What is Backtesting?

What does Backtesting mean in practice?

A backtest replays history candle by candle, applies your entry, exit and risk rules, records each simulated trade, and totals the result. It answers whether an idea was internally coherent over the period tested, and it is a cheap way to discard bad ideas.

What it cannot do is predict. It is a simulation of one past, under one set of cost assumptions, on one data set. Results are usually more optimistic than reality because history has no outages, no queue position and no rejected orders, which is why a second check on live prices matters.

A worked example

Illustrative only: a breakout rule is run over two years of 1-hour candles and produces 140 simulated trades. The report shows expectancy, profit factor and maximum drawdown. None of that says the next 140 trades will look the same.

How to read or test it

Read the report in layers. Start with trade count, since a result from a handful of trades says almost nothing. Then look at expectancy and profit factor, then maximum drawdown, and only then net profit. Change one parameter slightly and rerun it: a result that collapses is fragile. Finally compare it with a paper-trading run of the same rules, because the gap between the two is the honest measure of how optimistic the simulation was.

How AKQX handles Backtesting

AKQX backtests deduct estimated fees and slippage from each simulated trade, report expectancy, profit factor and maximum drawdown, and score the result out of 350 points. Results are simulated and are not a forecast.

See Backtesting with fees and slippage for the feature page.

A common mistake

Treating a backtest's profit as an expected return. It is the output of one simulation, not a promise.

Related terms