Back-Testing

Back-testing is a process in which a set of mechanical technical rules are applied to a period of historical price data of a particular financial instrument.

This is performed by analyzing the results of the back-test to gauge how profitable that set of technical rules would have been for the particular time period.

Thus back-testing aims to provide confidence over the system’s sustainability when trading in real-time.

For a financial trader to have confidence in their system or strategy, they need to be able to test their rules in their market of choice, before deciding whether to pursue this in a live environment.

Undoubtedly, the testing of one’s strategy in real-time is the most accurate in determining how feasible a trading system is.

This presents obvious issues however, since for some strategies, especially those involving longer timeframes, would take an extremely long-time to execute.

Additionally, forward-testing a system before the trader can attain the necessary threshold of confidence needed to trade it live, i.e. with actual money presents enormous levels of risk.

Why Use Back-Testing?

This is where back-testing comes in. Most broker platforms have a built in back-testing module, which allows the trader to test their rules on an instrument over a fixed time period.

The back-tester is then able to crunch this data, by applying all the rules to the asset and time period.

This results in a detailed report of how successful the system would have been, often all in a matter of a few minutes to a few hours.

Key statistics pumped out by the back-tester include, percentage of trades won, percentage of long trades won, percentage of short trades won, largest winning trade, largest losing trade, average length of trade, consecutive wins and losses.

The most popular platform that offers back-testing is the MetaTrader platform, where traders can codify their systems’ rules into an “Expert Advisor” (EA), followed by running that EA on MT4’s or MT5’s Strategy Tester.

In theory, back-testing provides an almost perfect solution to seeking out profitable, tradable strategies for the long term.

The problem is, back-testing makes a flawed assumption – that past results equate to similar future performance.

However, back-testing can arguably, at the very least, help optimize a promising system.