| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User
Joined: 12/17/2004
Posts: 3
|
Hi there, how is it possible to create a scan that works with weekly or even monthly charts? E.g working with weekly stoch/macd values rather than daily. Something simple like finding the 9 weekly macd value 5 weeks ago compared to 9 weekly macd value today.
Thanks,
Fraser.
|
|
Worden Trainer
Joined: 10/1/2004
Posts: 18,819
|
It is not very feasible to effectivly scan for indicators on weekly or monthly charts using EasyScan and/or PCFs. Good news - Visual Sorting makes this very easy to do! Watch these videos, I think you will see how easy it is.
Visual Difference Sorting
Actual Value Indicator Sorting
Visual Change Sorting
- Craig
Here to Help!
|
|
Registered User
Joined: 10/7/2004
Posts: 799
Location: Duluth, GA
|
Hi, Fraser:
MACD's, MACDH's, Stoc's and other PCF's can be created for weekly or monthly or whatever days/bar, which match what the chart shows. The *easiest* way to deal with it is to use SortBy, as Craig says. However, if you really need a weekly-based PCF for MACD, post the param's here, please.
MACD:
Days/bar = ? (weekly = 5)
Fast EMA bars = ? (typical = 12)
Slow EMA bars = ? (typical = 26)
I don't follow your "9 weekly" terminology. Do you mean you want a MACDH (histogram), with the third EMA term being 9 bars?
You said you want to compare the MACD value today to the one five weeks (i.e. five bars) ago. This is not a "standard" use of the MACD ... but certainly is do-able. Could you please explain a bit better what you're trying to accomplish?
I'll provide a PCF for you if you give me enough info to go on, to illustrate that this is possible.
Jim Dean
|
|
Registered User
Joined: 2/14/2005
Posts: 12
|
I am also interested in the Daily ,wkly and mthly scans.I want to do it with the Macd and Cci and Williams %R ? can it be effectively done and can they be done simultaneously (all 3 indicators same time) ? The basic Macd I am using is 3,22,100. What formula would I use ? Thanks
|
|
Worden Trainer
Joined: 10/1/2004
Posts: 4,308
|
As Craig mentioned above, it is not very feasible to accomplish this using scans. However, when you sort by these various indicators, you can change your chart to weekly, or monthly, and the sorts automatically adjust to whatever chart settings and time frames you have displayed. I encourage you to watch the videos that Craig recommended by clicking on the links in his post above.
- Doug
Teaching Online!
|
|
Registered User
Joined: 10/7/2004
Posts: 799
Location: Duluth, GA
|
Hi, John:
As I mentioned earlier, it's possible to write PCF's for virtually any barlength for most indicators ... but some of them (particularly for MACDH's) can get pretty long.
Williams %R is just an "upside down" Stochastic, which is a medium-length PCF expansion for multiday bars. The CCI for multiday bars is considerably longer, due to its Avg components. An exponential MACDH, particularly with your parameter settings, will be *really* long.
All of it can be done, and I will be happy to provide examples if you REALLY need them, but as Doug and Craig mentioned it is much better to custom indicators and sortby to do this, unless you are willing to put up with some really long recalc times.
In various places all three of the indicators you mentioned have been formulated or discussed ... the MACDH of course is a native plottable function, and you can substitue a native Stochastic plot for the Williams %R without introducing errors if you adjust your signal-logic accordingly. The CCI requires a Custom Indicator, which you can find described by doing a Search for CCI in the forum. Once these are plotted, you can set whatever barlength you wish, and SortBy will assist you with the scans.
Please try that out, first.
If, however, you really MUST have the PCF's, then you need to provide explicit parameters for each of the three indicators. They will be hard-coded for those parameters, as well as the barlengths you want. Also, you need to provide explicit LOGIC regarding what you want to filter for - i.e. crossovers or thresholds or slope or whatever.
Bottom line ... you will probably end up with an all-in-one PCF that is in the tens of thousands of characters long ... and will probably take the better part of an hour to recalc.
So, even though it is *possible* to do it in PCF-eze, it's just not *practical* to use.
Final comment. IMHO a MACDH spec of 9,22,100 is pretty absurd (if you don't mind my saying so :~). The final smoothing is so long that the relatively small variantion between the 9 and 22 EMA's is overwhelmed by the long signal-line average. But ... each to his own ... if it ain't broke don't fix it!
Jim Dean
|
|
|
Guest-1 |