After almost one entire year since my last contribution to Currency Trader Magazine, today I have the pleasure to announce a new article that has been published on the October issue of CT. This article is very important for me because – not only does it mark my return to full-time trading – but it is also the first article I write in the company of one of our most beloved Asirikuy members, Fd. The fact that this article was written as a collaboration makes me very happy because it speaks about the spirit of Asirikuy and what is achievable when team efforts come into play. Today I want to write a blog post about this article, about why I believe it is relevant for algorithmic traders out there and what is probably going to come after this first move back into the world of CT publishing. As always, the article is available for free so make sure you download it here if you want to. For those of you interested in reading more about my contributions you can also buy a full collection with many of my CT contributions here (please note that I do not get any commission from these sales).
When we were first discussing the topic of the article with Fd, he suggested the idea to explore algorithmic support and resistance and ways in which the validity of S&R methods could be determined statistically. This topic is quite important because trades (including me) tend to talk a lot about S&R and the importance of this technique when trading manually but there is certainly no agreed upon method to define what S&R is and how these levels might be used in trading. The idea to test which methods would work through statistical means was very appealing to me since it would allow us to design new trading methods as well as to validate if manual/algorithmic methods to define S&R levels had any substance to them. Another interesting thing is that we wanted to see whether S&R defining methods could in fact predict levels that would behave as S&R and not simply generate a profitable trading technique, this means that we wanted to know if there is truly some way to predict levels where price will bounce/breakout in the future.
Overall this wasn’t a very easy task and it took us about a month of bouncing emails and having conversations about the subject before the picture started to become clear. Of course it didn’t help that I was in the middle of a trans-Atlantic move but thanks to Fd’s hard work, we were able to keep our focus and take the article to its culmination. In the end what we came up with was a method to evaluate S&R algorithmic techniques based on bounce-breakout patterns that is able to tell you whether a method you have for defining S&R zones or levels does have some edge in predicting levels where price will bounce or breakout. The article doesn’t provide any S&R technique that works but it does in fact provide you with a tool that will allow you to test whether you actually have one.
In the article we also used an example deriving S&R levels through high/low levels to show that a technique that may sound good and put levels in the screen that may appear to make sense, doesn’t actually predict anything beyond random chance. This opened my eyes to the true problems of S&R defining techniques. While writing this article we found out that price has a natural tendency to “bounce” in the EUR/USD, so even though you may think you have a predictive edge, you are most likely predicting bounces that are expected from the natural characteristics of the trading instruments you are looking at. This has put my manual S&R definitions into perspective as well – as when put through this statistical testing – they show that what I am seeing is merely a product of what I want to see. For me this article was particularly interesting and its development process particularly enlightening in that it changed the way in which I now see S&R trading and definition methodologies.
With an article now explaining and providing a way to statistically evaluate S&R, it is now the task to evaluate other methods, polish the evaluation methodology and attempt to find a true algorithmic technique that does show S&R levels that predict points/zones where price is going to bounce/breakout in the future. There are already several candidates but it has become clear that this is not an easy thing to achieve. It seems that S&R levels/zones may change with time and therefore taking into account the age of S&R tests is fundamental to develop a definition methodology that may have some actual statistical edge in its predictions. Overall I believe the article shows that S&R definitions are not an easy thing to do and that showing that you have a statistical edge from an algorithmic S&R technique is actually quite hard.
In the future expect more article collaborations between me and Fd and possibly single articles from both of us in the years to come. This was a very nice collaboration and hopefully other Asirikuy members will have some nice ideas to collaborate in the future as well. If you want to learn more about our developments in algorithmic trading and how you too can increase your understanding about trading system evaluation and design please consider joining Asirikuy.com, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading in general . I hope you enjoyed this article ! :o)