My approach is a highly configurable supervised learning approach. Before writing any code, I have a good idea of what ideas I want to test and how to test them. The application just automates the effort. The product will ship with all of the indicators I use...it will just be up to the user to use them as they see fit (and of course write their own).
Adventures in Automation
- Thread starter fan27
- Start date
Nope. I average two to three trades a week. Some weeks there could be no trades. I am targeting high probability strategies which turn out to not occur very often. That being said, having enough of them makes it worth while.
I am trading 70 strategies. They are just low frequency, hence the low trade count. Some only trade when above the 200 day SMA and some only trade when the previous week is a 52 week high.
And your trading universe is just futures of ES, NQ and YM?
Have you looked at Buildalpha?
https://www.buildalpha.com/
I think your process is similar to Buildalpha. Buildalpha is mainly for daily data, yours incorporates intraday data. They can filter thru' more than 5000 signals in pretty short period of time.
Yeah...I have seen it. Looks like a cool application. The main difference between what I have and BuildAlpha is it is a desktop application with fixed functionality where as I have an SDK with application/source code. My solution is more targeted to users with programming abilities who want maximum flexibility.
Yes...for now. I have not decided yet if I am going to expand into other markets. At some point I might want to trade OPM and am wondering if it is better to have a focused approach.