Predicting randomness

Discussion in 'Trading' started by oddiduro, Nov 3, 2005.

  1. For the sake of argument, let's assume this probability assessment is correct. Thus, the chance of the DOW moving higher is 50/50 each day.

    Will the size of those moves be the same? If such is true, the DOW will be flat over the long-term.

    Direction is only one factor ... magnitude must also be considered.
     
    #241     Nov 7, 2005
  2. Taking in inflation the DOW does move flat over long-term.
     
    #242     Nov 7, 2005
  3. I think that much of this discussion is superfluous.

    I believe that most people who take the market seriously, and earn their living by trading it, are fairly confident that the markets are not uniformly random. What anyone else thinks or believes is only so much cannon fodder.

    I also believe that people who assume a normal distribution when assessing the so-called "probabilities" in respect of market outcomes have not yet encountered an important and painful learning experience. But if they continue with their ill-conceived assumption, I am confident that the lesson will be learned, and learned well, in due course.

    Just my opinion, of course.
     
    #243     Nov 7, 2005
  4. My prediction is that, imo the above prediction about prediction for predictions should be correct. :)
     
    #244     Nov 7, 2005
  5. I have not researched the DOW; I do know the S&P 500 has moved higher over the long-term on an inflation-adjusted basis.
     
    #245     Nov 7, 2005
  6. cnms2

    cnms2

    I've just read Trading Time Frame, link provided by OddTrader in an older thread. Interesting article although some might find it a little difficult to follow. It basically talks about the price signal immersed in the informational noise, and a few ways to better extract it. I recommend this article.

    The apparent randomness of the market is the result of the noise randomness. There is an objective factor we cannot avoid in this perception, but there is also a subjective factor that we can and should reduce.

    If you're trading dailies an immediate lower time frame may have a limited use to pin point entries, but generally intraday time frames will be a confusing noise.

    If you want to reduce the apparent randomness you have to correlate your input data time frame with your trading time frame.

    Input noise happens in two dimensions: time and price. Time frame correlation filters out the time component of the noise. To further reduce the noise effect the price amplitude has to be smoothed out too.
     
    #246     Nov 7, 2005
  7. #247     Nov 7, 2005
  8. MAESTRO

    MAESTRO

    What a boy! My favorite dog Thunder! You always make me laugh. God bless you.
    Cheers
     
    #248     Nov 7, 2005
  9. Pekelo

    Pekelo

    By definition, a Black swan event has 3 properties:

    - large impact
    - surprize effect
    - incomputable probability

    Thus by definition, we can not predict a Black swan event. Nobody said here we could. Two notes on this:

    1. Your original post was about the nonexistence and unpredictibility of trends and not black swan events.

    2. If I remember well, Taleb kept buying OTM puts so in case of a black swan event he would be handsomely paid off for all the wasted money while the event has not occured. Because of the sideways nature of the current markets (and that was easily predictable, at least I did) he would have been better of by shorting indexes.
     
    #249     Nov 7, 2005
  10. Yes, I know, but the discussion evolved in this direction, so I put up an informational article.

    I think trends only exist after they have occured. Sure I may have a bias towards being long or short, but just because the market has gone up for 5 days does not mean that that it will go up a 6th day.

    And there is no way to tell if it will.
     
    #250     Nov 7, 2005