Every trader, beginner or experienced, should run an EA on demo before committing real capital. Even when I test our own Golden Viper EA on a new broker, I still start with demo. There are just too many variables between your broker's execution, your specific setup, and real market conditions to assume everything will work perfectly from day one, and demo testing removes that uncertainty at zero cost. In this beginner's guide, I'll walk through what to test during demo, how long to wait, and when it's safe to switch to live trading.

Why You Should Run Your EA on Demo First

Skipping demo testing is one of the costliest mistakes in EA trading. Here is what demo testing validates:

• Installation verification: Confirms the EA is properly installed, shows a smiley face, and has no errors in the journal
• Broker compatibility: Tests that your broker's symbol naming, spreads, and execution work with the EA
• Settings validation: Ensures all parameters are configured correctly for your account size and risk tolerance
• Behavior understanding: Shows you how the EA trades, including frequency, holding time, and drawdown patterns, so you know what's normal
• Confidence building: Builds trust in the EA before real money is at stake, preventing panic during normal drawdowns

Think of demo testing as a dress rehearsal: same market data, same charts, same execution, just with virtual money standing in for real. There's zero downside to it and plenty of upside.

Why This Matters Even More for a Gold (XAUUSD) EA

Gold does not behave like a typical forex major. It reacts sharply to US CPI and NFP releases, Federal Reserve rate decisions, and safe-haven flows during geopolitical stress, and its intraday range can widen dramatically compared to a pair like EUR/USD. The World Gold Council and CME Group both track how gold volatility clusters around macro data releases and risk-off events, and financial news outlets like Kitco cover these moves in real time. A demo run that only covers a quiet two-week stretch will not show you how an EA handles gold gapping 400+ pips around an FOMC statement or spiking on a central-bank-buying headline. That's one more reason to prioritize the trade-count target over the calendar-day minimum: your demo window should span at least one higher-volatility stretch, not just calm, easy-to-trade conditions.

How to Set Up a Realistic Demo Test

A demo test only tells you something useful if it actually mirrors the account you plan to trade live. A demo account that differs from your live setup in broker, server, leverage, or hosting environment can give you a false sense of confidence, or a false red flag. Before you start the clock on your demo period, get these details matched up:

• Same broker and account server: Open your demo account with the exact broker, account type, and server you intend to use live. Symbol suffixes, spread models, and execution behavior can vary meaningfully between brokers, even ones that look similar on the surface. Our guide to choosing a broker for a gold EA covers what to look for.
• Same VPS, not your home PC: If you plan to run the EA on a VPS once live (recommended for uninterrupted 24/5 operation), demo test on that same VPS. Testing on your home computer and then switching to a VPS for live trading introduces a variable you never validated. See our recommended VPS specifications if you have not set one up yet.
• Realistic starting balance and leverage: Set your demo balance and leverage to match what you plan to fund live, not the platform's default (often $10,000 or $100,000). Leverage in particular changes margin usage and how lot-size settings behave, so testing at the wrong leverage can hide a problem or manufacture one that will not exist live.
• Follow the same installation steps you'll use live: Walk through platform setup exactly as documented, including template and chart timeframe, so you are testing the EA's actual behavior rather than a misconfiguration. Our MT4/MT5 setup guide and XAUUSD chart setup guide walk through this step by step.

What to Test During Your Demo Period

Technical Checklist

• EA shows smiley face on chart (not frown or X)
• No error messages in the Experts or Journal tabs
• EA opens trades automatically when conditions are met
• Stop loss and take profit levels are set correctly on each trade
• Magic number is assigned properly (check in trade details)
• Auto-trading button stays green after platform restarts

Performance Checklist

• Trade frequency matches documentation (e.g., 3-5 trades per week)
• Win rate is within expected range
• Average win size vs average loss size looks reasonable
• Drawdowns do not exceed stated maximum
• EA handles weekend gaps and market opening correctly

Run through both checklists at least twice during your demo period, not just once at the start. Some issues (a missed error message, a stop loss that occasionally lands a few pips off, a magic number conflict with another script) only become visible after the EA has placed a reasonable number of trades. Checking once on day one and then not looking again until you decide whether to go live defeats much of the purpose of testing.

How Long to Run EA on Demo First

The key is not just time, but number of trades observed. An EA that trades once per week needs more calendar time to accumulate meaningful data than one that trades daily. We recommend seeing at least 20 trades before making any live decision.

Week-by-Week Demo Testing Timeline

If you are not sure what to actually check each week, use this as a rough schedule. It is a guideline, not a strict rulebook, since trade frequency and market conditions will shift the pace.

Common Mistakes That Ruin a Demo Test

A demo test can technically "run" for weeks and still teach you nothing useful if it is set up or interpreted incorrectly. These are the mistakes we see most often:

• Testing on a different broker than you'll trade live: Results from a different broker's spread and execution model do not transfer. Match the broker before you start.
• Ignoring symbol suffix mismatches: Many brokers append suffixes like .m, .pro, or .raw to XAUUSD. Getting this wrong during demo can make an EA appear broken when it is only misconfigured.
• Manually intervening in trades: Closing trades early, adjusting stop losses by hand, or turning the EA off during a drawdown "to see what happens" defeats the purpose of the test. Let it run unmodified for the full period, exactly as it would live.
• Judging performance after 3-5 trades: A handful of trades tells you almost nothing statistically. A short losing or winning streak early on is normal variance, not a verdict.
• Ignoring journal error messages: A recurring error in the Experts or Journal tab (invalid stops, not enough money, trade context busy) will usually resurface live and can cause missed or malformed trades. See our guide to common MT4 errors if you spot one you don't recognize.
• Not testing the "no trades happening" scenario: If the EA goes quiet for longer than expected, don't assume it's fine, verify it against our EA not trading troubleshooting guide.
• Skipping demo entirely because a vendor says it's unnecessary: Be wary of any seller who discourages demo testing or pressures you to fund a live account immediately. Both the CFTC and the FTC list urgency and pressure to skip verification steps as common warning signs of trading-related scams. A legitimate seller has no reason to discourage you from testing first.

