Study ›
[Course] [Build an EA 7] Demo → small live → live verification
If the backtest looks good, it's time to check it in the real market. Not skipping steps is what matters most.
Step 1: 2–4 weeks on demo
- Attach it to a demo chart with the same settings as the backtest.
- Check the Toolbox → Experts tab and History every day.
- What to check
- Does it trade about as often as the backtest suggested? (Far more or less means a code problem)
- No order errors? (messages like `invalid stops`, `not enough money`)
- Are stops and targets placed where intended?
Step 2: keeping the computer on
An EA runs only while MT5 is open. Turn the computer off and it stops. Two options:
- VPS (virtual server): rent a remote computer that's on 24/7 and run MT5 there. A few dollars to a few tens of dollars a month; some brokers offer one free.
- MT5's built-in virtual hosting: pay inside MT5 to move the EA to MetaQuotes servers.
Step 3: a small live account
- If demo went well, start live with money you can afford to lose.
- Demo and live fill differently. Slippage (worse fills than requested) and spread widening (news times) are bigger live.
- Compare about a month of results with the backtest and demo. If they differ a lot, find out why before adding money.
Step 4: verify it publicly on EADesk
Anyone can make a "backtest screenshot," so it's hard to trust. With EADesk Verified → My strategy (live link), your MT5 account performance is published every minute, as is.
- On Verified strategies: Post a verification → My strategy (live link)
- Paste the reporter code at the top of your EA, or attach the separate reporter EA, as guided
- Once connected, the card shows total return, max drawdown and open positions live
Other members can follow your strategy with Paper copy, which is a great way to get your skill judged objectively.
When to stop
Decide in advance. E.g. stop if max drawdown exceeds 1.5× the backtest, then analyze why. Many accounts end with "just a bit more" on gut feeling.
Live results being worse than the backtest is normal. Plan on roughly 20–50% less.
