Quantopian how tos, page-3

Getting there slowly! To be perfectly honest it doesn't seem to be working too well. Over-fitting is a big problem (which makes sense because most patterns in price are due to chance). I need to find a way to filter out noise. Recently I read a blog where someone had had some success with using multiple RSI's of different time periods so will try something similar eventually.

Heres part 2 of above.
http://www.relativevalue.com.au/blog/what-about-quality
Some code at the bottom for dollar neutral version of ev/ebitda