Kedge Gold Scalper
- Experts
- Version: 1.0
- Activations: 5
Kedge Gold Scalper is an M1 scalping Expert Advisor for XAUUSD, with defaults built for a USD 500 account trading 0.01 lots. Every order is sent with a stop loss and a take profit. It holds one position at a time. It uses no grid, no martingale and no averaging down. It makes no DLL calls or web requests and does not use AI.
How it trades
- Trend filter (H4): it buys only when price is above a rising H4 50 EMA, and sells only when price is below a falling one.
- VWAP band (M1): the EA builds a VWAP from the start of the server day (tick-volume weighted, from completed M1 bars) with a standard-deviation band around it.
- Signal: the previous M1 bar closed outside the band (VWAP minus 1.5 standard deviations for a buy, plus 1.5 for a sell), and the last completed bar closes back inside it with a candle in the trend direction. The band must be at least 1 M1 ATR wide.
- Exits: the stop loss is 3 × M1 ATR and the take profit is 3 × M1 ATR. The stop is never closer than 8 spreads. A time stop closes any trade after 120 minutes.
- Timing: new entries are allowed from 09:00 to 21:00 server time, with at most 8 per day. Open trades are managed around the clock.
Drawdown protection
- Hard equity guard (default 50%): if equity falls 50% below the chosen reference, the EA closes all of its positions and stops trading. It writes the reason to the Experts log and shows it on the chart panel. The reference can be peak equity (the default), peak balance or the balance when the EA was first attached. The stop survives restarts. To trade again after you have reviewed the account, set ResetGuardOnStart=true once.
- The guard is a last-resort backstop, not the plan. In the gold backtest below, the maximum equity drawdown was 31.3% and the guard never triggered.
- Daily loss limit (default 4%): a 4% equity loss from the start of the server day closes the EA's position and blocks new entries until the next day.
Position size and filters
- Each trade risks 1% of equity at its stop distance. On a USD 500 account the 0.01 minimum lot usually risks more than that. The EA then trades 0.01 lots only when that stop loss is at most 3% of equity. Otherwise it skips the signal and logs the reason.
- A spread-spike filter skips entries when the spread is more than 2 × its recent average. An absolute spread cap is optional.
- A news filter blocks entries for 15 minutes before and after high-impact USD events in the MT5 Economic Calendar. It works only in live trading, because the Strategy Tester has no calendar data. The backtests below therefore include news periods.
- Before every order it checks money, free margin, margin level, volume limits and volume step. It checks stops and freeze levels before every order and close. It runs on hedging and netting accounts, and on netting accounts it doesn't trade if another position or order exists on the symbol. It always closes whole positions and never uses partial closes.
Backtest results (honest, one year only)
All tests ran in the MT5 Strategy Tester on XAUUSD M1 with real ticks, from 2025-11-01 to 2026-11-01. They started with USD 500 at 1:100 leverage on JustMarkets demo data, using the default inputs.
| Window | Net profit (USD) | Profit factor | Trades |
|---|---|---|---|
| Full year | +50.01 | 1.03 | 436 |
| Tuning window 2025-11 to 2026-06 | +169.00 | 1.16 | 258 |
| Held-out window 2026-07 to 2026-11 (not used for any choice) | −118.99 | 0.81 | 178 |
The maximum equity drawdown over the full year was 228.12 (31.33%). Equity peaked at 729.55 and the year ended at 550.01. The daily loss limit triggered on 23 days. 320 of the 436 trades used the 0.01 minimum lot, and 50 signals were skipped because 0.01 lots would have risked more than 3%.
Please note: the strategy was picked after a small prototype screen of several ideas on the tuning window. The defaults were then chosen in three tuning passes on that window only. The held-out months, which were never used for any decision, lost money. So the tuning-window profit is optimistic, and the full-year result is close to break-even.
Small-deposit checks on netting accounts (USD 500, 1-minute OHLC, 2025-11-01 to 2026-11-01, default inputs):
- XAUUSD with a D1 chart (signals still use M1): +45.05, PF 1.03, 437 trades, equity drawdown 31.57%.
- EURUSD with an H1 chart (not the target market): −243.04, PF 0.65, 545 trades. Equity reached the 50% guard on 2026-10-29. The EA closed its position and stopped trading, as designed.
- Neither run had any trade errors or a stop-out.
Risk warning
Scalping gold on M1 is very sensitive to spread, slippage and execution speed. Live costs, especially around news, may be worse than in these tests. One year of simulated history does not promise future results, and the held-out period was a loss. A 50% drawdown limit means you can lose up to half of the account, and gaps or slippage can push the loss beyond it. Only trade money you can afford to lose. Run it on a demo account first, and use it only on XAUUSD.
