Update on scans

I'm (Another Brian) back from a one week vacation. Thank you to "The Word" for the detailed posts and market insight provided for the TSX. It is interesting how the TSX does not track the US markets step for step.


I am in the process of back testing some of the scans that I have previously posted here on this blog. In addition, I am working on a system that I will be back testing, and posting results of the back test. I have been educating myself on the back testing process, it can be as complex as you want it to be. The first thing I noticed is the similarity to simulation. I have programmed discreet event simulation at work for material handling systems and I know that exactness is not that important. The idea of simulation is to have data input parameters somewhat accurate and the results will give you a great idea of the systems processing capabilities. If you can provide rough input, you can expect the output to be at least 80% accurate in simulation. I recently viewed a presentation on back testing that was very educational, and although the stock market is far from simulating material handing, there are allot of similarities. I am using Amibroker to do this.


Here is a link to the files for these presentations. The files include 2 presentations, power point files, and sample files. I recommend anyone who is back testing watch these two movies, even if you don't have Amibroker.
Presentation 101
Presentation 201

There is a free plug-in for Amibroker called IO. It is used to speed up the optimization process of your variables. The videos also discuss the results, the goal to set for results, what your goal should be based on (CAR, Sharpe, Ulcer, etc), the process of walk forward back testing, and tons of other good information.

For those of you interested in learning more about Amibroker, here are a couple of links. Yahoo also has a few groups.
Amibroker main site
Amibroker forum - a separate website, called amibrokerfan.com

I probably will not be posting regularly since I will be working on the system. Perhaps I will attempt to optimize the scans I have posted. I must first learn how to use this back tester and understand all of the information it provides on the output.