Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
If you are referring to my own trading, I use several markets/tickers side-by-side to trade ES. But to simplify them to solely "correlation" is a mistake. Sometimes they are correlated or inversely correlated, but other times they give you an idea of where money is flowing. I don't think it is very practical to track such things solely on an intraday basis, you need to look at the broader picture. These moves tell a story over time of days and weeks, letting you know how the market is shifting. The market is always moving in search of alpha, your job as a trader is to understand everything in a broad context and form hypothesis then trade them and learn if you were right or wrong.
I agree with most everything that Adam said from a general point of view, but I do not know his exact testing methodology and can't comment on it.
Here is a rolling 20 day correlation between the S&P and 10 year returns. Of course there is much noise in market data.
To me market regimes are basically sets of correlations between instruments that reflect how fundamental economic forces are playing out.
From the chart above for me I like to buy when things are hairy and sell when things are normal. Or at least not be selling when things are hairy. Like right now I like that the markets have not dipped much when the 10 year correlation has got completely out of whack since the last fed meeting. Now the edge would be in knowing if things will go back to "normal" or if things have structurally changed. That kind of thing is not testable or knowable IMO.
I would say correlation can be really advantageous of a tool. I also disagree that there is not "edge" to be found in it. Its a tool as anything in trading, it depends what your using it for. I am assuming that you are calculating the correlations correctly using ratio back adjusted contracts, and using log returns.
When people think of using correlation as an edge they think pairs trading, obviously correlation is going to be important. In this case though having co-integration and finding a unit root is more important, but your rarely going to find 2 instruments that are co integrated that are non correlated.
But more importantly I think correlation tools come more into their own when you look at portfolio building. Correlation between any 2 or more given instruments held in the portfolio, and correlation between any 2 or more strategies traded on the portfolio. No only does VAR risk management use a co variance matrix(related to correlation) but it is a requirement for all institutions. We can discuss the pros and cons of using VAR in the first place and tail stuffing but that's for another post. If you do not measure the correlation you can not gain anything from diversification, either in additional rewards or reduced risk.
Most people are solely looking at the correlation coefficient, is it .80 or .2 or -.73 etc. I think that's not painting the entire picture, you also need to look at its stability. As with all things in the market nothing is fixed, but is fluid. So you must measure not only their value but also their possible range of values over some look back. even 20 days, is a fairly short time frame given we are looking for larger market relationships. 90 days or 120 days are IMHO more important as single instrument event driven volatility will be averaged out more.
Take 2 correlation coefficient values for example:
Today A = .90, but over the last 180 days has ranged .45 to .95
or
Today B = .70 but over the last 180 days ranged between .65 and .75
Which one would you rather use as a baseline or investigate for a trading edge?
Most people would shun the .70 because it is only moderately significant (<.80) but it is much more stable. But you know exactly what you are getting into, while A at .90 is highly significant but can alter dramatically.
Also don't forget you can use linear regression or modelings as well to find the relationship between 2 instruments, its simple and easy. Correlation is not the only tool for this.
Very good post! I have followed the U.S. Dollar and its correlation with the s&p500 index. These normally follows each other, but the last months the correlation gone from +.6 to negative correlation.
My understanding, is investors are selling the dollar and going to other currencies less valued. The stock market in Brazil and China have going up very well the last months and the sp500 range bounding. My idea by selling the ES and purchasing the Brazilian stock market has been quite profitable.
Recently there has been a tendency for the ES to rally when the euro rallies, as the dollar is getting weaker, and vice versa when the dollar is getting stronger, ES tends to weaken.
But historically, the dollar and ES have a low correlation.
This is the main support thread for the atgPairsCorrelation ( indicator for NinjaTrader.
The atgPairsCorrelation indicator plots the correlated logarithmic returns of two data series. The inputs to the indicator are standard price series data. The …