I don't discount what either of you say. However, for kicks I've decided to spend a little time on it which may serve to validate what you say or otherwise.
Good for you. Would be interested to see your feedback here in future. Don't want to be preachy, but if I were you I would aim for a dispassionate analysis of your results. It's very easy to be convinced something is working just because you want it to.
Technical analysis to the best of my knowledge is statistics.
All the graphs, charts etc are statistics. There is maths behind all the squiggly lines.
The CSO, government etc all use statistics to try and predict the future based on what has happened in the past and present.
Sometimes it works wonderfully and sometimes it doesn't.
As Brendan said, it's not the same thing. Some signals contain useful information that allows the future to be predicted from the past. Some don't, because they are random. And some -- the so-called
chaotic ones -- are even worse than either of those. They may show long term trends (e.g. the stock market always goes up in the long run), they may oscillate between bistable or multistable states, they may gravitate toward or orbit around
attractor values, but will always be prone to unpredictably flipping into other metastable states.
And they don't even have to be random, like the Gaussian noise introduced by the Box-Muller transform in Brendan's example. It's possible to use perfectly deterministic mathematical equations (of the appropriate type, even if it's as simple as the equations of motion for the double-pendulum) and
still get chaotic results. The weather is the canonical example of a chaotic system, one which is completely unpredictable beyond the short term even though it stays within long term bounds of climate oscillations.
The stock market is almost certainly another example, since extreme sensitivity to initial inputs is one of the hallmarks of chaotic systems, perhaps along with another property called
self-organised criticality. It is in its very nature to erase predictively useful information about past movements, because the market always reacts to discount them, even if the inputs are coming from traders using technical analysis!