Example Trading System

Discussion in 'Strategy Building' started by sidinuk, Jan 17, 2005.

  1. I exhaustively tested "opening range breakout" for NQ several times in the past two years and it doesn't [edited] work.
     
    #31     Jan 22, 2005
  2. It did last week.:D
     
    #32     Jan 22, 2005
  3. Hi Sulong. The fact that I love you and Michael B. hates you should tell you something. I should have said that it didn't work in any of the 90 day backtests I did earlier. Of course it works, just like other old saws like fading opening gaps work. But not with positive expectation.
     
    #33     Jan 22, 2005
  4. I was just giving you some trouble. which is why I said "last week".

    I've tested the NQ orb to death by hand bar by bar.
    As long as you only enter on a expansion day, and the opening range contains price for a minimum of one complete or bar that follows the or, it can be profitable.
    Other than that, it's a waste of time.
    Frankly, too much work for my style, but I can't help but watch it.
     
    #34     Jan 22, 2005
  5. You blew me away with "the opening range contains price for a minimum of one complete or bar that follows the or". What bar period are you charting in to achieve that?

    Also, the thing that bothers me is that your definition of this "system" requires three "strong" rules, as Hypo used to say, and the testing of it is therefore highly suspect.

    My god! An intelligent discussion! What is ET coming to?
     
    #35     Jan 22, 2005
  6. With all due respect, the system not only uses an MA, it uses as a starting point a generic technique itself derived from extensive analysis of historical data. The article cannot legitametely claim to have done any out of sample testing.

    The Thursday filter does, in fact, turn a "losing system" into a winning one. Without that filter, the gain is small enough that mediocre fills on the MOC exit condition, a high probability event as every active trader knows, is enough to eliminate the profit of the simulated system. The "one tick" assumption is possibly acceptable on the entry, at least for initial testing purposes, but is completely indefensible on the exits.




     
    #36     Jan 22, 2005
  7. Sorry, I had some errands to run.

    I used 15m bars in my study, but I suspect 17m bars would be a better time interval, I just don't have the proper soft ware to make a 17m bar. I tried to count 1m bars but got lost.

    If you get a bar following the "or" bar that is not higher than the high of the "or" bar, and not lower than the low, then you have something to work with.
    If you get at least 2 following bars that are contained with in the "or", it's almost a slam dunk.
    As far as the rules, the "expansion day" was a wise crack, seeing as how at the time we won't know.

    I done my testing with an eye toward exiting at the completion of the ADR, or 9:00 (12:00 eastern time) or a stop out at the other end of the orb bar. whichever came first, unless there was 2 or more contained bars, then hold for the whole enchilada.
     
    #37     Jan 22, 2005
  8. Thanks for the explanation. So in my one minute context you are looking for a long consolidation which breaks out?

    That makes for five "strong" rules! That qualifies for "discretionary"!

    By my reckoning there are only three of us trading NQ here. Trade well!
     
    #38     Jan 22, 2005
  9. Yes. To my way of thinking, a significant BO needs to have a significant consolidation process to BO from.
     
    #39     Jan 22, 2005
  10. Is that why you're saying that you're better off after a couple of inside bars than on the BO on the very next bar after the "or"?

    Is the difference significant?

    Do you have any figures for this?

    Sounds interesting, anyway ...
     
    #40     Jan 22, 2005