Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User Joined: 4/1/2005 Posts: 1
|
what is the pcf for the indicator called the Chaikin Oscillator?
|
|
Worden Trainer
Joined: 10/1/2004 Posts: 18,819
|
Check out this topic:
Chaikin Money Flow
- Craig Here to Help!
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
ggh22650,
I assume that you have discovered by now that Chaikin Money Flow has little to do with the Chaikin Oscillator.
My first impression is that a PCF for the indicator is well beyond my endurance limits and I am sure even that of Bruce. There is good news. If the reason you want the PCF is to plot the indicator, that is within my endurance limits, and well within that of Bruce.
If a plot is what you want, let's talk to Bruce. Others may be interested in the indicator.
Thanks, Jim Murphy
|
|
Worden Trainer
Joined: 10/1/2004 Posts: 18,819
|
Plotting this will be very easy:
Create a new CUSTOM CUMULATIVE INDICATOR
Set the Period to 1
Use the following formulas:
Up Condition: C=C
Down Condition: C>C
Add/Subtract: (C-O) / (H-L) * V
UNCHECK the visible option.
Close the window.
Click the EDIT button on the chart.
Click the PLUS next to your new indicator (which is the Accumulation Distribution indicator, BTW)
Add a 3 period exponential moving average to it.
Click the PLUS again (next to the new indicator)
Add a 10 period exponential moving average to it.
This will plot the two moving averages that make up the Chaikin Oscillator. Normally this indicator is presented as a histogram but you cannot plot histograms as custom indicators and I don't know a way to create a custom indicator that takes the difference between two moving averages of a cumulative indicator.
When your short term MA (the 3) is above the long (the 10) it is the same as the Chaikin Oscillator being above zero. If the short is below the long, it is the same as the Chaikin Oscillator being below zero.
bustermu, am I mistaken in thinking there is no way to create a PCF for the difference of two MAs of a cumulative indicator (other than ones already in the language like price, MS or OBV)?
- Craig Here to Help!
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
Craig,
I am not ignoring your excellent post.
Some comments:
1) Value To Add/Subtract is more frequently V*(2*C-H-L)/(H-L) than V*(C-O)/(H-L). Both are correct.
2) The values of the EMA3 minus the EMA10 on your plot are the values of the oscillator, which is more than just determining the sign of the oscillator.
3) Since EMA's commute with the accumulation operation, one can plot the oscillator as a line chart. The problem is that the zero line has to be drawn for each new chart setting.
4) I am unaware of anything useful in TA that can be computed on a Symbol's data sequence that cannot be accomplished with the PCF Language. But, as previously stated, the Chaikin Oscillator is well beyond my endurance limits.
I will illustrate how all of this can be done if someone else does not.
Thanks, Jim Murphy
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
Addendum:
Statement 3) is in need of some (face saving) adjustment.
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
For a further explanation of Statement 3), please see my 06/25/05 12:41:32 PM post and that which follows at:
Applying a Moving Average to a Moving Linear Regression
In order to plot the Chaikin Oscillator as a line chart, you need only plot:
XAVG(PCF1,3)-XAVG(PCF1,10)
as a Custom Cumulative Indicator. The PCF1 is:
V*(2*C-H-L)/(H-L)
or
V*(C-O)/(H-L)
as you choose. The EMA3 and EMA10 have to be written as series.
Further details will be supplied if needed.
Craig. This is a way to create a plot for the difference of two MAs of a Cumulative Indicator. A PCF for the difference is a lot longer.
Thanks, Jim Murphy
|
|
Worden Trainer
Joined: 10/1/2004 Posts: 18,819
|
I am not sure I understand.
The MAs need to be for the value of the indicator. The PCF1 returns the change in the indicator, not the value of the indicator.
If the indicator had a value of 10 yesterday, to calculate its value today we would add the results of PCF1 to the value. PCF1 would not be the new value.
Does the difference between the average changes the same as the difference in the average of the values of the indicator?
- Craig Here to Help!
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
Craig,
Thanks for your response.
Please notice that:
XAVG(PCF1,3)-XAVG(PCF1,10)
is plotted as a Custom Cumulative Indicator, NOT as a Custom Indicator. That may clear up everything.
I am unable to interpret your question, but maybe the following will answer it. Let ACC represent an Accumulator. Then, in a qualified sense:
XAVG(ACC( C ),10) = ACC(XAVG(C,10))
just as, for example:
XAVG(AVG(C,5),10) = AVG(XAVG(C,10),5)
When the ACC is implimented as a Custom Cumulative Indicator, a constant must be added to the right-side of the first to account for initialization. This constant will change with chart settings.
This commutative property is demonstrated in the referenced post.
Did I answer your question? If not, please rephrase it and I will try again.
Thanks, Jim Murphy
|
|
Gold Customer
Joined: 5/4/2005 Posts: 74
|
I believe V*(2*C-H-L)/(H-L) is the only correct form. The intent is to accumulate an "earned volume" in the range -V < earned volume < +V. The end points of the range occur when C=L and C=H respectively. Also, when H=L, then C=H=L and the result is the indeterminate 0/0. Therefore, it would be appropriate to add the usual .001 in the denominator.
|
|
Registered User Joined: 1/1/2005 Posts: 2,645
|
My statement concerning the difficulty in writing a PCF for the Chaikin Oscillator is incorrect.
Please see:
Chaiken Oscillator
Thanks, Jim Murphy
|
|
Guest-1 |