Backtests and live trading almost always differ in cost. Measuring and closing that gap makes backtests more trustworthy.
Calibration process:
1. Run the EA live (or on a demo with realistic execution) for at least 100 trades.
2. For the same period, run a backtest with your current settings.
3. Match trades one by one. Compare entry price, exit price and cost.
4. Calculate average difference per trade.
5. Add that difference as extra slippage/commission in future backtests.
Typical sources of gap:
- Spread in backtest is fixed or based on historical data that doesn't match live.
- Slippage not modelled.
- Commission calculated differently.
- Swap on trades that run over.
If the gap is large (e.g. more than 30% of average profit per trade), the strategy is too sensitive to costs and may not be viable.
Repeat calibration every few months, as broker conditions change.
How big is the gap between your backtest and live costs?
Advertisement
Calibrating the live vs backtest cost gap
-
FXS London Scalper
- Posts: 806
- Joined: Sun Oct 18, 2026 12:25 pm
Calibrating the live vs backtest cost gap
FXS Official Team | London Session Desk
Recommended broker for automated trading & scalping
-
Shadow Trader
- Posts: 5438
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Calibrating the live vs backtest cost gap
The usual advice is to prove the strategy on demo, then go live with the same size. The part that gets skipped is that demo does not behave like live in the places scalping cares about most.
Demo fills are often instant and at the quoted price. Live fills come with queue time, partial fills and slippage on fast candles. Demo spreads can look cleaner because there is no real liquidity behind them, and news spikes frequently look gentler than they do on a funded account.
The bigger change is behavioral. A scalper who cut losses instantly on demo may hesitate when real money is moving. The strategy has not changed, but the execution has, and that is where most of the drop tends to come from.
A more honest bridge is a live account at the smallest size possible for a few weeks, with the aim of measuring the gap between demo and live fills, not making money. Compare entries, exits and spread at the time of each trade. That gap is the real cost of the strategy.
Scaling up should follow what the live numbers show, not what demo suggested.
Which part of the move to live was the biggest surprise, the fills, the spreads or your own reaction?
Demo fills are often instant and at the quoted price. Live fills come with queue time, partial fills and slippage on fast candles. Demo spreads can look cleaner because there is no real liquidity behind them, and news spikes frequently look gentler than they do on a funded account.
The bigger change is behavioral. A scalper who cut losses instantly on demo may hesitate when real money is moving. The strategy has not changed, but the execution has, and that is where most of the drop tends to come from.
A more honest bridge is a live account at the smallest size possible for a few weeks, with the aim of measuring the gap between demo and live fills, not making money. Compare entries, exits and spread at the time of each trade. That gap is the real cost of the strategy.
Scaling up should follow what the live numbers show, not what demo suggested.
Which part of the move to live was the biggest surprise, the fills, the spreads or your own reaction?