In a previous post, I wrote about how automated forex trading can help remove the emotion and requirement to constantly watch the forex markets.  One product that I listed was FXCM’s Forex System Selector, which is a web based automated trading tool that allows the user to select one or more trading systems, and have them automatically traded against an account.

I’ve now had an opportunity to start using this, and this article is about my first impressions of the product, and some of the systems on offer, and what a trader should know about the product.

At present, the markets are closed so I can’t comment on how it performs in real life, but we shall soon see since it is now running against a demo account.

Firstly, I would recommend viewing the video about the product before starting to use it.  It is not self explanatory, and there is no wizard or similar step by step guide.  That said, it is not difficult to use once you understand the concepts.  Essentially, you choose several trading systems from the large number on offer, set parameters for the systems (overriding stop losses and money allocation rules), then check the effect on the historical equity curve.

You should aim for an equity curve that is as smooth as possible, particularly without large drawdowns.

A number of the systems on offer are not suitable for trading.  Some of them have massive drawdowns (thousands of pips), or just lose money.  You cannot expect to pick some at random and magically become a millionaire - you need to understand what to look for.  Other systems have very few trades to evaluate historical performance against.  Overall, many systems do not have a historical record going back for more than a few months.  Trade Robot is better in this respect in that some of the system track records go back several years.

I chose my systems based on Profit Factor (which is a column on the systems report) - the higher the profit factor, the better.  At the same time, I eliminated some systems with huge drawdowns, or very low profitability per trade.  Generally, this seems to return some reasonable systems.  I selected 10 systems trading the following currency pairs:

• GBP/USD
• NZD/USD
• AUD/NZD
• CHF/JPY
• EUR/JPY
• USD/CAD
• GBP/JPY

The next step is to see how the systems perform together.  The idea is to obtain diversification through the systems using different parameters and trading a range of uncorrelated currencies.  You can do this by looking at the equity curve chart.  The one for the systems I chose looks like this:

You will notice that the system initially loses over $2,000 (or 20% of account equity) before increasing equity up to $43,000.  This result was obtained after tweaking - if you just choose random systems, you will get a far worse result.  The chart is over a limited time frame.  This is because many of the systems did not have a track record before the beginning of 2008.

One concern is that the process of selecting systems is done in hindsight.  If you choose systems with only a few trades, you run the risk of curve fitting the portfolio system - it performs well on paper, but may not do well in practice.  At this point, this is easy to do as few of the systems have long track records.  All that you can do is try to choose systems that have as many successful trades as possible, and diversify across systems and currencies.

There are some options to increase equity given to successful systems, while decreasing that for the less successful systems.  My primary concern with this is with small accounts, it may not be possible to fine tune asset allocation to this level.

System performance on the weekend is ok without being fantastic.  Even with a very fast connection, it is nowhere near the performance of a system running on your own PC but acceptable for a web application.  I would like to see how it goes during the week in the busy times.   The platform does take advantage of AJAX / Web 2.0 technologies, so it is interactive.

Overall, the system selector concept has promise, but I think that traders need to consider system selection carefully (targeting low drawdowns rather than the highest performing / highest risk systems) and carefully considering system track records and longevity.   I shall see how it performs against the demo account over the next month.

(No Ratings Yet)

 Loading ...

Tags: automated forex trading, Forex Trading, FXCM

This entry was posted on Sunday, April 6th, 2008 at 7:42 am and is filed under Automated Trading. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.