Over the weekend I spent more time on research and development of trading systems. Although certain ideas seem to be coming along, I am having trouble programming it correctly in to mechanica.
One trading model that I have been excited about is the counter trend model I talked about early. Although the code still needs some modifications, the initial batch of results are promising. I took the monthly returns from 1982 to 1992 and compared it to that of my core trend following system and found the correlation coefficient to be ~0.2, which is great news. The uncorrelated returns means that the two strategies will complement each other and definitely improve the risk/reward ratio.
I also came across multiple entry and exit ideas ranging from breakout to adx filters. To be honest, the more and more I research about quantitative trading models, the more I find facts that support the idea of “keep it simple.” Back in the days when I was still relying of discretion to extract profits, I would tirelessly research the perfect setup, the elaborate entry, and being right 100% of the time. IMHO, success in this business revolves around several core ideas of trading (breakouts and MAs). The further you venture out, the less robust your models will be as complexity brings forth curve-fitted models.
Also, I found that my system 1 is giving back too much profits which really is irritating. I need add on additional scaling out techniques to prevent this from happening. I was up more then 150 pips for both AUD/USD and USD/CAD but then I was stopped out at breakeven for the AUD/USD.
Anyways good trading, I will report back later.