Optimizing correlation limits
-
- Roundtable Fellow
- Posts: 98
- Joined: Fri Oct 20, 2006 2:25 pm
- Location: London, UK
Optimizing correlation limits
Classically to find robust parameters values during an optimization test, CAGR or MAR is used as the goodness measure. The goal of the test being to find the parameters values giving “highâ€
Can you say what you hope to achieve by installing correlation limits? Why are you doing it? What good do you expect will result?
Once you are able to state your objectives, you will have an easier time deciding how to measure the results.
Flip through some math books about optimization; you'll find that the very first step in an optimization is to identify "the objective function". My favorite such book is (link to Fletcher), but the chapter about optimization in (link to Numerical Recipes) is quite good too.
Once you are able to state your objectives, you will have an easier time deciding how to measure the results.
Flip through some math books about optimization; you'll find that the very first step in an optimization is to identify "the objective function". My favorite such book is (link to Fletcher), but the chapter about optimization in (link to Numerical Recipes) is quite good too.
-
- Roundtable Fellow
- Posts: 98
- Joined: Fri Oct 20, 2006 2:25 pm
- Location: London, UK
Sluggo,
Thanks for the links and your help. Your wise answer is indeed very helpful in pointing me in the right direction.
What I want to achieve indeed is testing a LTTFS with a “smallâ€
Thanks for the links and your help. Your wise answer is indeed very helpful in pointing me in the right direction.
What I want to achieve indeed is testing a LTTFS with a “smallâ€
- Attachments
-
- There is a lower zone of drawdown < 34 % in the lower left quadrant. That is where I would have picked my parameters. But the value of 6 max closely and 10 loosely correlated market described in the turtle manual are in fact in the max drawdown area.
S - Closely Vs Loosley Correlations.JPG (56.38 KiB) Viewed 3477 times
- There is a lower zone of drawdown < 34 % in the lower left quadrant. That is where I would have picked my parameters. But the value of 6 max closely and 10 loosely correlated market described in the turtle manual are in fact in the max drawdown area.