Registered User Joined: 2/14/2006 Posts: 15
|
Hi,
I'm attempting to produce a PCF which will give a measure of how well the daily price spread of a stock fits in/outside a fixed width channel over a defined period.
I'm not that interested in the direction of any trend, whether it is bullish, bearish or even a horizontal channel, just the spread of the daily price bar in/outside the channel over the period, e.g. a reading of 0% would indicate that the spread of the daily price bars all fall within the channel, 50% indicates that only half the spread of the daily price bars lies within the channel either because some daily highs lie above the channel and/or some the daily lows lie below the channel. I do not want the percentage to reflect the number of days where price lies in/outside the channel, but the percentage of the price itself which lies in/outside the channel.
e.g. I want to generate a 30 day linear regression line based on the closing price bounded by a fixed width channel which could be anything e.g. +/- (AVGH30 - AVGL30)/2, +/- (H - L)/2, +/- C*0.1 etc etc.
I think I've got the PCF for generating the slope of the 30 day linear regression of closing price:
( 14.5*C + 13.5*C1 ... - 13.5*C28 - 14.5*C29) / coefficients^2
which means I can go back and generate the channel bounds on any of the past 30 days to see what amount of price was outside the channel, then divide the sum for each day by the sum of (high - close) for each day.
However I can't help thinking that I am completely overcomplicating the PCF or that a standard indicator already exists (maybe based on linear regression channels, standard deviation of price?) to generate something similar to what I am looking for.
Any advice gratefully received.
Best Regards Richard
|
Worden Trainer
Joined: 10/7/2004 Posts: 65,138
|
I don't know a way to find out what percentage of a price range was outside an arbitrary channel without explicitly calculating it in the manner you outlined.
-Bruce Personal Criteria Formulas TC2000 Support Articles
|
Registered User Joined: 1/28/2005 Posts: 6,049
|
rwesley
Quote:"However I can't help thinking that I am completely overcomplicating the PCF or that a standard indicator already exists (maybe based on linear regression channels, standard deviation of price?) to generate something similar to what I am looking for."
I think you have answered your own question.
I don't think for all the complications your indicator will offer any more than a liner regression channel sort. You can also sort by the percent slope of the line.
The only way I can think of to "count" things in a channel is with a percent true indicator.
Something like C<AVGC10*1.03ANDC>AVGC10*.97 this will trigger every time the close is inside a +-3% band of a 10 day moving average. You can add from the indicator window a 50 day moving average(simple) This will give you a percent over the last 50 bars of how often it was true.
Of course the "50", "10", and +-3% can all be adjusted by you.
Hope that helps Good Luck
|