Tweet
Folks,
I have not been on the site for a little while, I have been working on getting my first EFS automated, and up and running. I started the process full speed around the beginning of November, and it has taken me about 6 months to get things to a point where I can release the EFS on the market.
I thought I would share some of my travels, and invite others to discuss.
I just read some of the posts from http://forum.esignal.com/showthread.php?s=&threadid=743 and I agree that not many want to disclose their code, and as soon as I can, I will post a sanitized structure-version of my code up in my share location on http://share.esignal.com/groupconten...er=&groupid=9, but it will be more of what I had to go through to make it work, and time permitting, a tutorial of sorts.
My code is bound to the company I work for, and as such proprietary. But here is how I made it all work. Just some ideas.
I started with this site:
They have TONS of indicators all written in EFS code that show trading systems based on any indicator out there. I have a set that I like as a trader, so I started with those. I looked through all 200 at the time in one weekend, and messed around with the ES futures, which I what I trade primarily. I narrowed down about 8 which work really well for what I trade, for other things something else may work. I personally hate stocks, and I trade futures due to their nature to adhere better to the math of formulas.
I also looked on this site:
I read anything a daytrader wrote, and thought about it. I love daytrading, it is my main goal, but it is a pain to code with safety.
Then I learnt EFS, and this took a while to become proficient. What was worst, was that I had to learn how the EFS worked, really the nuts to bolts. I did this by looking at code from tutorials, messing around, and LOGGING everything. For my EFS I have about 5 log files that log every aspect of the code's decision making process, and all key variables that control code loops. This took the longest, about three months to accurately debug.
So then I had trading ideas I wanted to program, an understanding of EFS, and all the log files which tell me whether the program is making money or losing money.
I worked to a goal of have three programs running concurrently; a fast trader, medium speed trader, and an overnight swing trader. Between the three, one will catch what the other misses, and so forth.
I use stops, but I read Brad's note, the larger the stop, the better, and I use a 10-point stop currently.
If I ever want to code a new program, I just start with either a chart or an EFS that already shows entries and sells. The next important thing I work with is to look at the underlying indicators on a simple chart, and see where they cross into overbought and oversold regions and see if I can formulate a rudimentary set of rules.
That is my first run for the code, and then I look at the trades the system makes on those rules, and refine, and refine, and refine some more. I design everything to start on a $3K account, with no more than a $1K drawdown, profitable after 8 weeks, and I test for 2 months.
Aside of that, I look at any chart any one ever posts and see if I can get some ideas to code something unique. I think the most important is to understand scripting, and how the code works. This just takes time and experimentation.
That is just the way I picked up my knowledge, maybe it can help some following behind. One other thing, I always have a goal for weekly profit. If I tell you that I making 20 points a day on the ES futures, well this may be, but if I am making 700 trades, that is not much use to me. I always shoot for no more than 20 trades a day if I am trading rapidly, and a profit of 20 points a day. I have never attained this reliably, but I will keep trying. I am close, but I never attain more than 13 points reliably, per day, averaged over 5 days.
Well, I hope this helps, I will keep reading, keep sharing. Good luck all,
Thomas.
I have not been on the site for a little while, I have been working on getting my first EFS automated, and up and running. I started the process full speed around the beginning of November, and it has taken me about 6 months to get things to a point where I can release the EFS on the market.
I thought I would share some of my travels, and invite others to discuss.
I just read some of the posts from http://forum.esignal.com/showthread.php?s=&threadid=743 and I agree that not many want to disclose their code, and as soon as I can, I will post a sanitized structure-version of my code up in my share location on http://share.esignal.com/groupconten...er=&groupid=9, but it will be more of what I had to go through to make it work, and time permitting, a tutorial of sorts.
My code is bound to the company I work for, and as such proprietary. But here is how I made it all work. Just some ideas.
I started with this site:
They have TONS of indicators all written in EFS code that show trading systems based on any indicator out there. I have a set that I like as a trader, so I started with those. I looked through all 200 at the time in one weekend, and messed around with the ES futures, which I what I trade primarily. I narrowed down about 8 which work really well for what I trade, for other things something else may work. I personally hate stocks, and I trade futures due to their nature to adhere better to the math of formulas.
I also looked on this site:
I read anything a daytrader wrote, and thought about it. I love daytrading, it is my main goal, but it is a pain to code with safety.
Then I learnt EFS, and this took a while to become proficient. What was worst, was that I had to learn how the EFS worked, really the nuts to bolts. I did this by looking at code from tutorials, messing around, and LOGGING everything. For my EFS I have about 5 log files that log every aspect of the code's decision making process, and all key variables that control code loops. This took the longest, about three months to accurately debug.
So then I had trading ideas I wanted to program, an understanding of EFS, and all the log files which tell me whether the program is making money or losing money.
I worked to a goal of have three programs running concurrently; a fast trader, medium speed trader, and an overnight swing trader. Between the three, one will catch what the other misses, and so forth.
I use stops, but I read Brad's note, the larger the stop, the better, and I use a 10-point stop currently.
If I ever want to code a new program, I just start with either a chart or an EFS that already shows entries and sells. The next important thing I work with is to look at the underlying indicators on a simple chart, and see where they cross into overbought and oversold regions and see if I can formulate a rudimentary set of rules.
That is my first run for the code, and then I look at the trades the system makes on those rules, and refine, and refine, and refine some more. I design everything to start on a $3K account, with no more than a $1K drawdown, profitable after 8 weeks, and I test for 2 months.
Aside of that, I look at any chart any one ever posts and see if I can get some ideas to code something unique. I think the most important is to understand scripting, and how the code works. This just takes time and experimentation.
That is just the way I picked up my knowledge, maybe it can help some following behind. One other thing, I always have a goal for weekly profit. If I tell you that I making 20 points a day on the ES futures, well this may be, but if I am making 700 trades, that is not much use to me. I always shoot for no more than 20 trades a day if I am trading rapidly, and a profit of 20 points a day. I have never attained this reliably, but I will keep trying. I am close, but I never attain more than 13 points reliably, per day, averaged over 5 days.
Well, I hope this helps, I will keep reading, keep sharing. Good luck all,
Thomas.
Comment