Generating Ideas for Automated Trading Strategies

June 26th, 2008

This is the third in a series of posts about how and why and how I started down the path of automated trading. Here’s the first, second, and third posts in the series (RSS). Our automated trading robot is available now. Trade your own strategy automatically with no coding required.

When I first started trading, I worried that I didn’t have enough market knowledge to come up with a strategy on my own. Since then I’ve traded a manual system that I’ve made some tweaks to to make it more consistent. There are some important changes that I made to it, but nothing that I would consider groundbreaking. I know people have differing opinions about this, but I think a lot of traders abandon strategies too quickly. There are so many variables that can be tweaked in a trading system. Of course, some are irrelevant to the strategy and modifying others just over-optimize your trading system. I don’t think there’s anything wrong with tweaking your system – perhaps a lot – if the changes make sense to the logic behind your overall strategy.

Everyone wants something that works right now and makes money every day without taking the time to develop a strategy. The problem with this approach is that if you trade a strategy profitably for a while and then you run into a drawdown (they are inevitable), you’ll give up on it precisely you haven’t done the work required to develop it. Drawdowns are valuable learning opportunities but you have to be willing to swallow your pride and be humble – if not then the market will gladly humble you forcibly.

There are plenty of strategies all around you that will work pretty well, at least as a starting point. Just think about what you see occur in the market every day. There is a tremendous amount of price movement everyday. Are some of those price movements tradeable? Probably so. Test them and see if they are. You’ll know pretty soon if something is worth more testing time. A lot of them won’t be, but some will be. A lot of it is counterintuitive as well – some ideas that you’re just sure will pan out won’t test well at all while others that don’t make any sense at all at first will test really well.

Once you start brainstorming for automated trading strategies, you’ll start gaining some momentum – those ideas will generate other ideas and strategies. A couple of my better automated strategies came from testing ideas that weren’t at all related to my initial concepts. One was purely accidental – I had a concept for a strategy and when I modeled it, it turned out that I modeled it incorrectly. I didn’t discover this until I had pretty thoroughly tested and refined the strategy. It turned out to be a pretty profitable system, but the event that I was trading wasn’t at all what I started out trying to capture.

I’ll write about how I’m doing testing in the next post.

3 Comments

  1. Atd Said,

    July 18, 2008 @ 5:13 am

    I’ve been thinking about automated trading but am still very limited in my programming capabilities so the automated trader seems very interesting for me. Can you give me some idea of how flexible this system is?

    Does the automated trader let you trade futures? How about fibonacci (generated based on the 30 min open range for e.g) based entry/exit? Can the system identify test re-test on an entry point?

    Thanks.

    Andy

  2. Dave Said,

    July 20, 2008 @ 4:03 am

    @Andy:

    The robot sounds like just what you’re looking for. It doesn’t trade futures yet but it does have Fibonacci capabilities. Take a look at the Trade-Ideas Odds Maker – it’s a backtester that is at the heart of the automated trading robot.

  3. Ran Automated Trading Robot Unattended For First Time - StockTickr Blog Said,

    August 7, 2008 @ 2:47 am

    […] Generating Ideas for Automated Trading Strategies […]

RSS feed for comments on this post