chetron
2,817 posts
msg #72709
- Ignore chetron
modified |
3/14/2009 3:06:00 PM
NOPE, IT IS TODAY'S XROC RELATIVE % TO THE XROC 60 DAY HIGH(100%) AND THE XROC 60 DAY LOW(0%)
|
glgene
616 posts
msg #72716
- Ignore glgene |
3/14/2009 4:25:47 PM
Chetron, You said ...."NOPE, IT IS TODAY'S XROC RELATIVE % TO THE XROC 60 DAY HIGH(100%) AND THE XROC 60 DAY LOW(0%) .
What does this get you? How do you analyze the results? In other words, what is good, what is bad, etc.....(If you go along with Appel's "xroc going above 4 is a buy signal for QQQQ, and going below 4 is a sell?"
Zack
|
chetron
2,817 posts
msg #72730
- Ignore chetron |
3/14/2009 8:52:05 PM
I AM SURE THAT Gerald Appel (creator of the MACD Signal), DID A VERY THOROUGH INVESTIGATION, BUT AS I SAID BEFORE, THIS IS MORE OF A PULLBACK TYPE SETUP THAN A CIRCLING THE 4 TYPE SETUP. 4 MAY BE THE LUCKY NUMBER. WHAT INDPOSITION CAN DO IS SHOW PULLBACK SETUPS BASED ON HOW ROC HAS BEEN RANGING OVER THE LAST 60 DAYS.
FOR EXAMPLE, IF THE ROC IS RANGING BETWEEN 100 AND 50, MAYBE A PULLBACK IN THERE MAY BE JUST AS VALID. JUST GIVING US THE TOOLS TO EDUCATE OURSELVES.
|
glgene
616 posts
msg #72733
- Ignore glgene |
3/15/2009 1:20:58 AM
OK, so how would you analyze INDPOSITION over the past 1 year. What do you see as the level for buying/selling? 50? 38? What? Just looking for something --- from you --- on a specific number that I can review. I want to see if I'm in the league for the possible number you may have in mind.
Does the number change every 60 days or so?
Zack
|
chetron
2,817 posts
msg #72755
- Ignore chetron |
3/16/2009 6:47:22 AM
i haven't seen any specific number come up consisently.
|
glgene
616 posts
msg #72835
- Ignore glgene |
3/17/2009 4:33:38 AM
Chetron,
After exploring the ROC numbers, and your suggested scripting, I would like to explore one other avenue. I would like to create a column (and display) of the 13-day EMA of closing prices, and apply a 21-day ROC to it (as suggested in "Trading for a Living" by Elder. I would like to compare this result vs. the "xroc" column, as shown in your scripting, in order to generate fewer buy/sell round trips.
I know how to show a 13-day EMA of price, and I know how to show a 21-day ROC of price, but I'm not sure how to combine the two (as described in first paragraph). Could you help me with that combo statement? Do I need another variable to include in the script?
I want to draw it and display it, as said before.
Thanks, again.
Zack
|
glgene
616 posts
msg #72836
- Ignore glgene |
3/17/2009 4:36:54 AM
Chetron,
Forgot to ask this question in the last msg. I posted. How can a ROC time span (21 days) exceed the EMA price time span (13 days)?
|
chetron
2,817 posts
msg #72839
- Ignore chetron |
3/17/2009 6:48:44 AM
maybe i am not understanding you,z, but the default ema for SF roc's is 13, so it should just be
add column roc(21)
and as for how you can have a 21 day roc on an ema 13 is that as long as there was an ema 13 result 21 day ago, you can roc it.
hth
|
glgene
616 posts
msg #72858
- Ignore glgene |
3/17/2009 1:52:33 PM
Chetron,
ROC(21), to my knowledge, isn't smoothed by a 13-day EMA.
I checked closing prices the other day ago, and ...
ROC(5) merely compares today's closing price to the closing price of EXACTLY 5 days ago. No smoothing.
Same with ROC(15) and ROC(25).
A smoothed ROC, instead, compares the value of an EMA instead of two points in a time. This from the book, "Trading for a Living" by Alexander Elder.
Alexander says that ROC has a flaw with simple (or exponential?) moving averages. It jumps twice in response to each piece of data. It reacts to each new price (today's price) and jumps again when that piece of data leaves the oscillator's window (Day 6 when it is replaced when new Day 5 now comes into being).
To create a smoothed ROC, you must (1) First calculate an EMA of closing prices. (2) The next step is to apply ROC to the price EMA.
Alexander's exact suggestion is.......Calculate a 13-day EMA of closing prices and then apply a 21-day ROC to it.
He goes on to say, "Some traders calculate the ROC of prices first, and then smooth it with a moving average. Their method produces a much jumpier indicator, which is less useful than the smoothed ROC."
Does that clarify it, Chetron? If so, can you provide me with the scripting to do it in the 1-2 sequence as shown above? I would like to draw it, as well as display the resulting number.
Many thanks.
Zack
|
chetron
2,817 posts
msg #72871
- Ignore chetron |
3/17/2009 8:23:46 PM
OK. YOU ARE RIGHT.
SF HAS A SMOOTHED ROC FUNCTION ALREADY AVAILABLE, SO I GUESS THIS IS IT.....
ADD COLUMN SMROC(13,21)
|