MT5 EA Not Making Money? 7 Reasons Why (and How to Fix Each)

Your EA passed a 5-year backtest with a profit factor of 1.6. You ran it on a demo account for two weeks — looked fine. You switched to live, and three weeks later you're down 8%.

This is one of the most common experiences among algorithmic traders. The good news: most causes are diagnosable and fixable. Here are the seven most frequent culprits.

---

1. Your Broker's Spread Is Higher Than Your Backtest Assumed

The problem: MT5 Strategy Tester uses historical spread data from the server you select. If you backtested on XM MT5 server but you're trading live on a different broker — or even a different account type on the same broker — the actual spread you're paying may be 2× to 4× higher.

For XAUUSD, typical spreads range from 1.5 pips (RAW/ECN accounts) to 6+ pips (standard accounts during off-hours). A gold EA with a 20-pip average win needs to overcome the spread on every trade. At 5-pip spread, that's 25% of the profit gone before the trade closes.

The fix: Check your live trade history in MT5. Look at the difference between the open price and the first price after entry — that gap is your actual spread. Compare it to your backtest assumptions. If it's materially higher, switch account types or brokers, or increase the MaxSpread filter threshold in the EA.

---

2. Wrong Symbol Name

The problem: This sounds trivial but it causes real losses. XM Trading lists gold as GOLD, not XAUUSD. If you attach an EA coded for XAUUSD to a GOLD chart, or vice versa, the EA may misbehave — entering on the wrong instrument, calculating lot sizes incorrectly, or throwing errors.

The fix: Always verify the exact symbol name on your broker's Market Watch. For XM: use GOLD. For Exness, HFM, IC Markets: use XAUUSD. Hard-code a symbol check in your EA's OnInit() if possible.

---

3. Your EA Is Over-Optimized (Curve-Fitted)

The problem: If your backtest profit factor is above 2.5 and your parameter set was found through a broad optimization grid, there's a good chance the EA is curve-fitted to historical data. It "learned" the specific quirks of the 2019–2024 gold market rather than a genuine trading edge.

A telltale sign: the EA performs brilliantly on the optimization period but falls apart on data it hasn't seen before.

The fix: Always do out-of-sample (OOS) validation. Split your historical data — optimize on 60%, then test the same parameters on the remaining 40% without touching them. If performance collapses on the OOS period, the strategy is not robust.

At fxea365.com, every EA must pass OOS validation before publication. A profit factor of 1.4–1.6 with consistent OOS results is worth far more than a curve-fitted 3.0.

---

4. Slippage Is Eating Your Edge

The problem: In backtesting, orders fill at the exact bar open or signal price. In live trading, your order competes with thousands of others. During news events or illiquid hours, slippage of 5–20 points on XAUUSD is normal.

For a scalping EA targeting 15-pip moves, consistent 10-point slippage cuts your edge by more than half.

The fix: Set a MaxSlippage parameter in your EA (our EAs default to 20 points). Avoid trading during major economic releases (NFP, CPI, FOMC). Use a VPS located near your broker's server to minimize network latency.

---

5. You're Running It on the Wrong Timeframe or Session

The problem: Most gold EAs are tuned for specific market sessions. An H1 trend-following EA designed around the London-New York overlap (13:00–17:00 UTC) will behave very differently during the quiet Asian session where gold barely moves.

If you're running an EA during the wrong session, it may over-trade in choppy conditions or miss the moves it was designed for.

The fix: Check when your EA generates most of its historical trades using the MT5 Strategy Tester "Entries by hours" chart. Match your live trading hours to those patterns. Use a time filter if your EA doesn't already have one.

---

6. Your VPS or PC Keeps Going Offline

The problem: MT5 EAs only trade when the terminal is running and connected. If your home PC sleeps, loses internet, or restarts during an update, your EA misses trades — or worse, stays in an open position without the ability to exit.

The fix: Use a VPS (Virtual Private Server) that runs 24/7. For gold trading, you need a VPS in the same region as your broker's server — ideally within 10ms ping. Many brokers (including XM and Exness) offer free or subsidized VPS if your trading volume meets a threshold.

See our VPS setup guide for recommended providers.

---

7. The Market Regime Changed

The problem: No EA works in all market conditions forever. A trend-following EA built on 2020–2024 gold data saw one of the strongest gold bull markets in history. If the market transitions to a range-bound regime, a trend EA will churn through losses.

This isn't a bug — it's a structural limitation of any strategy. Even the best discretionary traders underperform when their edge is out of phase with the market.

The fix: Monitor your EA's drawdown versus its historical max. If it exceeds 1.5× the historical max drawdown, pause trading and investigate. Don't let a regime change turn a 10% drawdown into a 40% one.

---

Summary

Reason	Quick Check	Fix
High spread	Compare live vs backtest spread	Switch account type or broker
Wrong symbol	Check Market Watch symbol name	Match symbol exactly
Over-optimization	Run OOS validation	Use conservative parameters
Slippage	Check order fill prices	Add MaxSlippage filter
Wrong session	Check entry hours chart	Add time filter
PC goes offline	Check EA logs for gaps	Use a VPS
Market regime change	Compare current DD to historical	Pause at 1.5× max DD

---

If you're looking for a free gold EA with conservative, OOS-validated backtest results, download the GOLD EMA ATR EA — PF 1.45, DD 7.2%, 680+ trades over 5 years of real tick data.