Allow myself to introduce myself,
I have been reading these forums at ForexFactory for a few weeks now, and am very impressed with the level of knowledge being shared. I have already found several great ideas floating around that I would not have thought of otherwise. My main interest is in autotrading, and due to this, I thought it would be good to give a short background on myself and start a thread where those interested in autotrading could shared ideas.
My main goal is to LEARN as much as possible, and it seems that by presenting some of my ideas, getting feedback and starting discussions is a great way to do this. I have read some of the other threads on autotrading, and it seems that these use commercial platforms, while I am taking a different approach…
I traded FX manually for about 3 months demo and 6 months live. I ended up at zero profit, due to very poor money management techniques. I was making a lot of small gains, and getting wiped out by over leveraging myself on the wrong ends of trades. Eventually, I decided not to trade manually anymore since I have other full time pursuits. Now I am studying a PhD in computational biology, where I make machine learning models of genetic and regulatory networks. I also studied engineering and worked in software development.
For the last 2 years, I have been working on creating an autotrading system. I have the structure and control system created, but have not gone live due to inadequate demo results. Now I am working on two different models that I will test simultaneously, based on two different forms of machine learning; both look very optimistic. The minimum profit I consider safe to run the system live is 100 pips per month, with a target of 400 per month. Any less than 100 could be dangerous or unreliable in the long term, I believe.
1) I think that it is essential to code everything from scratch. The reason for this is that it is very difficult to control code creation using commercial tools and know exactly what is going on. Even backtesting is tricky if one cannot be certain how it works. I use C++ and R for my work, and highly advise everyone doing this to learn a few languages and control every part of the process. Python is much easier and faster to learn than C++ and Matlab is simpler and more intuitive than R (although expensive, unless you are a student). I should also mention that my strategies probably can’t be coded in EasyLanguage or something similar.
Any comments, or should I continue?
Thanks for you attention,
-Duke
I have been reading these forums at ForexFactory for a few weeks now, and am very impressed with the level of knowledge being shared. I have already found several great ideas floating around that I would not have thought of otherwise. My main interest is in autotrading, and due to this, I thought it would be good to give a short background on myself and start a thread where those interested in autotrading could shared ideas.
My main goal is to LEARN as much as possible, and it seems that by presenting some of my ideas, getting feedback and starting discussions is a great way to do this. I have read some of the other threads on autotrading, and it seems that these use commercial platforms, while I am taking a different approach…
I traded FX manually for about 3 months demo and 6 months live. I ended up at zero profit, due to very poor money management techniques. I was making a lot of small gains, and getting wiped out by over leveraging myself on the wrong ends of trades. Eventually, I decided not to trade manually anymore since I have other full time pursuits. Now I am studying a PhD in computational biology, where I make machine learning models of genetic and regulatory networks. I also studied engineering and worked in software development.
For the last 2 years, I have been working on creating an autotrading system. I have the structure and control system created, but have not gone live due to inadequate demo results. Now I am working on two different models that I will test simultaneously, based on two different forms of machine learning; both look very optimistic. The minimum profit I consider safe to run the system live is 100 pips per month, with a target of 400 per month. Any less than 100 could be dangerous or unreliable in the long term, I believe.
1) I think that it is essential to code everything from scratch. The reason for this is that it is very difficult to control code creation using commercial tools and know exactly what is going on. Even backtesting is tricky if one cannot be certain how it works. I use C++ and R for my work, and highly advise everyone doing this to learn a few languages and control every part of the process. Python is much easier and faster to learn than C++ and Matlab is simpler and more intuitive than R (although expensive, unless you are a student). I should also mention that my strategies probably can’t be coded in EasyLanguage or something similar.
Any comments, or should I continue?
Thanks for you attention,
-Duke