Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
My PyArb stat arb module

Howdy Quantopian folks!

I just uploaded my PyArb python molule to GitHub here. This is basically a generalization of pairs trading for an arbitrary number of stocks. It's also "model dependent", i.e. instead of trying to find mean reverting portfolios from eigenvalues of a covariance matrix, it's assumed that the prices are "interacting" according to a vector geometric brownian motion.

I'd be interested in trying out the model in Quantopian, but I don't yet know the Quantopian code very well... so if anyone wants to help with that, let me know!

2 responses

Hello Heikki,

Would you be willing to provide an outline of the approach? What are the major elements of the code? How does it work?

Grant

Hi Grant,

the IPython Notebook html version of the intro should explain the approach, so I won't repeat that stuff here. I understand that I still need to clarify several points, which I intend to do as soon as I can find the time for it. I'd be happy to answer any more specific questions here though, so feel free to ask!