Williams %R

Discussion in 'Technical Analysis' started by Pekelo, Feb 28, 2006.


  1. As do the rest of the 95% of "traders" who eventually fail.

    st
     
    #31     Mar 1, 2006
  2. KS96

    KS96


    The problem is that
    when you define the profile
    of the noise you want to filter out, it changes.
    Slippery business, isn't it?

    The neural net inside your head is capable of adapting faster
    than any other adaptive filter you may want to use.
    (Let's also stress here that the indicators/filters we started
    talking about in this thread aren't even adaptive...
    Pick your parameters, and live with them.)
     
    #32     Mar 1, 2006

  3. Looking at the failure rate of 'traders', it becomes clear that approximately 95% never "understand" that principle. I'm simply amazed at how many people follow the losers before them by thinking they have somehow found the secret to an indicator.

    st
     
    #33     Mar 1, 2006
  4. Pekelo

    Pekelo

    So far 1 signal, 20 points gain with the YM. Nice....
     
    #34     Mar 1, 2006
  5. just21

    just21

    So how does price tell anything? If it is expensive it is overbought? How price reacts to news? How price reacts to support/resistance? How price reacts to yesterdays high/low/close?
     
    #35     Mar 1, 2006
  6. KS96

    KS96


    Good for you! You buy beers for us tonight.
    We will buy for you another day :D

    Let's see... Stoch(14,3,5) gave a great sell yesterday
    at 13:55CET in FESX: 55 full points in a few hours until stochs
    crossed above 50 again! Trading is easy!
     
    #36     Mar 1, 2006

  7. Even a blind squirrel finds the occasional acorn. :D


    st
     
    #37     Mar 1, 2006
  8. KS96

    KS96

    There is no such thing as "overbought".
    Price-derived indicators cannot tell you
    how many traders are still queued to buy.
     
    #38     Mar 1, 2006
  9. ST, understanding Kalman filtering gives one a unique understanding of the mathematical nature of indicators.

    IF:

    the signal in price could be described by a physical model,

    the noise in price could be described statistically,

    and the noise in the measurement of price could be described statistically,

    THEN one could create a multi-dimensional filter which would estimate the future value of the signal component of price to the greatest theoretically possible accuracy.

    A corollary to the theory of Kalman filtering is that EVERY mathematical operation (like an ma) on a time series (like price) is an ideal Kalman filter for some potentially infinite number of combinations of the above three conditions. The sad part of indicator invention is that the people who do this don't realize that. Every indicator is a perfect solution to an unknown problem.
     
    #39     Mar 1, 2006


  10. Already discussed at length with concensus that it provides no tradable edge.

    http://www.elitetrader.com/vb/showthread.php?s=&threadid=49502&perpage=6&pagenumber=1


    st
     
    #40     Mar 1, 2006