Technical studies and correlations

L

Lonewolf

I'd be curious to know if there's a place where I can find out the correlations between the various technical indicators. For example, I'm using the following in my trading program:

Linear Regression Channel
Rate of Change
Moving Average Convergence Divergence
50 Day Simple Moving Average

Obviously the goal is to avoid relying on signals from indicators that aren't independent, thx for the colour.
 
Quote from Lonewolf:

I'd be curious to know if there's a place where I can find out the correlations between the various technical indicators. For example, I'm using the following in my trading program:

Linear Regression Channel
Rate of Change
Moving Average Convergence Divergence
50 Day Simple Moving Average

Obviously the goal is to avoid relying on signals from indicators that aren't independent, thx for the colour.
More...

Nice crutches.

Why don't you just buy at support & sell at resistance?
 
Quote from Lonewolf:

I'd be curious to know if there's a place where I can find out the correlations between the various technical indicators. For example, I'm using the following in my trading program:

Linear Regression Channel
Rate of Change
Moving Average Convergence Divergence
50 Day Simple Moving Average

Obviously the goal is to avoid relying on signals from indicators that aren't independent, thx for the colour.
More...

This is how I did it in the past when I was a newbie using indicators. Maybe not so rigorous but it gave me an indication of the correlation.

1. Determine the number of trades you get from two indicators separately, say X1 and X2

2. Add the two indicators and again determine the number of trades, say X12

The correlation is c = X12/(X1+X2)

c = 1 , no correlation

c = 0, anti-correlated

c = 0.5 fully correlated

and so forth, test another pair, add another indicator, etc.
 
Thinking about correlations is a key to a deep understanding of the markets.

Thinking about correlations of "indicators" in the sense mentioned here is helpful only to realize that they are all based on the same input: Past price.

From there on one can start thinking about what other information can be extracted from data feeds.
 
Correlations in the market is important, but that wasn't really my question. Every model, be it an indicator, VAR or monte carlo simulations is based on past events/occurrences.. If building a trading system its important to know whether certain indicators are more correlated or less correlated particularly when using multiple indicators under the assumption that they won't all show you the same thing.

In either case, its been a nice chat, thanks for the comments.

LW
 
If you happen to own a copy of Tushar Chande's book "The New Technical Trader", you'll be pleasantly surprised to find that its Chapter 1 is called An Abundance Of Indicators and you'll enjoy his discussion. Chande performs exactly the studies that the OP was asking about.

It's not all that hard to do it yourself either. Just have your Technical Analysis program print out the values of the various indicators you're interested in, over some long period of time (tens of thousands of bars), and import the data into Excel. Use the builtin correlation function in Excel (called CORREL, naturally). Problem solved.
 
You must log in or register to reply here.
Back
Top