To: gbr who wrote (7710 ) 1/10/1998 10:33:00 PM From: ftth Read Replies (2) | Respond to
Hi Gary, I've played with Insync some. I don't know if you've read the article from which it is based (worthwhile if you haven't), but the basic idea is to generate a composite signal of a group of indicators, determine how they interact (easier said than done), define a "setup" region for each, and assign a weight to each one as they enter their setup region. I'm greatly simplifying a 7 page article, but I like the general concept and the method. I just don't particularly like the components they chose. For me to use an indicator, I have to completely understand why it does what it does, and know how to read false signals. I could never get to that point with this indicator because it has 10 indicators blended together, which makes it very difficult to develop an intuitive feel for (I found myself constantly dissecting it to try and understand why it looked the way it did). It probably could be done over a long period of use, but since I don't really like the components, I didn't bother. They had reasonably good results with the settings you're using (I assume you downloaded it from the Equis site), trading the DOW with it. However, I'm a firm believer that fixed indicator settings don't perform universally well (or poorly) across all stocks. If you're using it to trade the DOW, or any basket of stocks that correlates well to the DOW, it will probably work relatively well. But, all the components of INSYNC have time/amplitude/phase parameters that are optimized for DOW-like cycles, trends, and overall movements. My view (FWIW) is that a TA system is a multistage filter, and like any filter, it needs to be designed around the specific system which it will be applied to, which means you have to know the characteristics of the types of stocks you will be applying it to, and design the filter from there. Otherwise, you're just randomly designing filters and looking for a system to apply it to which will produce good results. I could give some geek examples of highly complex adaptive filters which work miracles on the signals they are intended for (because they were designed to work with that class of signals), but fail miserably on other classes of signals--even in the same frequency range (but I won't bore 'ya--you see my point). Also, I've just got this "idiosyncracy" about using canned indicators set the way everybody else uses them. That's my 1 cents worth (adjusted for deflation). Please comment if you think I'm full of it.
