in Search

Re: Announcement

Last post 02-18-2005, 19:26 by didier. 14 replies.
Sort Posts: Previous Next
  •  01-28-2005, 17:02 383

    Re: Announcement

    Thank you Didier, are you French by any chance?
    I'm actually from Marseille .... 8)
    My Site
  •  02-03-2005, 14:55 395 in reply to 383

    Re: Announcement

    Oui, je suis français. Je vis et travaille à Paris. Mais je me fiche de football si tu vois ce que je veux dire...
    Merci pour la qualité de tes réponses. Tu sembles maîtriser Metastock en profondeur. Je continuerai, sauf avis contraire de ta part, à te poser mes questions en anglais pour que les anglophones puissent profiter de tes astuces...
    Si ce n'est pas trop indiscret, travailes-tu uniquement sur l'évolution de Metastock ou développes-tu également des systèmes de trading ?

    A bientôt !
  •  02-03-2005, 15:46 396 in reply to 383

    Re: Announcement

    On peut continuer en Anglais, pas de probleme. Ca fait plaisir de voir un compatriote de temps en temps et je te souhaite la bienvenue sur le forum :D

    On my personal time I do work on trading systems and tried to develop a few of my own. Not that I'm trying to sell plug ins for $300 a piece :roll:

    I find technical analysis very interresting, It just seem to me that people keep using the same technique and changing a little something here and there and then just stick their names on it :)

    If you have any ideas that needs coding do not hesitate to ask me, I'm always interrested ...
    My Site
  •  02-03-2005, 19:20 399 in reply to 383

    Re: Announcement

    In fact, I'm a newbie on trading systems. As you seem much more advances than I, I wanted to know if you tried to code something based on some basic principle :
    the trend is bullish if higher highs and higher lows.
    the trend is bearish if lower highs and lower lows
    But I wish I could avoid to predefine the time span.

    So I was considering something on peaks and throughs based on three "active" barcharts. The problem is then to be able to define "active" barcharts backwards. It is much more straightforward by just considering the first bar of the chart as an active bar and then to see is the new one adds some information (i.e. lower low or higher high). If so, we have the two active bars that are necessary to find some peak and through progression :
    - the third bar will be considered as active if it goes above or below the Max (Active Bar 1, Active Bar 2) or Min Active Bar 1, Active Bar 2)
    - peak : if Close Active Bar 3 < Open Active Bar 1
    - through : if Close Active Bar 3 > Open Active Bar 1
    Doing the work backwards is more difficult, because of inside bars.
    I'm not sure that my explanation and my english are clear enough. So I quickly drew the "system" on a chart. There is, a least, one error (the second sell signal could have been avoided, since the close is above the open of on October 1st), but it should clarify what I would like to do.

    Thanks for your comments !
  •  02-04-2005, 16:25 405 in reply to 383

    Re: Announcement

    I actully did something similar looking for basic reversal patterns.
    Unfortunately when testing it, it was not as good as I hoped.

    I will take a look at what you gave me but it might take me some time.
    I hope you will be patient with me :)
    My Site
  •  02-07-2005, 17:33 419 in reply to 383

    Re: Announcement

    No problem Patrick !
    I appreciate your help. In fact, I don't want to use such "swing points" alone. I would like to couple such basic trend following with indicators.
  •  02-08-2005, 21:30 428 in reply to 383

    Re: Announcement

    Ok here is a chart of what I got so far.
    Take a look at it and tell me if that seems right ...
    If it looks correct than let me know what The other conditions are
    Thanks
    My Site
  •  02-09-2005, 8:39 439 in reply to 383

    problem to see your attachment

    Hello Patrick,

    I cannot open your ".bmp". Can you please try again or send me the graph on my private email
    Thanks a lot !
  •  02-09-2005, 16:35 442 in reply to 383

    Re: problem to see your attachment

    Let me know if you can see this one ...
    My Site
  •  02-09-2005, 17:27 443 in reply to 383

    Re: problem to see your attachment

    Hi Patrick,

    I can see your chart now. Unfortunately, it does not fit to my vision of active barcharts. It begins well but after the very wide bar, the trouble begins. I tried to label on several "decisive" bars why I don't agree.
    I see that your bands do not reflect the Min and Max of the two previous active barcharts. But you seem to be quite close...
    Shall I try to explain again the kind of swing points I wish to modelize or was it clear the first time ?

    Good evening !
  •  02-11-2005, 4:56 460 in reply to 383

    Re: problem to see your attachment

    I think I understand what I missed.
    Let me take a look at my formula tomorrow and see if I can adjust it.
    My Site
  •  02-15-2005, 16:55 484 in reply to 383

    Re: problem to see your attachment

    I did not forget about this ... I'm just having a hard time doing it :oops:

    I don't give up easily though.
    I will let you know if I find the solution.
    My Site
  •  02-16-2005, 7:43 489 in reply to 383

    Re: problem to see your attachment

    Thanks Patrick for your efforts. Keep trying !
  •  02-18-2005, 17:26 500 in reply to 383

    Re: problem to see your attachment

    I don't know if this is correct, but I figured I would post this :

    CD:=Inside();
    CD1:=If(CD,0,
    If(CD=0 AND
    (H>Max(ValueWhen(2,CD=0,H),ValueWhen(3,CD=0,H))
    OR
    L<Min(ValueWhen(2,CD=0,L),ValueWhen(3,CD=0,L))),1,0));
    CD2:=If(CD1=0,0,
    If(CD1 AND
    (H>Max(ValueWhen(2,CD1,H),ValueWhen(3,CD1,H))
    OR
    L<Min(ValueWhen(2,CD1,L),ValueWhen(3,CD1,L))),1,0));
    CD2
    My Site
  •  02-18-2005, 19:26 503 in reply to 383

    Oups ! I was on the wrong file !

    Sorry, I did not notice that my first mail was not in the right folder...


    As the formula does not seem straightforward, I started by creating an Expert with your code.
    Bars in blue are for CD2=1 and bars in red are for CD2=0.
    If I understand, CD2=1 are for "active" bars and CD2=0 are for the other bars.
    Your code is quite perfect. I noticed very few errors (labelled "w") on the chart. I will try to use it as it is, as you already helped me a lot.

    I will be away for two weeks, but as soon as I come back, I will try to see if the system I had in mind can be succesful. Be sure that you will be the first to see the results... or to receive some mail if I am not able to build it !

    Thanks a lot !
View as RSS news feed in XML