Market Profilers - help please (CQG)

Discussion in 'Technical Analysis' started by mightyginger, Jan 11, 2004.

  1. I have. My understanding is that you are supposed to build it from bars (30 minutes in what I read). Just calculate the high and low for each bar and plot all of the prices in between. Then you won't have to worry about missed prices.
     
    #11     Jan 12, 2004
  2. Yeah, that would make sense actually doing it that way. I just construct them from last trade and build it that way. I suppose that way would resolve the problem.

    Thanks.
     
    #12     Jan 13, 2004
  3. Calypso

    Calypso

    I have found excel does not capture all the ticks in a fast moving mkt.
     
    #13     Jan 13, 2004
  4. Calypso I don't have that problem - i think it's good enough. it's just to do with the speed of the vba and how often it's run. i think mine would get away with it but i can see now that the best way is to use the bar approach and just try to work out the max and min of the current time period. i'm just interested in where that approach came from - I've only got the dalton book and it doesn't describe building it that way. It's very vague but you get the general idea. I was perhaps just being too strict by only including that which traded - now i'm refining my code to build it the new way.

    When i've got a chance i'll try and get it running properly.



    Neil.
     
    #14     Jan 13, 2004
  5. I attended one of Peter Steidlmayer's Market Logic seminars back in the 80s. They say he's changed the method somewhat since then but not in any substantial way. He was building his TPO distributions from 30-minute bars. I'm sure he didn't care a hoot whether any particular 1-tick TPO had actually traded. Good luck with your spreadsheet! Once you get the hang of it, you get a new and different angle on price action. You also see empirical proof that efficient markets and independence are not to be found in the real world. His whole idea is that the markets are always striving to make price and value converge, and never quite succeeding.
     
    #15     Jan 13, 2004