Hi Dennis, > These two are just a few of the possibilities. They are based on > combining two lines that have a probability of being a support or > resistance line to get a time target where both could be true at the > same time, thus increasing the probability of a reversal.
I don't know if that is the explanation but the projected turning points are right on the pivot bar with uncanny accuracy. There are several problems. 1. The accuracy with which the trend lines are placed is critical, even if the position is off a small amount, it can throw the projection off especially if the lines cross a large distance from their starting points. AB has a snap tool that lets you snap drawn studies to prices but I could not figure out how to make it work for trend lines. 2. While most of the projections are right on turning point bars, sometimes the turn last for only 1 bar and other times it indicates a major change in trend. I have drawn several hundred square and angle CIT's (the dual angle ones occur much less frequently so I have limited experience with them) by hand and it works well enough that I want to study it in more detail. It seems to work on every trading instrument, market fractal and chart type (time, tick, volume, and range) I have tried it on so it must be tapping into some underlying market symmetry or cycle. > I have thought about how to do various study like indicators in AFL. > I do not have the AFL for you, but If I were to try to do these two, I > would approach it with using a ZigZag indicator to find the > significant peaks to draw from. Then using a simple solution of two > linear equations to find the crossover points. > > The question is how many of these do you want to draw? For researching this technique, I want to draw them all but for trading purposes, I only need the ones in the future. > The latter would require a lot more smarts in > the AFL to decide how to use the information. It should not be too hard to analyze the data, I need to calculate all the CIT points; catalog all the pivot point, even the very smallest; and correlate them. Then see if there is some (hopefully obvious) way of filtering out most of the less important ones. > However, you could also just use the support and resistance lines from > different methods and add together all the probabilities of a reversal > of each. That would automatically increase the probability as the > price approached two or more lines that are converging. That does not seem to work. Frequently more that 1 set of reversal lines cross at the same CIT point, unfortunately, that does not make the point any more significant, i.e., it is no more likely to be a major pivot than any other point. > Understanding the underlying principles can lead to simpler solutions. Agreed, that is one reason I want to study it. Bill
