Demo accounts lie, politely

[Fairman]

Demo accounts lie, politely

Instant fills. Any size. No rejection, no last look, no partial fills, no queue. Nobody is on the other side of the trade, so the platform simply says yes.

A scalping approach can look beautiful on demo for a month and then fall apart in the first live week. The strategy did not change. The execution did.

Demo still has uses: learning the platform, building hotkeys, testing entry logic. It cannot say anything about how a real liquidity pool treats an order, and pretending otherwise is how accounts get blown.

Best approach is to move to the smallest possible live size quickly and treat the first month as an execution test, not a profit exercise. That month's cost is simply the tuition for learning how fills behave.

The deeper problem is psychology. A strategy that wins on demo builds confidence that does not transfer. Position size creeps up in the first live week, the first rejected order feels personal, and the first slipped fill gets blamed on bad luck instead of structure.

A cleaner path: trade the live account at the minimum lot for two to four weeks and compare each trade with what a demo would have produced at the same moment. Run both in parallel if the platform allows it. The gap between the two equity curves is the real execution cost of the strategy, expressed in money.

That number tends to be bigger than expected, and it is usually the number that decides whether a scalping approach is viable at all. Better to learn it at 0.01 lots than at 1.00.

Some traders also find that the demo-live gap is not constant. It widens at news, around rollover, and during thin sessions, which are exactly the moments a backtest quietly assumes are fine.

How wide was the gap between demo and live results the first time it was measured?