From Demo to Live: Use a Demo Account Well and Know When to Go Real
A demo account copies prices, not fear or real order fills. How to trade demo like real money, a checklist before moving, and a step-by-step plan to start small.
- A demo doesn't simulate emotions, and on some brokers it doesn't simulate slippage or the wider spread in fast markets either
- Trade demo with the same balance and lot size as the live account you plan to open, not the 100,000 the platform hands you
- Use a checklist written in advance, such as 50–100 rule-based trades, a positive average per trade after costs, and drawdown within your limit
- Go live with a cent account or 0.01 lot, compare results trade by trade, and expect live to come in lower
A demo account uses pretend money on real market prices. It is a good place to learn the platform and test rules without paying for it. But it can’t simulate everything, and the biggest thing missing is how it feels when real money disappears in front of you. This article covers how to get trustworthy information out of a demo, a checklist before moving, and how to start live in steps instead of jumping.
What a demo doesn’t simulate
1. Emotions. On demo, five losses in a row are just numbers. On a live account they are money you worked for. Most people start doing things they never did on demo: closing winners early, moving the Stop loss further away, or hesitating to take the next trade after a losing run.
2. Order fills. Some brokers fill demo orders at the price you see, every time. On a live account you may get slippage, a fill at a different price from the one you asked for. It usually happens when price moves fast or around news, and it usually goes against you. Stop orders that enter as price breaks a level are hit hardest (more on order types in Order types explained).
3. Spread at certain times. The spread is the gap between the buy and sell price. A broker’s demo and live accounts may not use the same spread, especially at the market open, around the server’s day change and during news.
Points 2 and 3 depend on the broker and account type and aren’t different everywhere. Point 1 happens to everyone.
Trade demo like it’s real
One main rule: the balance and lot size on demo must match the live account you intend to open.
Example: you plan to open a 500 USD live account and risk 1% per trade.
- Risk per trade on the live account = 500 × 1% = 5 USD
- If you use the 100,000 USD demo balance the platform set, the same 1% = 100,000 × 1% = 1,000 USD per trade
- 1,000 USD on demo means nothing emotionally, and it doesn’t train you to size a small account either, where the 0.01 minimum lot may already be too big for some Stop loss distances
Better to open the demo with 500 USD from the start and size every trade from its Stop loss distance (see Position size from your Stop loss and Risk 1% per trade).
Also:
- Write down your entry, exit, Stop loss and daily loss limit rules before you start, and don’t change them midway. If you do change them, restart the trade count
- Log every trade in a trading journal, including a column for “did this trade follow the rules?”
- Trade during the hours you will trade live, not whenever you happen to be free
A checklist before going live
Write this checklist before you start the demo, not when you are itching to open a live account. The frame below is a sensible set of criteria, not a formula that guarantees a result.
| Item | Example criterion |
|---|---|
| Number of trades | At least 50–100 trades that followed the rules completely |
| Average result per trade (expectancy) | Positive after spread and commission |
| Maximum drawdown | Within the limit you set, for example 10% of the account |
| Discipline | Very few rule breaks, and no day past the daily loss limit |
Expectancy is the average result per trade. Measure it in R (R = the amount risked per trade) so it stays comparable when lot size changes.
Expectancy = (number of wins × average win − number of losses × average loss) ÷ total trades
Example: 60 demo trades, 27 wins averaging 1.8R and 33 losses of 1R each.
- Total = 27 × 1.8 − 33 × 1 = 48.6 − 33 = 15.6R
- Expectancy = 15.6 ÷ 60 = 0.26R per trade
- Risking 5 USD per trade, that averages 0.26 × 5 = 1.30 USD per trade
- Win rate 27 ÷ 60 = 45%, and the break-even at RR 1.8 is 1 ÷ (1 + 1.8) = 35.7% (the formula is explained in Win rate vs break-even)
It passes. That still doesn’t prove an edge.
Why 50–100 trades is only a minimum
A win rate measured over a few dozen trades swings a lot. A way to estimate the range the true value probably sits in (a roughly 95% confidence range) is win rate ± 1.96 × √(p × (1 − p) ÷ n), where p is the win rate and n the number of trades.
From the example above, p = 0.45 and n = 60:
- √(0.45 × 0.55 ÷ 60) = √0.004125 ≈ 0.0642
- 1.96 × 0.0642 ≈ 0.126, so ±12.6 points
- Range = 32.4% to 57.6%
That range dips below the 35.7% break-even, so after 60 trades you can’t tell an edge from good luck. If the win rate is still 45% at 200 trades, the range narrows to 38.1–51.9%, entirely above break-even.
That is why the first live account has to be small. 50–100 demo trades only tell you there are no warning signs yet. They don’t prove the system is good. Trading more often builds the sample faster, but costs then take a bigger bite per trade too.
A step-by-step plan for going live
Step 2 matters most. A small live account isn’t there to make money. It measures two things a demo can’t: the real fill cost for your system, and whether you follow your rules when the money is real. A cent account or 0.01 lot keeps the risk per trade to a small amount, but it is still your own money.
