A few days ago I made a post about the first Amachay portfolio I devised and how this setup brought to us very good profit and draw down targets for this trading system. However after several tests it was evident that the Amachay logic was quite complicated and several measurements were taken to simplify Amachay in order to make it a much more elegant counter-trending strategy. Within today’s post I am going to share with you the reasons why I found the system to be over-complicated and why better results were in fact achieved after I carried some extensive simplification of the strategy’s framework. Within the next few paragraphs you will learn why the new version of the strategy is much more elegant and why its results are better.
As I have mentioned before Amachay is a counter-trending strategy based on Fibonacci lines. The first version of this strategy I evaluated used a combination of two positions one which had a move-to-break even trailing stop and the other which aimed for a higher-level profit target. I also said – within the past few posts about this system – that this trailing stop mechanism seemed to be the “best choice” for this strategy even though for most strategies it is not the best decision. The first tests I carried out did show it to be better although the problem seemed to be that I was not carrying out the proper optimization procedure to let the strategy show its potential under a more simplified approach.
After I started my thorough evaluation of the first portfolio I made for this strategy it started to become apparent that the trailing stop was indeed more of a problem than something positive. On several currency pairs the end-level placement eliminated the trailing stop completely (making the move to BE point and the profit targets the same) so it became obvious that this move-to-break even wasn’t the best solution after all. I subsequently removed the second position and made the system take only one position per entry signal, however this didn’t show better overall results. Later on I realized that the problem was related to the range of Fibonacci levels I was testing and after increasing the range (almost doubling it) I finally started to see results which were much more profitable than for the first developed portfolio overall.
Amachay was therefore reduced from a complex two-order system with a move-to-break even stop mechanism to a single position system without any trailing stops. The complexity of the system was greatly reduced and this made it become much more “elegant”. Once this was done optimizations were done much quicker and upon testing my new optimization procedure I found out that I could consistently get better out-of-sample results than with the first version including the move to BE. The move to BE therefore proves again to be detrimental to mechanical strategies as it reduces the probability of most trades to reach their full potential and just causes premature exits on many trades that would have been successful. I have commented about the fact that the move to BE is a bad decision for algorithmic systems overall (long term profit sacrifice for psychological easiness) and these observations and changes on Amachay make me even more fond of this conclusion.
The results of the simplified Amachay portfolio are actually even better than the results for the first one although the number of profitable pairs remains the same (6). The system achieves an average compounded to maximum draw down ratio in the range of 2-3 (although higher than the first portfolio) and the trading logic and system complexity are enormously reduced. I have to say that I am much happier with this Amachay portfolio because the strategy now has the level of simplicity which makes most Asirikuy systems such elegant trading solutions with simple entry and exit logic sets that make them able to exploit inefficiencies with great transparency.
It is also very interesting to note that Amachay is still a strategy with an unfavorable risk to reward ratio (about 2.3:1) and a relatively high winning percentage making it a very good match for Sumaq (another strategy released next week which I also discussed on a blog post last week) which has a very favorable risk to reward ratio with a very low winning percentage. Overall these two counter-trending strategies approach the problem of fading the trend in very distinctive ways with Sumaq attempting to catch big reversals and Amachay attempting to capture small retracements. Note however that although the risk to reward ratio on Amachay is unfavorable its average profit is still above 10 times the spread so interpolation errors do not account for profitability. The system also passes the 3x spread widening tests meaning that its inefficiency is not derived from the assumption of low spreads.
Next week Amachay and Sumaq will be released to the Asirikuy community, giving us two very useful additional tools to trade the Forex market. With Amachay now being such a simple and elegant trading strategy with totally opposite trading characteristics when compared with Sumaq, the potential for very interesting portfolios as well as an increased level of diversification on current Atinalla ones opens up. Of course we will have to spend a few months testing Amachay and Sumaq portfolios before venturing into any sort of “joint effort” but certainly having 2 new counter-trending strategies with such opposite characteristics will be a great addition to our trading arsenal.
If you would like to learn more about my work in automated trading and how you too can design your own systems and evaluate them using formal statistical tools 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)