What to Watch for During Demo

Green Flags (Good Signs)

• Consistent trade frequency matching documentation
• Win rate within expected range (do not expect 100%)
• Smooth equity curve with manageable drawdowns
• Clean journal log with no recurring errors
• EA recovers from drawdowns within expected timeframes

Red Flags (Investigate Further)

• No trades opening for extended periods (check symbol configuration)
• Frequent error messages in the journal
• Drawdowns significantly exceeding stated parameters
• EA opening trades at strange sizes (lot size misconfiguration)
• Platform freezing or crashing (hardware or VPS issue)

None of these red flags automatically mean the EA is broken. Most of the time they trace back to a setup issue, wrong lot sizing, a missing symbol suffix, an unreliable VPS, rather than the EA itself. Treat a red flag as a signal to investigate and fix, not an automatic reason to abandon the test. If the underlying cause turns out to be your broker's execution quality rather than the EA, that is exactly what demo testing is supposed to surface before you commit real capital.

Slippage, Spreads, and Execution Quality: What Demo Can't Fully Show You

Demo accounts are a good proxy for an EA's logic, but they are an imperfect proxy for execution. Slippage, the difference between the price you expect and the price you actually get filled at, tends to be minimal on demo and more variable live, especially during fast markets. MetaTrader's own strategy tester documentation notes that even historical backtests only approximate real execution conditions, and a live demo account, while closer to reality than a backtest, still is not identical to a funded live account at the same broker.

The broker's execution model has a lot to do with how big that gap ends up being:

This is one of the main reasons we recommend demo testing on the same broker, and ideally the same account type, you intend to fund live. If you want to dig deeper into this specific topic, see our guides on what causes slippage in a gold EA and how to reduce it, minimizing slippage for automated gold trading, and ECN vs STP broker models.

When to Go Live: The Transition Checklist

Switch from demo to live when you can check every item on this list. This is essentially a forward-testing exercise: you are confirming the EA behaves consistently on your specific setup before committing capital, which is the same principle covered in more depth in our forward-testing plan for a gold EA before live deployment.

1. EA has run on demo for your minimum period without unresolved issues
2. You have observed enough trades to evaluate performance (20+ minimum)
3. Results are within expected parameters for win rate, frequency, and drawdown
4. No recurring errors or technical issues
5. You understand the EA's trading style and what constitutes normal behavior
6. You have set proper lot sizes for your live account balance
7. You are emotionally prepared for drawdowns (they will happen)

If you are still uneasy about committing to a live account after a strong demo, you can bridge the gap further with a very small live deposit rather than jumping straight to your full intended balance. Our guide on testing a trading robot without a large deposit covers this approach in detail.

Demo vs Live: Key Differences to Expect

A demo account, by definition, simulates trading using real-time market data but without placing real orders in the market. That simulation is genuinely useful, but it is still a simulation. Demo trading is close to live but not identical. Here are the differences you should expect:

• Execution speed: Demo fills are often faster than live. Live orders may experience slight delays.
• Spreads: Demo spreads can be tighter or more stable than live, especially during news events.
• Slippage: Demo has minimal slippage. Live trading may fill at slightly different prices during volatile moments.
• Psychology: This is the biggest difference. Real money creates emotional pressure that demo does not. Drawdowns that felt acceptable on demo may trigger anxiety with real capital.

Overall, expect live performance to run somewhat below demo results, since real slippage, variable spreads, and requotes are not fully captured in a simulated environment. Treat any specific demo return as an upper bound rather than a guarantee of what live trading will produce. This gap is normal and should be factored into your expectations. Understanding how XAUUSD works helps set realistic expectations, and choosing the right broker minimizes the demo-to-live gap. Our guide on backtest results vs real trading covers the same widening pattern one step earlier in the process, from backtest to demo to live, and is worth reading if you also plan to backtest before demo testing.

None of this means demo results are meaningless, quite the opposite. A demo period is what lets you separate an EA's actual logic from your broker's execution quirks. When you later look at a publicly verified account, such as a Myfxbook-tracked live results page, you are looking at outcomes from one specific broker and account setup. Demo testing on your own broker is how you find out whether your setup is likely to produce similar outcomes, rather than assuming it automatically will.

When to Restart Your Demo Clock

A demo test is not a one-time box to check. Certain changes reset the clock because they change the variables you already validated:

• You update the EA to a new version: Even minor updates can change behavior. See our guide to updating your EA version and re-run a short 3-5 day, 5+ trade demo check afterward.
• You switch brokers: Different execution and spreads mean your prior demo results no longer apply. Run a fresh 1-2 week demo period on the new broker.
• You migrate from MT4 to MT5 (or vice versa): Platform migration can affect symbol handling and order execution. Our MT4 to MT5 migration plan covers what to re-check.
• You change your VPS or hosting setup: Latency and uptime affect execution timing. Re-validate on the new hosting environment before trusting live results from it.
• You significantly change risk settings: Adjusting lot sizing mode or risk-per-trade parameters changes your exposure profile enough to warrant a short re-check before scaling back up to your target size.