Also, ET is one way to stay mentally stimulated (or just plain mental) during those 5 minutes of downtime which occur so often.
Yes, but why is someone of your stature and with your connections not managing wads of money as well? Surely that is where the real money and fun would be for someone as expert as you, no? (As an aside, I reassert that you and NickelScalper are quite the tag team.)
thanks for the link earlier. we must define random as it seems everyone here has a different take on it.
As in our context we are basically dealing with signal theory, or at least trying to mimic it, there can be only one way to go. A simple exponential moving average as commonly applied in TA is already an outgrowth of this theory. Finding 'Trend', whatever it may mean, is an attempt to 'extract' or perhaps 'predict' something on the basis of 'signals' or 'time series' generated by our dear markets. Of course, if you don't know what you want or are unable to explain it to others, including yourself, you obviously run into a problem in trying to fall back on rigorously defined concepts. I think that this is in essence the bewilderment of many when confronted with seemingly endless pro and contra 'Trend' threads.
Nononsense makes an interesting point, but I fail to see how you would arrive at that conclusion based on his post. In my opinion, if trading the markets only came down to a matter of signal processing, then it would be a relatively straightforward engineering problem. However, that does not appear to be the case.
To discuss this, you will have to describe the way you generate your time series. I'll try to do this in a simplified manner. (1) deterministic time series where the value at a particular time can be represented by a function; (2) random or stochastic time series where a value at a particular time cannot be represented by such a function. In fact these random signals are generated through standardized processes well defined through their probabilistic properties. Popular ones are the 'Wiener-Levy Process', 'Poisson Process', 'Brownian Motion'. 'White Noise' can be defined by a (tricky) limiting process on the above but is commonly defined as having a flat power spectral density function yielding the delta function as autocorrelation function. Passing this 'White Noise' random signal through a linear (even non-linear) filter generates a new signal, equally random, called 'colored noise'. The spectral density function is no longer flat and the autocorrelation function is no longer a delta function (spike). This makes this new but still perfectly random signal predictable in the sense that a filter can be found yielding the optimal approximation in some mathematical sense (minimum integral RMS error). Of course, the probabilistic parameters of the random signal have to satisfy the requirements of the theory. Wiener developed his filtering theory. Other more advanced methods exist, eg Kalman filters. These are linear filters. Some work has been done on non-linear filtering but this becomes rapidly involved. When dealing with markets, it is useful to have an idea of these theories. This does not mean that these will lead to any useful result. In fact, most often one will remain far removed from satisfying the requirements for applying these theories.
If a series is random within known parameters, then it is predictable to some extent. It is also nonrandom to that same extent.