Forward-testing on history without fooling yourself

A backtest is only worth anything if the trader running it knows as little as a trader living through the day. That sounds obvious. In practice, information from the future leaks into almost every manual backtest people run, and it leaks in the same five ways.

1. Seeing the candle that is still forming

On a normal charting platform the current candle is drawn as it builds, which means that at 14:20 you can see where the 14:00 hourly candle *currently* is. A trader on the day could not; the hour had not finished. That one glimpse is enough to make a bad entry look fine.

TradifyFX never draws a candle until it has closed at your clock time, on every timeframe at once. If you want to watch candles build, tick-by-tick printing shows the current one growing minute by minute — from the one-minute candles inside it, not from the finished hour.

2. Scrolling ahead

The most common leak, and the hardest to admit to. You scroll forward “just to see”, then scroll back and place the trade. The replay clock only moves when you move it, and stepping back does not undo a trade — your record is your record — so there is no free look.

3. Choosing famous dates

Everyone knows what gold did in April 2013 and what the pound did in March 2020. Testing on those weeks tests your memory. Pick dates you have not studied; the date picker goes back to 2005 on the majors, and the less you remember about a month, the more honest the result. Use the well-known weeks for what they are good at — stress-testing your rules — not for measuring your edge.

4. Changing the rules mid-test

A strategy that gets adjusted after every loss is not one strategy; it is a sequence of strategies, each fitted to the last few trades. Write the rules into the session before you start — number of trades, risk per trade, the hour you stop — and let the discipline score measure you against them. If the rules need changing, change them, give the new version a new strategy name, and start the count again.

5. Ignoring costs

A test without spread flatters every strategy and destroys the honesty of any that takes small, frequent profits. The replay charges the spread on every trade by default — you buy at the ask and sell at the bid, half on the way in and half on the way out — and the session summary shows you what it added up to. Leave it on. The only reason to turn it off is to measure how much your costs are costing you.

The rule that covers all five

Trade the replay the way you would trade a live account, with the future genuinely hidden, and let the journal record what you did rather than what you meant. The engine closes the mechanical leaks. The behavioural ones are yours to close, and the journal will show you whether you did.