http://wikiposit.org/p?futures
It seems pretty amazing that they even have historical term structure data.
Has anyone worked with their data and care to comment on the data accuracy?
Yes. If you look into the Matlab code, the hedge ratio is simple the slope of the regression line. So in excel you can just add a trendline to the regression and set the intercept to zero. I have verified this with Chan's example of GLD/GDX.
It seems like PTF bases its calculations heavily on the "ratio" of two equities. However, isn't there a problem with the arbitrary choice of which stock is the numerator and which is the denominator?
I agree it's a good idea to keep things simple and I don't intend to get too overly...
When people talk about correlation, they usually mean "return" correlation, not "price" correlation. So it's quite possible that returns are correlated but the prices deviate substantially. I find gummy's website a very good reference to understand some basic concepts...
Just a general newbie question to the PairTradeFinder users.
Most of the stat arb papers use "cointegration" as the test for mean-reversion to find potential pairs. However, PTF uses only simple linear correlation to scan. What's your experience with mean-reversion behavior from pairs found...
If you are really that good with C++, why don't you try to get into a more famous school like UCLA or USC instead of CSULB? Even though I don't necessarily believe in brand-name, sometimes it just makes things easier, especially when all the career ads emphasize "top tier" schools.
Are you implying your stat arb model is actually scalable in terms on holding period? Wouldn't the expected profit decrease with longer holding period since the model ideally works with higher frequency?
My understanding is that high-frequency strategies are either market-making or stat arb. Do you agree or are you aware of any other type of HF strategy? I am apparently a newbie trying to satisfy my curiosity.
Since high-frequency trading requires very low commission and very good execution...
I agree with you in general. I think what matters is the ratio of expected holding period and the backtesting length. Since high-frequency strategies can generate more trades for a given period, the sample size is larger and the backtesting length does not need to be as long.
Just out of...
This is a little off-topic, but can you describe, in general terms, any precaution that needs to be taken when backtesting with high frequency data? I assume you are working with tick data from the orderbook. How do you make sure that the trades in backtesting would actually be executed in real...
Very interesting. Some schools even offer stat arb as a course:
http://cims.nyu.edu/~almgren/timeseries/
http://public.tepper.cmu.edu/Shared/ViewCourseDetailsStyled.aspx?CID=5470
I wonder if they teach any profitable strategies in those classes. :p
Anyone here taken those classes and...
I am not sure how high-frequency your data needs to be, but AmiBroker is the fastest backtesting retail software that I know and it supports tick data.
http://www.amibroker.com/guide/versions.html
This is great insight.
The more I work on my retail trading "business plan," the more I realized that the odds are stacking up against the retail traders due to various reasons that you are well aware of(infrastructure, support, among others). That's why I try to go the opposite route as...