What is forward testing and why isn't a backtest enough?

A backtest tells you how your strategy behaved in the past. Forward testing tells you whether it works when you don't know the future — which is the only test that truly counts. Skipping this step is how strategies that are "perfect" on paper die in the live account.

Why doesn't a winning backtest guarantee anything?

Because the past already happened and you know it. It's easy, even unintentionally, to tune the rules to fit exactly what occurred — and that "perfect" strategy only describes history, it doesn't predict the future. Forward testing breaks that trap: you test on data the strategy has never seen, in real time, without the advantage of knowing how the candle ends.

What exactly is forward testing?

It's applying your strategy going forward, trade by trade, as prices arrive. There are two ways:

• On demo (paper trading): you execute the trades in a practice account, with fake money but real market conditions. Zero risk, ideal for validating before you risk capital.
• Live with minimum size: you trade for real but with the smallest possible position. It adds the one factor demo can't replicate: the emotional pressure of having money on the line.

What does it reveal that the backtest hides?

Things the clean history doesn't show: how wide the spread is at your actual entry, how much slippage you take on news, whether your rules are executable live or require decisions you don't make the same way under pressure. And above all, your own behavior — whether you really follow the plan when the money is real, or whether fear and greed break it.

How long should you forward test?

Long enough to build a sample that means something — not three lucky trades. The same metrics that matter in the backtest apply here: if the forward test's win rate, R:R and profit factor resemble the backtest's, you have a strong signal the edge is real. If they collapse, the backtest was overfitted — and you just saved yourself from losing real money finding out.

The bottom line

The backtest is the hypothesis; forward testing is the experiment. First you validate against the past with an honest backtest, then you confirm going forward on demo or minimum size, and only then do you scale up. It's slow, which is why almost nobody does it — which is exactly why it works.