The Geek Forum

[Capitalist]

The project on technical analysis conditions is finished and I have moved it into the Lab Archives:

http://www.market-geeks.com/labarchives.php

As a result of this project, I now have a nice little research tool that I'm calling the Conditional Analysis Test Scanner (or CAT Scan for short...cute huh? LOL).

Anyway, this is a tool on MS Excel which, given a set of TA conditions such as "RSI > 80 and rising, 20-bar moving average below 50-bar moving average" or whatever, tests that set of conditions to see if there is any statistically significant change in price behavior during periods when those conditions are met. The CAT Scanner is set up to run this standard statistical hypothesis test 21 times using 3 time-frames for each of 7 FOREX currency pairs.

It takes anywhere from 30 seconds to a few minutes for me to code in the TA conditions, depending on how complex they are. Really complex conditions are not a good idea anyway for a lot of reasons. At any rate, it then takes me about a minute to run the 21 tests using macros I have set up. This provides a set of 12 "scoring values" which give an indication of how price behaves when the given TA conditions are met as opposed to when they are not. So I can test several sets of conditions in under an hour. There are concrete examples of this entire process in the project notes.

One benefit of this tool is that it's great for generating ideas for trading setups and other strategy ideas that wouldn't have otherwise occurred to me. For example, I was running routine tests of how price behaves when price crosses a moving average, and I ended up with a method for knowing when I could use stops that are much tighter than usual. This example and others are all documented in the notes.

A theory that I talk about in one of the Geek Academy classes has to do with how the effectiveness of different TA indicators might change continually over time:

http://www.market-geeks.com/academy/research/frontiers/frontiers1.php#paradigm

I think the CAT scanner (as primitive as it is) may help enable us to watch those changes and keep on top of them. I could periodically run scans of a set of popular TA conditions and see which ones are still working well and which ones don't seem to be very valuable currently. I think I'll set something like that up in the Trading Operations Center, since there's no tools or anything there right now. I'll put the next research project off for now while I work on that.

Out for now...

Cap