Hello, I have developed an indicator external to tradestation that only trades on certain days, like 100 days a year. I backtested it by hand and found it to be up to 86% accurate, and a real money maker. The problem is that there are inherent biases in manual backtesting because the parameters that trigger the trade can be a bit subjective. I want to take this technique and then filter the trades with an indicator and then backtest it. I hired an easy language programmer to develop a system for me but it didn't actually work, and now I am out $200. The concept is fairly simple, I will produce a series of dates, up to 100 in the future and then easy language will be programmed to search through the dates and only trade on the days that are allowed. I think you just use an array, and some simple commands. Can anyone help point me in a direction that will help? Thanks, Andrew
If your trading days are based on some common factor that can be programmatically determined using calendar/date functions, that is a better way to go (e.g., first trading day of each month). Otherwise, hard-coding the dates in an array is the next best option. Any easylanguage programmer should have been able to code this for you.
I bought Tradestation Made Easy and Building Winning Trading Systems books, read them and have a basic grasp of the language. Unfortunately, they don't give much info about arrays and I am not sure how to code them correctly. If I could find some code for a single array, I am sure I could work it out. Thanks for the reply.
Programming is a complementary skill to trading. And programmers are resourceful, they know how to use Google to find code snippets. If you're not going to pay a programmer, you need to spend some time acquiring programming skills before you proceed.
Yes, exactly. It should take 2 minutes to write what he wants and he's going to need to write it a thousand times before he realizes what he should b writing instead of what he's thinking of at this moment.
Thanks for your assistance in this matter. I don't mind paying a programmer financially, I just don't want to throw cash at bad code. My technique took a long time (Thousands of hours and years of life.) to develop so I don't want to give it away either. I did get it backtested by hand, and it does work with a high degree of accuracy. My other concern is that I gave a guy one of my other techniques, and he turned $15.00 into $2,500 in about two months. He quit because he got a letter from his broker accusing him of insider trading because every trade made money. He also ended up selling a bunch of my materials on EBAY. One has to be careful. Regards.
At some point during all that time u spent u should have made an effort that exceeded the effort u put forth. Best thing u can do at the moment is get the knowledge u need. U won't succeed otherwise. U can't buy your way in.
If this is the concern, then you can always prepare a modified version of your system that does not reveal your secret sauce. Create some phony rules that require similar logic as your real rules. When the programmer delivers the code, you will have a working program that you should be able to easily modify.
Thanks Guys. The backtesting was an afterthought. In hindsight I should have allocated more time to learning programming. Suggestions are quite helpful thanks..