Historical Testing

by Scott Owens

Risk assessment is a critical element of money management. To gauge risk in advance of live trading, sophisticated investors employ a historical testing system. However, not all historical systems are capable of delivering complete risk assessments, and choosing the wrong tool could lead to poor live trading results.

Content

ANALYSIS

Use historical testing as a risk assessment tool.
Learn how historical testing works and can be optimized.
Understand the potential pitfalls of historical testing.

ACTION

Test your systems to get baseline metrics for risk performance.
Optimize your systems to minimize drawdowns and consecutive losses.
Trade your systems live with a few caveats about historical testing.

RELATED MATERIAL

Test-drive FX Engines for free online at www.fxengines.com to see the power of system building, system testing, and system automation.

About this Report

The Forex Report is a periodic publication that investigates advanced strategies for superior trading performance in the foreign exchange markets. These reports utilize advanced statistical and econometric modeling techniques to create new insight into the trading strategy of the average trader. This Core Concept Brief, Historical Testing, is intended for traders with all levels of forex trading experience and technical analysis understanding.

To learn more about The Forex Report or to register for delivery of all future reports by email, including Case Studies & Data Briefs, please visit www.fxengines.com

Analysis

Historical testing is one of the most powerful tools in the trader’s arsenal. Using a large amount of historical data allows the trader to build systems that are fundamentally sound and expand upon them. Along with the benefits of historical tests come a few warnings about the dangers of using historical results irresponsibly.

ASSESSING RISK
Each time a trader places an order, a number of factors combine to form a risk profile for that particular trade. Chief among these factors are position size, volatility, drawdown potential, and recent events. For a trader with a long history of live trading experience with a particular system, these dynamics are well known. But the trader who deploys a new system usually does so without the advantage of this live trading perspective. For these traders, historical and live tests are the best substitutes for actual trades.

HISTORICAL TESTS
Historical tests, in particular, provide a rich analytic framework for ascertaining a system’s ability to cope with the factors that influence risk. A historical test is a tic-by-tic re-enactment of a trading system’s performance over time. The best systems, like the one offered by FX Engines, use multiple years of tic data, employ that data in a real re-enactment with real trading constraints, and work in a way that the trader can replicate in a real-time, real-money account.

A good historical test provides a wealth of data which must be scrutinized by the trader to identify patterns that can be used advantageously in further optimized tests. The most telling metrics are net pips, maximum drawdown, consecutive losses, and success rate. These metrics give the trader enough information to make a rapid determination of the system’s worth. If the system is obviously bad, another direction can be chosen. If the system is good or looks to have potential, further scrutiny is needed.

The statistics of a back test are convenient, quick ways to gauge a system’s value, but there is no richer data to mine than the trades themselves. By looking at each trade the trader can get an idea of what happened, even with only a handful of data points. Using this data and the metrics from the test, a course for system optimization can be devised.

In some cases the methods for optimizing a system are obvious – adjust a stop, change an exit signal, change the entry schedule, etc. In other cases the methods are not so clear, and that’s when an automated optimizing system is of tremendous value. FX Engines, in particular, has such a tool, the Back Test Multiplier. This tool takes a number of different engines, breaks them down into their component parts, then recombines them into many more engines. This method creates systems that the trader might not have had the time or creativity to discover otherwise.

Once a system has been created and optimized through historical tests, a period of live testing is required. Only after verifying the conditions predicted by the historical test should a system be traded in a real account. Even then, real trades can deviate from the historical trades in many ways...