How do live results differ from backtests?

When it comes to trading strategies, two types of results are often shown: backtest and live trading. At first glance, both look like confirmation that the strategy is working. But there is a very important difference between them.

What does the backtest show?

A backtest is a test of a strategy using historical data. In other words, the algorithm is run through the past market and seen how it would trade if it worked during those periods.

This is a useful stage: it helps to understand whether the idea has at least some logic, how it behaves in history, where it has weak points. Backtest is needed. But this is still a model, not real trading.

What does live show?

Live results are real trading on a live account, in a real market, with real quotes, delays, execution and risks.

This is where the strategy encounters not the “ideal story”, but the real conditions: slippage, spread changes, news, technical nuances and market unpredictability. Therefore, live results are usually valued higher.

Why are the results often different?

• Over-optimization: the parameters were adjusted too closely to the history.
• Real trading costs: spread, commission, slippage, execution speed.
• Changing market environment: what worked historically does not have to work the same way later.

Does this mean that backtesting is useless? No. It is useful as a screening and initial screening step. But the problem begins when they try to pass off a backtest as full-fledged proof of the quality of a strategy.

The healthiest approach is to look at both the backtest and the live test, but understand the difference: the backtest answers the question “is there logic in history”, the live one answers the question “does this work in the real world”.

Bottom line

Backtest is a useful model of the past. Live results are a test of the strategy by the real market. If you choose what to trust more, then, other things being equal, it is live, real trading that is more important.