Algorithmic Trading Strategies Online

skynetrading · Post by **skynetrading** » Wed Oct 17, 2012 11:18 am

Hey Artemiso

I have picked up Johnson's book and it is going to be my next book to read in my list. I am going to start using python to play around with different types of strategies with Quantopian's data. I understand it is limited data but I believe it is a great place to start gaining some hands on practice. After I gain familiarity with the concepts and posting some results of my skills and determinations on a blog than maybe we can talk about furture collaborations. I will keep you updated =)

dandunn · Post by **dandunn** » Wed Oct 17, 2012 5:12 pm

Just to clarify - the Quantopian data is minute bar data, not daily bar. We have the last 10 years of US stocks. We don't have survivor bias - it includes all traded companies. It also includes mergers and spinoffs, which can be pretty tricky. One example is IAC, which Google and Yahoo both make pretty confusing, but Quantopian corrects nicely: https://app.quantopian.com/posts/smooth-data-for-iac

We'll add more data sources in the future.

allmyid · Post by **allmyid** » Fri Oct 19, 2012 4:39 pm

Maybe you may find our project interesting: we are developing a tool that should help systematic and discretionary traders identify strategies and get fresh ideas.

It is based on graph+network theory. We launched a small blog to show our approach and the kind of ideas our tool could help you generate, as well as to get feedback on the path the project will take.

Feedback, in particular the negative one, is highly appreciated.

You can check the blog out at www.genchi.io

ricko · Post by **ricko** » Mon Oct 22, 2012 9:40 pm

One very interesting approach is the one followed by Michael Harris in his blog. He uses a tool to identify price patterns with sufficiently small profit target and stop-loss to avoid curve-fitting returns. You can possible develop your own tool for that - it is not necessary to use his. Then, he checks the significance of the patterns he identifies against a portfolio of instruments to both increase sample size and reduce survivorship and data-mining bias. This is an example for DBC ETF and this is another example for the ASX 200 Aussie index.

I find the approach useful because it involves well-defined exits and the aforementioned tests.