include conservative assumptions or segmented analysis. A strategy that survives only the minimum historical spread has little margin for error.</p><h2 style="margin:26px 0 12px 0; line-height:1.3;">Log execution details</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">For live or demo monitoring, record requested price, filled price, delay, spread, rejection reason and symbol specifications. Compare observed friction with the assumptions used during validation.</p><h2 style="margin:26px 0 12px 0; line-height:1.3;">Handle failure explicitly</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">An EA should define what happens after a rejected order, requote, missing price, invalid volume or unavailable market. Repeating an order blindly can create risk that was absent from the strategy design.</p><h2 style="margin:26px 0 12px 0; line-height:1.3;">Reject edges that are too small</h2><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;">Sometimes the correct engineering decision is not to optimize execution further. It is to reject a strategy whose advantage disappears under ordinary market friction. Robustness needs economic margin, not only statistical significance.</p><p style="text-align:justify; margin:0 0 16px 0; line-height:1.65;"><em>Historical simulations cannot fully reproduce future execution conditions.</em></p>
