Let me elaborate my point. I know that the trends exist. They can emerge and subsist due to to many different reasons like economy growth, herding behavior, positive feedback, etc. The usage of trend is absolutely necessary for a long-term strategies. But I do not understand how the people try to predict the trends, using the statistical (hystorical) data. IMHO, one can not do so for the timeframes longer than one hour, otherwise the (two-point or n-poit) autocorrelation functions would not be zero for such timeframes.
To summarize my point:
1) The trends do exist, and one can use them.
2) After (15min.-1 hour) the trends can not be predicted by using the statistical data, mechanical trading systems, plots, etc. This is a matter of principle — the prediction implies nonzero autocorrelations.
3) One can try to predict the trends by using the fundamental data, macroeconomic analysis, behaviour patterns of market participants, etc.
I ask the trendfollowing people to let me know if I am wrong at point 2.
СТА-директории
What is it and where one can have a look at it? Thanks
