Statistical arbitrage a.k.a. Pairs Trading

With my limited Zorro knowledge accumulated thus far I assume that "pairs trading" is possible with Zorro.

That is : taking at least 2 assets, whose spread is cointegrating (mean reverting), buying the under performer and selling the over performer. That old gag.

For that technially I need to asset() calls
create the spread series (diff of the 2 assets close prices)

And do some linear regression / kalman-filtering to get the legs correct.

Is this strategy feasible to implement ? I mean ever building block is given by Zorro as far as I see. Please correct me If I am wrong.

Best,
Attila

Yes, that sounds like a serious strategy that should be straightforward to implement.

Ok, I'll dive under, I'll keep you posted on the results.

When using kalman filter for dynamic hedge ratio of a cointegrated pair, will you consider C++ libraries such as KFilter, or R packages such as KFAS, DLM, or code one by yourself?

Also, it's greatly appreciated if someone can demonstrate a simple example since I got stuck on this issue for a long time .

Somebody post any examples or link to see how to build the pairs strategy?

This sounds interesting. What are some example assets?
I might give it a try and will post my code if it works, like I did with the Dual Momentum script.

Dax and cac40 are two good assets for pairs trading. For my vision is also possible with eurusd and gbpusd.

We can build it together. We can try!

I seem to find several different hits for these abbreviations. Could you post yahoo finance links to the exact assets?

Those are stock indices. Either trade index futures or equivalent ETFs.

EDIT: Oh yeah, there's also some CFD's.

You can find it with the symbols below from yahoo.com

GDAXI
FCHI

The pairs trading is it possible to apply it at different instrument correlated and cointegrated

EURUSD + GBPUSD --> EURGBP.
It could stay stationary, but it could easily trend as well.

Ok, looked at GDAXI and FCHI. A couple questions:

- First thing I notice is that they are rather expensive. Are there cheaper etfs following the same curves?
- Their price is different in magnitude, while % change might be cointegrated. Still, which indicator do you propose to decide when the two prices are "in line" and when they are "apart"?
- I guess the general strategy then would be - sell bigger and buy smaller one, close position when they align again to win the initial discrepancy?

Ratio (spread) indicator between BB is very interesting to make pairs trading

You can use all symbols which are correlated and cointegrated between them

sell the strongest symbol and buy the weakest symbol till reach the medium line of BB. This is my main rule

Thank you for the replies. Excuse my ignorance, but what is BB? Could you describe a bit more details on how you propose to use a ratio to determine spread?

BB are the Bollinger Bands.
Later I post a screen on my use