Predicting randomness

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

  1. cnms2

    cnms2

    This ties into another discussion / thread about being either aggressive or defensive. Neither extreme seems to be profitable.
     
    #361     Nov 18, 2005
  2. cnms2

    cnms2

    "Pyramiding RSI" is a very simple script taken from Wealth-Lab's documentation. If you ran it on low volatility issues like DELL, MSFT, DIA, SPX it will give returns over the last 1,500 trading days better than the buy and hold approach. It will miserably fail for higher volatility issues, i.e. QQQQ, but in my opinion it is a good example of a successful use of a very simple technical analysis indicator.

    It is very simple to try it: clicking "Pyramiding RSI" you'll get to this script's page on the Wealth-Lab site, where you just type in the symbol you want to check and in a few seconds you'll get the results.

    For those interested this is the code:
    Code:
    { This Trading System buys whenever RSI crosses above 30, 
      and closes all open positions when it crosses below 70. }
    var BAR, P: integer;
    for Bar := 15 to BarCount() - 1 do
    begin
      if RSI( Bar, #Close, 14 ) < 70 then
        if RSI( Bar - 1, #Close, 14 ) >= 70 then
        begin
          for p := 0 to PositionCount() - 1 do
            if PositionActive( p ) then
              SellAtMarket( Bar + 1, p, '');
        end;
      if RSI( Bar, #Close, 14 ) > 30 then
        if RSI( Bar - 1, #Close, 14 ) <= 30 then
          BuyAtMarket( Bar + 1, '' );
    end;
    
     
    #362     Nov 18, 2005
  3. #363     Nov 19, 2005
  4. surfer

    I am not going to defend TA here.I have seen yours and oddi's posts about Gann here and also at TS forums, so I think its useless to argue about value of TA

    My problem with the article had to do with the following points.

    Indicators? Value is ascribed?
    Whoever told him that you regular indicators are suppose to have value.
    Sounds like this guy
    http://elitetrader.com/vb/showthread.php?s=&threadid=58084
    Someone ought to tell him that reading Magee and Murphy is not the end of TA.They are just primers.Perhaps his protege(?) Crabel might give a few pointers to show how actual testing is done.

    VN and the team apply the scientific method in an exacting fashion to the market, testing massive amounts of data, across lengthy time frames, using the latest and best programs--- like it or not, the results speak for themselves.[/QUOTE]

    All this to test what? Canned indicators? Head and Shoulders?

    So there Surfer

    hermit
     
    #364     Nov 19, 2005


  5. All this to test what? Canned indicators? Head and Shoulders?

    So there Surfer

    hermit
    [/QUOTE]

    Okay, what I think VN is saying is that the usual suspects as far as indicators go...MACD, RSI, Stochastics, Head and Shoulders, Double bottoms, Triple bottoms, Average True Range, Doji, Dark Cloud Cover, etc. ,all have thier issues with reliability.

    I will say again that I think indicators (including the squaring of price and time) should be used only to trade within a framework of rules.

    These rules are mainly in place to limit risk, not to predict profit.

    It is not possible to predict direction with any statistical reliability.

    Those people who are saying that the market is non-random are mistaking their system of limiting risk for predicting the market, and they are two different animals.

    My trading education is fairly well documented on elite, and on the TS forums. I have used indicators successfully, but I have never felt comfortable enough to fully automate any system. The market is too complex.

    The complexity is due to it's chaotic nature.

    That is my assertion.

    This thread continues because I believe that many very experienced traders are aware of this "apparent randomness"

    If someone could predict within a margin of error of 0.1% the closing price of any market entity, for the span of one week, it would end the randomness talk for good from me.

    You need not reveal your methodology, just post the results by no later than 9:29 am on the week of the test.

    That would shut me up:)

    Best Regards
    Oddi
     
    #365     Nov 19, 2005
  6. Actually, Pekelo, that post is a good one.

    When you think you can predict the market, you are in effect trying to predict the future, which, if effect, makes you think you are a psychic.

    I don't believe in psychics, nor do I believe that the future is predictable, not in life, and not in the markets:)
     
    #366     Nov 19, 2005
  7. Very nice contribution!!!
     
    #367     Nov 19, 2005
  8. Thanks...

    Be sure to tell Vic that we are talking about him (again):D

    And could you ask him, what does he mean when he says, "good stock" and "good price"?
     
    #368     Nov 19, 2005
  9. I think I should post a little history.

    I started trading in 1999. My first stock was RMBS. For those who are newer, RMBS was then what GOOG is now.

    Nothing new under the sun.

    Then the bubble burst, I had to learn how to be a bear. I had to adjust, and readjust indicators, some canned, some custom written.

    Then the bubble deflated, and volatility with it, more adjusting of indicators. Some indicators work in some markets, some in others.

    Indicators assume that the market is a static entity.

    It is in fact a dynamic entity, and to rapid for a static indicator to maintain for any length of time. This is why people wind up using a suite of indicators. They are tring to define a circle with straight lines.

    I come to the traders today to bring this to the table, and offer a friendly challenge, which I stated in the earlier post.

    Thanks for all who have read and this very Socratic debate.
     
    #369     Nov 19, 2005
  10. What happened to randomness? I want a refund.

    :)

    (there that's better)
     
    #370     Nov 19, 2005