Backtesting Guide

This page explains how to test Orion Gold Scalper yourself in MetaTrader. It does not publish fixed win rates, profit factors, or guaranteed historical returns — those depend on your data, broker model, spreads, and settings.

Use the performance evaluation page for what to judge on demo, and read the risk disclosure before going live.

What a useful backtest needs

Your broker’s XAU/USD symbol (including any suffix)
Realistic spreads / commissions for the account type you will use
Enough history to cover quiet and volatile gold periods
Settings you would actually run — see the settings guide
Demo confirmation after the tester — fills and news spikes differ live

How to run a backtest (MT4)

Step 1: Prepare historical data

Use quality XAU/USD history from your broker or MetaTrader’s data tools
Check for large gaps before trusting results
Prefer longer samples over a single “perfect” month

Step 2: Open Strategy Tester

Press Ctrl+R or go to View → Strategy Tester
Select Orion Gold Scalper from the Expert Advisor list
Choose your gold symbol
Pick a scalping timeframe you plan to use (often M1 / M5 / M15)

Step 3: Configure the test

Period: enough history to include different regimes
Model: Every tick when you need higher fidelity (slower)
Deposit / leverage: match how you intend to trade
Inputs: start from documented defaults, then change one variable at a time

Step 4: Run and review

Click Start, then review the report and equity curve. Treat the numbers as a relative check of settings and data quality — not a promise of live results.

Metrics worth reading (without targets)

Profit factor — gross profit ÷ gross loss; sensitive to a few large trades
Maximum drawdown — peak-to-trough stress; decide what you can tolerate
Win rate — alone is meaningless without average win vs loss
Trade count — too few trades makes any metric noisy
Equity shape — smooth vs jagged tells you more than a single percentage

There is no universal “good” win rate for gold scalping. Compare runs under your own constraints.

Backtesting limitations

Past data does not guarantee future performance
Fixed-spread models understate news and session spikes on gold
Slippage, requotes, and VPS latency are often missing from tester results
Over-fitting settings to one period usually fails out of sample

Always confirm on a demo account with your broker before risking capital. See strategies and installation.