Why do I see "Trends" in Randomly Generated Data?

Discussion in 'Data Sets and Feeds' started by Rahula, Feb 21, 2008.

  1. MarkBrown

    MarkBrown

    there are patterns all over - mb
     
    #51     Feb 22, 2008
  2. MAESTRO

    MAESTRO

    Hey, Thunder, good to see you too! You are the only guy here on ET who makes me laugh. Your sense of humor is ranked number 1 in my books. I will find for you an interesting article that deals with “Zero Intelligence Trading”. This article puts our egocentric point of view on how much really our “intelligent decisions” worth. You will be surprised to find that a large amount of “intelligent” traders who act with their best intentions and use their best wisdom produce an environment that is extremely similar to the behavior of a lightly biased coin. I will look for the article in my bookmarks for you.
     
    #52     Feb 22, 2008
  3. MAESTRO

    MAESTRO

    So are illusions!
     
    #53     Feb 22, 2008
  4. Thanks, I look forward to it. And once again, welcome back.
     
    #54     Feb 22, 2008
  5. ramora

    ramora

    Wishing you the best of health Maestro!

    This may be the article you referenced:

    http://www.santafe.edu/~baes/jdf/papers/science2.pdf

    "The Predictive Power of Zero Intelligence in Financial Markets" from Santa Fe Institute (great center for work on time series studies).

    Good trading.
     
    #55     Feb 22, 2008
  6. MAESTRO

    MAESTRO

    Yes, that is the one! Thank you so much!
     
    #56     Feb 22, 2008
  7. Okay, thanks. Looks like I've got some reading homework for the weekend.
     
    #57     Feb 22, 2008
  8. Are you uggesting that the process that generates market prices is similar to an experiment in probability of tossing an unbiased coin?

    Since S&P retuns over the last 50 years have outperformed investments in risk-free assets, then at least for the last 50 years or so either the market is not a random number generator or you are an empirical fool.

    There are too many parameters that can introduce a positive bias in market returns. One of those parameters is innovation through hard work. Yes, in your simplistic analogy there is no hope making money in the markets but this is not how the economy and markets work. there are events that produce bias and reflected in all sorts of patterns in the market. Whether you like it or not:)

    Bill
     
    #58     Feb 22, 2008
  9. http://wilmott.com/messageview.cfm?catid=3&threadid=58450

    For those interested in some quant views on randomness, I thought this thread was quite good. While the main thread is dedicated to debating on Taleb's views, if you sift through all the pedantic pontificating, you will find some well drawn conclusions that I have advocated on the boards for some time.

    Namely, that markets are not random, they are worse than random (unless you are privileged to inside information and order flow, which is why trading desks always seem to be so successful). Nor is it possible to predict markets. It's similar to predicting the weather; yes, there are millions of deterministic forces that cause a deterministic outcome from a physical perspective, but the complexity and determination of those forces renders predictability mostly useless.

    Many argue that the markets are based on cause and effect human behavior. This is true, unfortunately there are so many complex interactions going on that to one who looks purely at charts, for all intents and purposes, while the activity may be deterministic, the complexity renders it chaotic and non-deterministic for chart readers.

    Based on this view, the majority of erudite traders on that forum rely on statistical betting methods over chart reading (i.e. approaching markets as if they were a random, not deterministic phenomenon).

    Final note: over the long run, being long is in your favor. But, this argument is simply intended towards short term trading chart predictability. If you plot a distribution, although the results are not completely gaussian random (ignoring the outliers), the bell portion shows a slight positive skew, which is why over the long run the bias will be up (as the slightly positive edge will compound upwards over the long run).
     
    #59     Feb 22, 2008
    beginner66 likes this.
  10. Ridiculus and worthless paper. First they assume that players place orders at random and then they conclude that the process places constraints on strategic plays. Do they understand what "rational investor" means?

    It's like one assumes it rains, puts on a raincoat and then compains that places limits on his activities when it does not rain.

    Bill
     
    #60     Feb 22, 2008