Thanks for the kind words and thanks for letting me know about Adaptrade! The app looks very sophisticated. It is feature rich and a good product for someone wanting a fixed feature set. The one thing that is missing from all of these similar products is a lack of ability to strategically combine strategies into a portfolio of strategies. I think that is and edge my app and approach has.
"The one thing that is missing from all of these similar products is a lack of ability to strategically combine strategies into a portfolio of strategies. I think that is and edge my app and approach has." That's a very good point.
Exited the following trade at 6651.25 just now: Long 1 NQ contract at 6643.75 on 1 hour chart Fixed stop at 6519.50 Exit after 30 bars While it gave up most of the gains, the epic winning streak continues with a fantastic month so far! Will post full results at the end of the month.
fan27, Since you are trading Index Futures, did you program or the ability to see if there was any above average intra-trade DD on holidays/half days or FOMC days? Curious to know if historically those days caused any spikes in risk exposure to your portfolio of systems.
Any Tradestation/Easy Language developers here, interested in system generation, may find this article of value. Easy Language code is provided. http://www.adaptrade.com/BreakoutFutures/Newsletters/Newsletter0907.htm "Automated System Development In the July 2006 issue of Futures magazine (ref. 1), I read an interesting article by Mike Barna. He described a computer program he developed that automatically generates trading systems using a genetic programming algorithm. Very few details about how the algorithm works were provided, and the article was laced with technical jargon. However, the basic idea was clear. He was using a computer algorithm to construct trading rules and combine them in different ways in order to find the set of trading rules that worked best. It was optimization of the trading system development process, rather than optimization of trading systems per se. He followed that article with one the next month that provided an example of a trading system produced by this process (ref. 2)."
I'm not understanding how this is done in practice. I can understand developing a system and then using the software to come up with "best" stop method for example. But snooping for trading systems sounds wrong to me. This is more of a question than opinion.
Did you read the entire article? Perhaps the question would be best directed to Mike Bryant, author of the article. mrb@adaptrade.com If you are concerned about curve fitting, walk forward testing could address that.
My approach is to first come up with a thesis of market movement (i.e. Breakout, mean reversion, etc) and then determine the technical or price conditions that MIGHT profitably model my thesis. I let the software/machine test the various conditions that I determined might make up a strategy. So in my approach, the "snooping" is very targeted.
I now have AlgoTerminal running on an AWS instance. Also, I made a major change in that all live trading events are logged as JSON and the resulting log data is added to AWS CloudWatch. Having the log data as JSON will allow me to query log data in very granular ways. For example, I could query on all order events for a particular portfolio and strategy. Also, I can set alerts on any number of events or errors. In case of some sort of critical issue, I have the AWS console app on my phone and I can stop the running instance within seconds. The instance is an m5d.large and it's cost will be about $50 per month to just run it during RTH (9:30 am - 4:15 pm Eastern).