Let’s admit, the fundamental factors advancing a market position, long enough remain invariable (or vary a little and slowly). Then dynamics of quotations will be advanced by irregular effect of not too essential circumstances. It is possible to make a hypothesis (which by all means requires check) that these circumstances have casual character.
The elementary model of observable movement of the quotation will represent so-called «a one-dimensional Brownian motion». Thus it is supposed that quotation change for any period of time is a random variable with the normal law of distribution, and the population mean of this change is equal to zero (that is the increase is so probable, as well as reduction), and the dispersion (disorder of parameter) is proportional to a root square of length of a considered time interval.
Thus the future movement of the quotation is advanced by current significances of parameter and does not depend on background. The last property takes place in physical processes more often. It is possible to assume that for dynamics of quotations it is executed not too well, at least because traders in the decisions are guided by background. As the model considering this circumstance, it is possible to try to use «fractal Brownian motion».
It is advanced as a usual Brownian motion with unique difference: the dispersion of change of the quotation in time t is proportional not to a root square from time, i.e. t0.5, and is proportional t H, where H – any number concluded between zero and unit.
It appears that at significances H which is distinct from one second, the increment of the quotation for two next intervals of time already are correlated, and if H there is more than half correlation is positive and if less – it is negative. Thus, if significance H with certain reliability (which that more than is stronger H differs from one second) it is possible to predict the future change of quotations on the basis of last changes is known.
There is a question how to find significance H? It appears that H it is connected with fractal dimension of the schedule of quotations d the formula d=2-H. Dimension of the schedule can be evaluated, as we described above (covering with its squares). Detailed enough exposition of this model can be found in R. Cronover’s book. In the same place necessary computer algorithms are resulted. It does not mean that the offered model will make for you profit, and, accordingly, it is not necessary at once to rush to program. The model is based on a number of the hypotheses most essential of which are formulated above. First of all it is necessary to find out, whether these hypotheses in your case are executed.
As blanket recommendations it is possible to be guided by following reasons. Long-term tendencies are advanced by the factors lying out of a financial market. On the scale they are most significant, therefore first of all it is necessary to consider them, and it is area of the fundamental analysis. Faster processes arising at change of long-term tendencies, most likely, are described by smooth models, like models of classical technical analysis. And only if you have desire and possibility to consider even more thinly effects, it is necessary to address to fractal models. Thus it is necessary to mean that corresponding models it is essential not lined, therefore, that them correctly to identify (that is to advance significances of all parameters of model), it is required to consider significant information contents at the analysis.
People who took the decision to participate in forex trading should start from learning the basics of this market to make sure you do not have problems with this industry.