Hello, for my StatArb setup I am also looking for stable intraday stock pairs where you can generate consistent profits from them doing pairs trading. Beside known ones like PEP & KO, which are other known pairs ? Is there also a kind of website here ? I know I can look for same (sub)industries, similar market caps too, and look for high correlations among them. But I would like to know if that work is already done somewhere and already published anywhere ? As I could not find any good source here. I am looking for stock pairs which are good for intraday StatArb. Of course I could do all the work myself but with over 1000 of possible combinations but it can take quite a long (computing) time. That kind of intraday especially should work on the 1 or 5 minute time frame, what I am looking for, but also higher times frames are welcomed if they really can produce consistent results. So all in all they really must be strong fundamentally related. Otherwise it does not make sense I am guessing. Do not know if looking for shares with the correspondent ADRs are good combination here in addition ? Anyone can contribute to this topic/these questions something please ? Thanks in advance.
Some inversely correlated tickers: BOIL and KOLD SOXL and SOXS I hadn't time to apply pairs trading in practice yet. But I did an IMO important research on it. I mean I wrote a program to find the best correlating tickers among all US tickers (more than 4000 tickers). I did this some months ago, and made IMO an important discovery: there are mainly 3 methods to use for the correlation analysis: 1) stock price 2) change in stock price 3) change in stock price in percent My research result is saying that #3 is the best. The surprising thing was this: the result of #3 differed very much from the standard correlation methods (Pearson, Spearman). For example it found very good correlation (and it indeed was correct when verified manually) whereas standard methods said (wrongly) that there is barely a correlation btw. them... Or something that, can't remember all details exactly now after such a long time of many months, but essentially such a surprising discrepancy in the method results I did find. And: I used EOD data, not intraday, and my research was private, not published. Unfortunately I haven't found time yet to apply these IMO important research findings to real-world practice, as it's a little bit too much work. I've reserved this for my future team of co-workers, which I anticipate to form after making some money by solo trading as I'm currently doing. If someone is interested in trying this out, PM me for clearing the details.
What I didn't do yet (and will soon test as well) is this case: 4) ln(thisClosePrice, prevClosePrice) ie. using the logarithmic ratio, which also gets used in the Black-Scholes-Merton (BSM) option pricing formula. And: I'm using SMA to smooth the time-series data, so that outliers (spikes) don't falsify the overall result.
%% Some thing like that. SH with a simple ma; SPY + or SPYG which is a bit different...........................................
What counts is not whether others can generate consistent profits, but whether TrAndy2022 can and more importantly does. In general non-specific terms, What's your method Anderoo22? TrAndyMan-Can?
I don't understand what you're saying here - you used percentage returns but didn't use correlation? Isn't the correlation of percentage returns the standard way of doing it?
Yeah, you are right. It's indeed confusing. I'll have to look into my results and notes and will then reply. I think I just came up with a new correlation method different from the standard one, b/c of the research results proving that the standard one did not work and that there is a better method... Will check...
StatArb. And yes I can and do already, but I also look for further improvements. It can always get a bit better than now is my thought. It is always the same thing as when you do some money you want more or could do more, so it can always be better than actual. If you are happy with a fixed income let me know. So you are not looking for more growth and stay there where you are ? It should get really pretty boring when you always make the same amount over years. That is I always want (a little) bit more. Now it is your turn, in general non-specific terms, what's your method ?
Never stop refining, yep. I missed the part where you state in general non-specific terms what your pairs trading philosophy is. Do you look for counter trending pairs? Counter proportional moves? Do you trade the product or it's options? or ...? Examples?
It would get too specific I think (and too much lines/writing here to tell you). See also no reason to be more detailed here. By the way what is your method or strategy non-specific ? But in short I am looking where I can get the best edge or highest sharpe (with StatArb). Hope that helps. If you know the topic about it you might know what I am doing already.