Compare trade by trade, not total profit. If you trade with an EA, run the demo or a backtest over the same dates and match trades one to one, looking at how far the actual entry price was from the intended one. If you trade by hand, write the intended entry price in your journal before you click, then compare it with the fill you got. The average gap per trade is the cost the demo couldn’t see.
Scale up in steps, for example 0.01 → 0.02 → 0.03, staying at each size for a while. Not from 0.01 straight to 0.10 after one good week. A bigger size brings emotions back into play.
Expect live results to be lower
Live results coming in below demo or backtest are normal, and not always a sign you did something wrong. The technical causes are covered in detail in Why live results usually trail the backtest.
In our own planning, we assume live results may keep roughly 40–60% of the backtest for systems that make little per trade. That figure is an estimate, not something measured on a live account. It comes from a gold breakout system that enters with stop orders. In our backtest over 5 years (real tick data only for the recent period; earlier years use ticks generated from bar data) it averaged about 0.77 USD per trade at 0.01 lot. We estimate that slippage plus spreads worse than the data would cost about 0.20–0.55 USD per trade.
| Extra cost per trade (estimate) | Left per trade | Left as % |
|---|---|---|
| 0.20 USD | 0.773 − 0.20 = 0.57 USD | 74% |
| 0.35 USD (middle case) | 0.773 − 0.35 = 0.42 USD | 55% |
| 0.55 USD | 0.773 − 0.55 = 0.22 USD | 29% |
The 0.773 average also includes older years run on generated ticks, which have made this same kind of system look better than it really was, so live results could fall below this table too. Backtest results do not predict future results.
Notice the extra cost is a dollar amount per trade, so a system that makes little per trade loses a larger share. A system making 5 USD per trade with the same 0.35 extra cost still keeps 4.65 ÷ 5 = 93%. When planning, think of costs in money per trade and subtract them from the average per trade, rather than knocking off an overall percentage.
The same works for the demo example above. If live trading adds 0.10R of cost per trade (hypothetical), the 0.26R expectancy drops to 0.16R, or 61.5%. At 0.20R it drops to 0.06R, or 23.1%. Your system’s real number only comes from the trade-by-trade comparison in step 2.
If your demo average per trade is already close to zero, a live account is likely to go negative. Don’t move until the system has a clear margin for costs.
A system that wins often for someone else may not for you
Another case from our testing. We saw the trade history of another trader’s gold scalping system: 66 trades, 84% won, a Stop loss 5 dollars from entry, about 85 trades a day. We coded the rules as far as we could work them out and backtested them on gold in 2026 with real tick data, a 3,000 USD balance and 0.10 lot per trade.
It lost 2,928 USD, about 98% of the account, over 1,954 trades, winning only 35.7%. The main cause was trading costs: the real spread was about 5–6% of the Stop loss distance on every trade, across thousands of trades a year. The original may have had tighter spreads, or exit rules different from our guess. But the lesson is that a tight-stop, high-frequency system depends more on fill quality than on its rules, and 66 trades is too few to trust an 84% win rate (by the same formula the range is ±8.8 points, before any costs). The original itself had a 61% drawdown.
All results in this article are backtests or estimates. Past results do not predict future results, and they are not returns you should expect.
Common mistakes
- Practising on a 100,000 demo, then going live with 500. The lot size and the feeling are completely different, so the demo results barely compare
- Moving because of a short winning stretch. Two weeks of profit from 15 trades can easily be luck. Use the checklist you wrote in advance, not a feeling
- Opening a full-size live account straight away. Skipping step 2 means you never learn the real cost of your system, and you meet the full force of your emotions from the first trade
- Expecting live to match demo. When results come in lower, people keep changing the rules, even though some gap is normal. Measure first whether the gap comes from costs or from not following the rules
- Going back to demo to win it back. Taking big trades on demo after a live loss to feel better trains nothing. If you hit your loss limit, stop as planned. More in Daily loss limit and revenge trading
- Panicking at the first losing streak on live. Good systems have losing streaks too. Check whether it is still within what the demo showed. Read Losing streaks are normal
Frequently asked questions
How long should I trade demo before opening a live account?
Count trades, not weeks. A sensible frame is at least 50–100 trades that followed your rules completely. At 3 trades a week that takes months, not two weeks. And that number is a minimum, not proof that the system makes money.
My demo was profitable, so why is my live account losing?
The usual causes are a different lot size, emotions that stop you following the rules (closing winners early, holding losers), and real costs such as slippage and the wider spread around news, which a demo may not show. Logging every trade on both accounts tells you where the gap comes from.
What is a cent account?
An account that shows its balance in cents. A 10 USD deposit appears as 1,000 USC, and 1 lot on a cent account is far smaller than on a standard account, so you can trade with a small amount of real money. Spread and contract size vary by broker, so check the contract specification for that account.
This is general education, not investment advice. Trading is risky and you may lose all of your capital.