Weekly audit of partial closes in an EA

[FXS London Scalper]

My EA uses partial closes: half at the first target, the rest with a trailing stop. In backtests it looked good. In live trading, the weekly numbers didn't match, so I started a short audit every Friday.

What I compare:

• Number of partial closes expected vs executed. If the first target is touched, was the partial actually done?
• Volume closed. Half of 0.03 is 0.015, which is not a valid volume. Did the EA close 0.01, 0.02, or fail?
• Price of partial vs target. Slippage on partials is small but adds up.
• Remaining position's stop. Was it moved to break-even as the logic says?
• Deal tickets. On MT5, partial closes create separate deals. Make sure your reporting script picks them all up.

What I found in the first month: about 8% of partial closes failed silently because of volume rounding on small positions. The EA logged a warning nobody read, and the full position then ran to the stop or the second target.

Fix: minimum position size so that halves are always valid, plus an alert when a partial fails.

If you run partial-close logic, do you audit it regularly? Anyone found other failure modes?