Elliottware In a nutshell

The idea of Elliottware is to use the human pattern recognition in order to precise where to place the optimization windows and until when the trading window would be OK.


The idea is about similarity. The Elliott Wave theory is very precise about identification of the market stages. Here and by now I would assume that you have an understanding what it is.


For example we have a trend pattern. The market behaves in a trendy way. We can isolate this period and use it for training window. It would be a non sense (well in my perspective) to use the range that precedes the break - out to train the expert (or the neural net what ever). Well we may not have sufficiant data but the data we would have woul lead to a robust solution.


There is a problem more we have data with precise the same market conditions we expact that the change in the market conditions is imminent.


A lot of data for optimization - less robust (the market is expected to change)


Quite a few optimization data (e.g. we just had a break - out) - quite robust (we expect that the new market conditions are likely to continue).


The ideas for the Lyapunov exponent should be added here. There is a constant loss of information as the market is chaotic, that should be taken into account.


There are other ways to make an optimization for example some want to make a more global training that takes into account both trending and ranging patterns. This can be done easily using the Elliott Wave theory. We can on the past data isolate quite sucessuflly 5 impulse followed by correction. LOL


1. The firts training range will be the immediate visible count wave.

2. The second training range would be the bigger wave.

3. The third training range would be the previous wave.

4. The last would be the biggest wave.