Can you post the statistics of each time series? Like mean, standard deviation, skewness, kurtosis? An interesting experiment is to try different bars on pi then redo the statistical measure. The statistics should be nearly identical if large number of values are used.
I don't have the data at the moment, I've posted some raw bar files online if you're interested: https://www.sendspace.com/file/oikydm Edit: Values here are shifted so that 0 is the minimum. NN doesn't like negative numbers
The question would not be which one is PI--Rather it would be a decision based upon which item is a step up from the descendent one. In mirror like quality and qantity it would be verbose and futuristic in nature. The bars et al that are ensconced within these margins are derived from noisy attributes of the participants. Moods grow from this area and serve notice as to the direction of desirable pomposity. Once the decision has been taken a virtuoso with strong virtue as well as incomparable technique can and will lay claim to the bounty foretold in the many lines of coded vociferous suspended animation. The spoils are to be kept in a one dip box for future remunerative possibilities.
From my perspective a number of the "breakouts" from "consolidation" in the PI chart are non-typical while the other chart seems more true to the real world. Also the trends/S&R seem to be more akin to typical market action in the real chart.
Hey Tavurth An interesting thread you started here I like it, very abstract. I would like to think that chart B showed signs that it was reaching for liquidity a certain points but then again am I being fooled by randomness. I would certainly like to participate in an experiment where the trader had to try and ascertain which charts were produced by a random walk and which were actual market data over a larger sample, the results would be very interesting. Was there a particular hypothesis you wanted to pose with this thread? As a topic i am very interested in time series analysis and how it can be applied to baskets of instruments to create heavily cointegrated synthetics. Thanks
A blind selection would be interesting. PI looks so tradable, I think it would also be interesting to see if the market is more efficient. I wanted to suggest how easy it is to be fooled by our own minds. To me the charts of PI look far cleaner than the marketplace. The trends are longer, and with a greater run. And this number is supposed to be random! A much better use of my time perhaps, than this posting of yours.