Adapting to Part-Time Trading: Automate

Well, I sat in a courtroom for most of the day ready to pick a jury and get going with a trial just to be excused at 4 p.m. and told to come back on a certain day in November. What a waste! I knew that being a lawyer would conflict greatly with trading and I've contemplating several ways to deal with the problem and today, I was reminded of how bad it could affect trading. First, I thought I would trade on longer time frames. This might work and I'll probably be able to use it to some extent but using monthly, weekly and even daily charts does not quite fit my trading personality. I do have some great daily and weekly strategies but I only see them as a side project. The intra-day (day trading) time frames are what really work for me.

For quite a while now I've been toying with automated trading. Recently, I've spent a lot of time getting familiar and experimenting with Meta Quotes' Meta Trader program. This is very versatile and programmable forex software that allows you to program indicators and trading strategies (i do not have any affiliation with this company nor do I receive anything for recommending them) and it even lets you back-test the strategies. To top it off, you can set the software to execute your automated trading strategies in a live account, which is what I'm very interested in.

So far, I've had some success with trend following and support/resistance trading programs but I must admit, the trend following systems seem to work the best. I've been using the old Donchian 5/20 ema cross over system and tweaked it some to get it to fit with my intra-day time frame. I've named my system the "Donch Crosser" and below is a screen shot of a chart that's being traded in real time and a back tested equity curve below. The back-tested equity curve began with a balance of $5,000 and it traded from 1-01-08 through 10-27-08.
As you can see, the initial $5,000 turned into approximately $17,000 in 10 months. Now that's some trading!!! The max draw down for the period was 22.84% and the win ration was only 42%. This just goes to show that they don't all have to be winners, you just have to make sure the winners run and the losers are cut short. I've had the most success with this program by running it in the hourly (1h) time frame. Although the hourly is the best, because it has a GREAT return with a relatively small draw drown in that time period, the system has also been very profitable in the 1, 5, 15, and 30 minute time frames and also the 4 hour time frame. In fact, the only tested time frame that was not profitable was the daily time frame and I think this probably has more to do with a fluke in the parameter settings or data than it does with the system.

What this testing shows is that the system is not getting stellar returns simply from curve-fitting the data. It is at least somewhat robust which gives the system a fighting chance for being successful in the future. Although this system looks incredible on paper and it's at least somewhat robust, it will still likely incur more volatile swings and draw downs in the future compared to the last 10 months of back-tested trading. The key is to determine what is statistically normal and expected through back testing in multiple time frames and instruments so that you can be prepared to face what the future brings. This is much easier said than done.

Only time will tell and I'm sure there will be plenty of unexpected surprises but it seems that the "Donch Crosser" is showing some promise and hopefully it will continue to prove itself as a robust and profitable automated system.

Good luck out there.

TLT