Back to Community
Implementing recognition of technical patterns question


I assume that proper implementation of technical patterns and other elements of price action trading could prove to be very powerful in terms of algorithmic performance. I was wondering if anyone help me learn how to code trading logic that would be able to recognize and account for such things e.g. a cookie-cutter 3 push bull channel with a spike up and 2 swing points forming a trendline.

I understand that this might be asking for a donation of too much "secret sauce" so I was hoping someone could point me in a general direction of what I would have to study and learn in my pursuit of such trading logic.

I would presume that machine learning would play a significant role in this; I'm just not sure where exactly to start.

Any and all help is appreciated!

3 responses

For (future) users of quantopian who also frequently google "quantopian -relevant search terms-" in the hopes of discovering something on the forums that could help them with an issue:

If you also are interested in learning how to write trading logic capable of recognizing technical patterns, I have found another thread making good progress on a related topic.

If I catch wind of anything or make my own significant progress, I will update this thread out of goodwill to the many people like me who do not know of any other starting point in research other than google :)

For price patterns there is excellent software from PAL that generates Python code for Quantopian IDE. I have tested it and runs very well.

Thank you for this Sol Marte, this is quite a magnificent find. Just curious, is there any other software you use or know of that you would recommend? "Paying to play" is of no inconvenience (hopefully).