Paper Trading: Useful, Until It Is Not.

Most demo accounts open with a balance nobody would ever deposit. A hundred thousand units of account currency, granted instantly, replaceable with a button. Every decision made on that account is made with an amount of money the trader does not have, cannot lose, and can restore in ten seconds. The results that come out of it are real in one narrow sense and misleading in every other.

None of which makes paper trading useless. It is the right tool for a specific and fairly short list of jobs, and the mistake is asking it to do the rest.

What it genuinely tests

Platform mechanics come first. Where the order ticket is, how to attach a stop and a target, how a pending order behaves, what a partial close does to the position, how the platform displays swap and commission. Learning that with real money is an expensive way to discover a mislabelled button, and any trader moving to a new platform should spend a week on a demo before funding.

The second job is rule mechanics. Does the entry condition occur often enough to matter. Does the stop distance fit the instrument. Does the exit rule leave you in trades you would have wanted to close. These are logical questions, and a simulator answers them adequately.

The third is forward testing a frozen ruleset. Once a strategy has been tuned on historical data, running it forward on data that did not exist during tuning is the only test that hindsight cannot contaminate. That is the last step described in why backtests overstate results, and a demo account is a perfectly good place to run it.

The execution gap

What a simulator does badly is fill orders like a real venue. Demo servers commonly fill at the requested price, with no queue, no partial fill, and either no slippage or a token amount. Live, your market order joins a book where the price you clicked may already be gone.

The difference concentrates in exactly the conditions a strategy is likely to trade. A breakout entry at the London open, a stop triggered during an economic release, an exit into a thin Asian session: each of these produces fills on a live account that a demo would have handled cleanly. For a system averaging a few pips per trade, that gap alone can be the difference between the two equity curves.

Some demo environments also quote a fixed or narrow spread while the live book widens with conditions. Anyone testing an intraday method should check whether the demo feed carries the same variable spread as the live one, because a strategy built against a permanently tight spread has been built against a market that does not exist. The broader comparison is in demo versus live accounts.

The gap the money creates

The larger problem is behavioural, and no simulator can fix it. Watching an unrealised loss on a demo is a mildly annoying number on a screen. Watching the same percentage on money you earned is a physical experience, and it changes decisions. Stops get moved. Positions get closed early because the profit feels too good to risk. Losers get held because closing makes the loss real.

The direction of the change is consistent. Traders take winners sooner and losses later once the money is theirs, which quietly inverts the risk-to-reward ratio their method depended on. That mechanism is set out in trading psychology basics, and the practical consequence is straightforward: demo results tell you what the strategy did, not what you will do.

This is why the common advice to "go back to demo after a losing run" rarely helps. The problem was not the rules. It was execution under pressure, and the pressure is absent on the account you have returned to. What usually works better is continuing live at a size small enough that the pressure is manageable while still being real.

Running a demo phase that counts

If a demo phase is going to inform anything, treat it as a controlled test rather than a rehearsal.

• Set the balance to the amount you will actually deposit, not the default. Position sizing on a fictional balance teaches sizing you will never use.
• Cover more than one market condition. A trend, a range, a scheduled event and a quiet holiday week beats three busy weeks in a single regime.
• Take every signal the rules produce, including the ones you dislike. Skipping trades on a demo is how a mediocre method produces an excellent record.
• Record each trade in a journal with the reason for entry and the emotion at the time. The emotion column is what you will compare against live later.
• Add a cost penalty to the results. Assume worse fills than the platform gave you, and see whether the method still stands up.

Then set a stopping condition in advance. A defined number of trades, or a defined period covering different conditions, after which you either fund a small live account or discard the method. Without that condition, demo phases extend indefinitely, and an indefinite demo phase is usually a way of postponing the part that matters.

Going live small is the bridge. A position size that produces a real but survivable loss reintroduces the pressure at a scale you can handle, and it is the only way to find out how much of your demo record was the strategy and how much was the absence of consequences. Trading leveraged products carries a high risk of loss, and starting small is a risk decision, not a lack of confidence.

"A demo tells you whether the rules work. It tells you nothing about whether you will follow them when it is your rent in the position."

— Alex Onta, Executive Director, SINGUARD

Frequently Asked Questions