| jfh123 |
|
Gold User, Member, TeleChart
|
| Registered User |
|
|
|
|
| Unsure |
|
| Friday, December 17, 2004 |
| Thursday, August 25, 2011 11:26:05 AM |
143
[0.05% of all post / 0.02 posts per day] |
|
|
Thanks, anyway.
|
QUOTE (Bruce_L) I don't know how Envelope Channels are calculated in TeleChart and have been unable to reproduce them using formulas. The Envelope Channel Width calculations used in TeleChart appear to be very similar to the Linear Regression Channel Width calculations:
Linear regression channel calculations
That said, I have not been able reproduce the Envelope Channel Width calculations exactly and cannot provide any additional details. You can use Channel Sorts on Envelope Channels however. You may wish to review the following:
Sort using envelopes, regression channels and Bollinger bands
B-
I want to APPROXIMATE the formula for envelope channels. For my purposes, LR Width calculations would be similar enough. But, when I tried the LR settings you provided in the link (60,10), and an envelope with the same settings, they didn't seem to agree. Would you help me find my error?
Here's an example: On SP-500, on 7/13/09, LR upper = about 935 and LR lower = about 913, but the lower env chnl = about 801 and the upper env chnl = about 1002.
Thanks!
J
|
|
Thanks.
|
B-
Would you remind me how you get the constant which begins the formula?
J
|
|
Exactly. Bummer. Thanks for looking at it.
|
I can't begin to figure out a workable way to write a PCF for this. What I tried was to divide the price range, over a time frame (let's say 20 periods), into pieces (let's say 10 deciles). I can easily find the 20-day high and low. Next, I'd assign each day's volume to a decile via (C>MINL20+(MAXH20-MINL20)*.3 AND C<=MINL20+(MAXH20-MINL20)*.4)*(-1)*V (assumes today is in the 4th price declie from the bottom). Then I would do the same for C1 thru C19, and sum them, to get total volume in that declie. Then, I'd have to do the same for the other 9 declies to know which had the most volume, then I'd use the top and bottom of that price decile as the boundaries of my support/resistance line (based on whether price is above or below it. But, there has got to be an easier way, no?
In case anyone's unclear on what I mean by a "volume by price overlay", an example is at the following URL: http://stockcharts.com/school/doku.php?id=chart_school:technical_indicators:volume_by_price
But I don't want a logarithmic chart, and I expect it'll be most doable if all "slices" are the same "height". All I want is to know the top and bottom of the horizontal slice of the chart which has the most volume.
I'd appreciate any help anyone can provide.
|
Could it be that, for the first time ever, I can solve something asked of Bruce, instead of asking Bruce to solve something for me? 
If I understand the request correctly, here's my "solution" to plotting any horizontal line: (C>0)*(-1)*18
This assumes you want the horizontal line to be $18.00. Unfortunately, I don't know of a way to limit it to certain days; it goes across them all, as long as C>0.
You can approximate the 52wk high as MAXH252.
Hope this helps.
|
|
Now that I think about it, 4 cents is not gonna make or break any strategy. Thanks for your reply.
|
Bruce,
I'm using your Wms %R, and it's working fine. I'd also like to program a SL related to it, and am confident it'll be easy for you to tell me what I'm doing wrong.
I want the following CI to give me a zero if the three conditions aren't all true, and yesterday's low if they are all true:
(STOC30.1.1>80 and c>h1 and L>L1)*(-1)*L1
It does, but when I multiply it by 100, it gives me -.02 (see stock = MA, daily, on 5/2/07) when 1 or more conditions is false.
Shouldn't it be giving me a zero? Because my ulitmate goal is to add to it the same test on 2-5 days back, it can end up being 4 cents off. Has TC2K got a rounding function I need to use, so that a zero is really a zero?
|
B-
These formulas match the CPC exactly. Now I can now test it. If it performs well, it will be what I want. 
Thanks.
J
|
|