Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Efficient Frontier: what is it actually optimizing?

Hi team,

I followed the following in optimizing a portfolio:
https://www.quantopian.com/posts/the-efficient-frontier-markowitz-portfolio-optimization-using-cvxopt-repost-cloning-of-nb-now-enabled

See my version:
https://github.com/yanwu26/public_code/blob/master/python/Efficient%20Frontier.ipynb

I am trying to understand what it's trying to understand what the optimizer is trying to optimize. The QP returns a weight vector. I calculate the return by using weight.T.dot(pbar) and then I calculate the standard deviation by using weight.T.dot(S).dot(weight). When I plot the intersection, it's not anywhere on the efficient frontier.

I would expect that the optimal optimal portfolio is based on the risk free rate, that is tangential to the efficient frontier curve:

https://www.glynholton.com/wp-content/uploads/2013/06/exhibit_capital_market_line_1.png

Does anyone know why the portfolio thats generated doesn't seem like it's correct? Thank you.