A System for successful Range Trading of Vix Options

zenostiffler

zenostiffler

The Vix is mean reverting. Here is a simple system, drafted in Python, which buys Vix Puts when the Vix is over 40 and buys Vix Calls when the Vix is below a certain level. CAGR 20%, Max DD 27%. Scroll down for charts of the equity curve. Download the code from my Gist

 
I don't have the data. Its expensive stuff to buy from the CBOE. Also it very much depends how you concatenate the data into a continuous series. So I'm not getting too excited. I have done a huge amount of work on all this over the past year or so but I fear I am no closer to trading it!
 
I'd prefer to see more panda's data frames and fewer lines low level execution code. Do the funds who fund those online things think about it like this I wonder. 'factor analysis for predicting annual returns, a critical look, using log likelihood tests'. Que?
 
Although keep in mind the fundamental theorem of quant funds. For fixed assets as the number of strategies approaches O(1/Pe), no strategies matter therefore diversification is a free lunch.
 
TommyR said:
I'd prefer to see more panda's data frames and fewer lines low level execution code. Do the funds who fund those online things think about it like this I wonder. 'factor analysis for predicting annual returns, a critical look, using log likelihood tests'. Que?
More...
The purpose was to make it simple. So that all people need is contained in the Notebook. As it is a great deal has been abstracted away through libraries. The idea was that people can download this and extract the Pandas dataframes to Excel. I do not think there would have been much point in displaying a few lines of the dataframe in the notebook.

Or perhaps I am missing the point. Are you suggesting I should replace the looping code with vectorisation? If so, I agree, obviously, that vecotorisation where practical is the better approach.

I'm not at all sure what point you are making here. Would you mind expanding and clarifying?
 
Last edited:
TommyR said:
Although keep in mind the fundamental theorem of quant funds. For fixed assets as the number of strategies approaches O(1/Pe), no strategies matter therefore diversification is a free lunch.
More...
You still have to devise the strategies before you add them into a mega strategy. Unless of course you use a random strategy approach such as a nature inspired algo which simply tries anything and everything and gradually learns. Assuming learning of a complex system is possible. Such an approach is useless if you can not predict the outcome of a deterministic complex system. But yes, diversification. Agreed of course. But that is not a real option for the small guy.
 
TommyR said:
Although keep in mind the fundamental theorem of quant funds. For fixed assets as the number of strategies approaches O(1/Pe), no strategies matter therefore diversification is a free lunch.
More...
Incidentally, I am not convinced that a "trading" approach to the financial markets is sensible or "correct" at all. I'm not convinced by this range trading strategy either. The range will shift and change. If you take a look at the chapter I wrote for Harriman House and their New Book of Investing Rules you will see my scepticism writ large!
 
You must log in or register to reply here.
Back
Top