A refreshing look at technical analysisConventional technical analysis is based on the combination of two methods for predicting price movements: one method attempts to "clear" them of noise in order to track their direction, and the second method attempts to predict changes by detecting obstacles in its path and evaluating the momentum of a current direction. In addition to these two methods, it is common to rely on patterns that are signs of things to come. These are mainly patterns based on Japanese candlestick arrangements that can indicate a trend reversal, such as a head and shoulders pattern, etc.
The problem, which is familiar to everyone, is that with any security or currency pair, you can also see shorter periods of decline in a period of rise and vice versa. That is, the price chart has fractal properties (fractal is the name coined by mathematician Benoit B. Mandelbrot in 1975 to describe repeating or similar geometric shapes).
Classical or Euclidean geometry fits perfectly into the world that man has created. However, it is less suitable for the structures found in nature. Clouds are not perfect spheres, mountains are not symmetrical cones, and lightning does not travel in a straight line. Nature is not smooth but rough, and until recently it was not possible to measure how much. Mandelbrot developed a mathematical representation of complex patterns that repeat at any scale, and thanks to the invention of the computer, he proved that fractal geometry can represent patterns even under conditions of irregularity in the natural world. In his book: "The (Mis)Behaviour of Markets: A Fractal View of Risk, Ruin and Reward" he shows that fractal geometry can also represent market movements.
The idea can be understood without mathematical knowledge by looking at the course of a river. A river does not flow in a straight line, but in a channel where the resistance to water flow is lowest. The actual length of the river divided by the air distance is defined as the coefficient of curvature, and its average value for rivers in the world is 1.94. The coefficient of curvature is also a special case of the fractal dimension. The curvature coefficient 1 of a straight line is actually one dimension, while the curvature coefficient close to 2 is actually almost two dimensions. In the general case of two- and three-dimensional shapes, the less smooth the contour of a shape, the greater its roughness - its fractal dimension is greater.
Mandelbrot claims that the "noise" - the fluctuations of the price around the general trend - is not random ("white" noise) and therefore cannot be suppressed by a moving average that smooths it out while ignoring the fractal properties.
Many attempts have been made to find the regularity of market movements using fractal properties. What I do is to form multiple channels:
Within the "game board" whose boundaries are defined, I sketch the longest parallel channel, and within that channel I sketch an intermediate parallel channel, and on top of that I sketch the optimal parallel channel for trading in a time frame that provides the best balance between too early and too late, and between too painful fluctuation intervals and insufficient trend lengths.
