Combining multiple systems

Discussion in 'Strategy Building' started by Arrow, Sep 4, 2008.

  1. I bumped into this interactive chart ( NewYork times web link below) showing how this Bear market compares with previous bear markets in terms of drop and length.
    I am planning a small research to build a programs ( chart and data ) to how the GOLD and Crude Oil traded in all those bear market periods compared to Dow Index .

    How This Bear Market Compares
    http://www.nytimes.com/interactive/2008/10/11/business/20081011_BEAR_MARKETS.html


    I believe this exercise will give good clues how GOLD, Crude behaves in those periods relative to Dow. I think by taking average of all those bear market periods for GOLD and Crude we can see pattern which can be good source for trading ideas on GOLD and Crude Oil . I think that will help to predict till some extent the direction of GOLD, Crude for next year or two.

    Let me know if anyone is interested in this small project , so that we can pool the resources and share the results and insights.

    Cheers
    Reddy
     
    #141     Oct 20, 2008
  2. You can combine multiple systems within TradersStudio using a "Trade Plan"
     
    #142     Nov 7, 2008
  3. MGJ

    MGJ

    As you could in 1995 within Trading Recipes using a "SIMULTEST". Other testing engines that will run multiple systems simultaneously include Stratasearch, Trading Blox, PowerST, Ultra-10, and Mechanica.
     
    #143     Nov 7, 2008
  4. Some great info here, thanks to the contributors.

    That ratchet effect of Parrondo's Paradox is a nice example of how stop losses that are too tight can affect your equity in what would otherwise be winning trades.

    Unless I read the article incorrectly...

    I currently use Amibroker for designing and implementing automated trading strategies, however now that I am combining multiple strategies, are there any recommendations for alternative software? Ie Stratasearch?
     
    #144     Nov 11, 2008
  5. An alternative for amibroker is to take the resulting equity curve and put it in excel, index it (so you can compare it to other equity curves), and then add it as a chart back into amibroker.

    I use mainly SAS for all my analysis. Excel is a good start because there are some good add-ons for statistical manipulation on the internet or you can do your own programming.

    Think of what you are trying to do before you jump on some tool set, then ask what you are looking for. I see so many people using sophisticated statistical tool sets: neural nets, evolutionary algorithms, etc., when they have no idea how they work or the assumptions behind them.
     
    #145     Nov 11, 2008
  6. What the Monty Hall example really shows is that people do not understand the conditional probability. If they did, the Monty Hall example is like asking someone to sum 1+1 (which is 2)! The example is literally equivalent to summing 1+1, since when a door is opened the probs on the door non-opened and not chosen doubles (1+1=2)/3. The latter is the conditional prob of the last door.
     
    #146     Nov 27, 2008
  7. 0.75
     
    #147     Nov 27, 2008
  8. mind

    mind

    1. A dice is thrown showing heads and then another
    time showing heads.

    2. A dice is thrown showing heads and then another
    time showing tails.

    3. A dice is thrown showing tails and then another time
    showing tails.

    4. A dice is thrown showing tails and then another time
    showing heads.


    given the dice showed one head, what is the probability
    of another head?
     
    #148     Nov 28, 2008
  9. (Most important point is 3.)

    1. A. to your question as stated is 0.5.

    2. That guy has other assumptions in him mind. Answer depends on understanding of assumptions, which I am not sure I understand as he has them in his mind.

    3. In fact, trading as would be done by a monkey (I mean real monkey not a human insulted as a monkey) using stops would be:

    Lose $1 two thirds of the time, to make at least $1, one third of the time.

    Could someone provide the proof/answer/disproof to question 3.?

    I asked the question on RFT's financialtraders blog, and not a single person answered it, despite the blog receiving between 800 and 900 page views per day!
     
    #149     Nov 29, 2008
  10. mind

    mind

    the answer of 0.5 is only obvious, because i called the
    object at hand "dice", everything else is the same as
    in the stock example.

    there is no other assumption and no one her provided
    one throughout the discussion if i recall correctly.

    there are only three outcomes after two repetitions:

    -2 with 25% chance
    0 with 50% chance
    +2 with 25% chance

    unfortunately this is all very simple. the difference to
    the paradoxa that were mentioned is that the stock
    price is identical repetition with the same odds for the
    first and the second round, while in the monty hall for
    example it is not. the only thing that would change the
    odds for stocks would be an additional assumption, like
    "the stock move of today depends on yesterday's move",
    but nobody mentioned an addition like that and it is
    relatively needless, since it is selfexplaining and has
    nothing to do with a paradox.

    IMHO the whole discussion is pointless and i (still) do
    not understand why we are discussing such obvious
    stuff ... actually i know. i read reasonable stuff from
    MAESTRO on other topics, otherwise i would not have
    written anything here.
     
    #150     Dec 1, 2008