Can a Computer Design a Trading System ? Part Three : Some Interesting Results Tackling JPY crosses

If you have read my blog for a while you may have noticed the fact that I have never developed an official strategy to trade JPY crosses. The problem with these pairs are many but certainly a very important reason why it has been difficult for me to develop such long term profitable strategies is the fact that the behavior of these pairs seems to have changed dramatically after the carry trade unwinding we saw in late 2008. Every time I have attempted to develop such strategies – except for some success with MA crosses – I have encountered systems that trade beautifully till late 2008, only to be pulverized later on as the behavior of the pair starts to shift from carry to a more “regular” mode. For this reason I decided to first test my genetic programming strategies to tackle JPY crosses to see if I could develop a set of daily strategies which could be used to deal with these pairs  in order to get long term profitable results that would work pre and post 2008. On today’s article I want to share with you some of the first finding I have made, showing that a computer can in fact design quite effective trading strategies (especially for pairs like these, whose behavior I do not fully understand).

If you haven’t read my previous posts on this matter I would encourage you to read this and this post where I explain the idea behind my genetic programming approach and how this can in fact lead to the creation of sound and robust daily trading strategies without over-fitting or generating some obscure trading strategies we cannot understand. Using the concepts highlighted on the above mentioned articles I used my genetic workshop to make my computer face the problem of achieving profitable JPY systems that remain profitable through the past almost 11 years. The pairs I chose for this task were the EUR/JPY and GBP/JPY as I have enough Alpari UK 1 minute reliable Bid data to do the whole analysis. (this article will only discuss the GBP/JPY results)

-

First of all, I have to say that I was completely surprised at the results generated by my computer and the way in which it decided to tackle the “JPY cross” problem. After an extensive “evolution” of the trading strategies being used, my computer decided that the best way to achieve profitability was to use trend and counter-trending techniques in a very complimentary way. The computer’s experimentation finally led it to a strategy based on the Stochastic and RSI oscillators in which the extreme levels of one are used to trigger trending behavior while the extreme levels of the other are used to trigger counter-trending behavior.

-

The above images show you the behavior of the system under a typical rangy period for this pair before 2008 and through the financial crisis and the unwinding of the carry trade. As you can see the system manages to succeed under both as under the rangy conditions trades that follow counter-trending behavior are most triggered and successful while under the unwind the trend following techniques start to take center fold. Definitely what I like the most about this approach is that the final trading techniques are absolutely transparent and easy to understand, just some very simple criteria which seem very logical to any experienced trader analyzing what the genetic workshop “came up with”.

As for performance, the system achieves quite good results with an average compounded yearly profit to maximum draw down ratio of almost 1:1, generally doing a very good job at preserving equity under draw down periods and letting profits run when favorable conditions start materialize. To tell you the truth, my computer was able to develop a strategy for the GBP/JPY which was more successful than any strategy I have attempted to develop for this pair, a strategy that is just as good as any strategy I could have come up with by manual analysis.

-

Certainly this genetic analysis tool is not a “replacement” for manual system development but it is certainly a great way to guide me across pairs which I do not understand very well. The computer is able to develop transparent systems in a very efficient way but it is always up to me to analyze the strategies and understand what they do and why they are successful in getting long term profits from a given pair. Not only does this give me insight into the inefficiencies present in other pairs but it also allows me to trade the strategies successfully under live market conditions since I gain a good understanding of what the system does.

This genetic programming framework will probably be released to Asirikuy during the beginning of next year as right now I am still experimenting and refining the techniques used to increase the transparency, efficiency and robustness o the results given by the workshop. If you would like to learn more about automated system development and how you too can create you own systems based on sound trading tactics please consider joining Asirikuy.com, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach to automated trading in general . I hope you enjoyed this article ! :o)

Print Friendly
You can leave a response, or trackback from your own site.

3 Responses to “Can a Computer Design a Trading System ? Part Three : Some Interesting Results Tackling JPY crosses”

  1. Winter says:

    Hi Daniel,

    Looking forward for your framework and your web site is simply amazing. Thank you for sharing your experiences and knowledge in the world of forex.
    Keep up the good work.

    Winter

    • admin says:

      Hi Winter,

      Thank you very much for your comment :o) I am glad you like the article and my website ! Hopefully the genetic programming framework I am working on will be another milestone in Asirikuy system development. Thanks again for your comment,

      Best Regards,

      Daniel

  2. […] with minors using my genetic programming framework (you can read more about it here, here and here) and what preliminary results I have been able to achieve. I will talk about the different pairs I […]

Leave a Reply

Subscribe to RSS Feed Follow me on Twitter!