| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User
Joined: 3/29/2011
Posts: 10
|
Is it possible to weight different return dates to come up with a cumulative ranking for a watch list? For instance, out of all the commodity ETF's in a watch list I wish to weight the 20 day return .5 and the 30 day .5 to come up with a cumulative statistic. Is this possible and if so how would I approach it?
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
You could write an Indicator Formula to come up with the a value combining the 20 and 30-day returns for the individual symbols in a WatchList as:
50 * (C / C20 - 1) + 50 * (C / C30 - 1)
But it is not currently possible to then calculated an average of this value for the entire WatchList in either TeleChart 2007 or TC2000 version 12. You could do so using Market Indicator feature in StockFinder however.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 3/29/2011
Posts: 10
|
If I wanted to add volatility as measured by either 20 day standard deviation or possibly ATR and still rank them the same way, likely just applying my scan to a column. What would be the best way to write that code? Keep in mind I would want low stdev and I would want to weight them into the equation we talked about earlier. Thanks for your help.So it would be something like40*(c/c20)+40*(c/40) 20*(the lowest standard deviation or ATR)Ideally I want to look at high relative strength low vol stocks and ETFsThanks for all your help!-Nate
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
You would need to start by determining if you want Standard Deviation or Average True Range.
You would then need provide a period for the SD or ATR.
You would then need to specify a mechanism for making small SDs or ATRs provide larger values than small SDs or ATRs. This method would preferably provide results which are comparable to 20 and 30-Period Price Percent Changes and which return similar values so the results would neither overwhelm the SDs or ATRs or be a negligable portion of the result.
I really do not know what you would want to do for that third step.
Expressing the SD or ATR as a percentage of price would probably return comparable values. They would be larger values for large SDs or ATRs and smaller values for small SDs and ATRs however. In addition, while being a percentage and comparable, the values are not guaranteed to be similar in range to the Price Percent Change figures.
You could divide the previously calculated value by the SD or ATR expressed as a percentage of Price, but this would through out the idea of weighting the SD or ATR relative to the Price Percent Changes.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 3/29/2011
Posts: 10
|
Bruce,A 20 period standard deviation would suffice. I think your idea for creating a mechanism that makes low vol appear high is exactly what I want.Maybe I complicated the last step more than it needs to be. I'm essentially just looking to create a weighted average of the metrics mentioned where my various return periods carry different weights and volatility carries a weight.How would you go about writing the formula if I just choose to use 20 day stdev and wish to weight it 20%?Thanks for your time
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
I will repeat that I really do not know what you would want to do for that third step. I explored why this is the case in the last two paragraphs of my Wednesday, August 17, 2011 3:55:17 PM ET post. If you have something specific in mind, I will do my best to help you implement it.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 3/29/2011
Posts: 10
|
I believe I could just subtract the weighted standard deviation.like...40*((c/c20)-1)+40*((c/c40)-1) - stdev of (last 20 closes)the higher the stdev the more it will take away from the weighted averageI believe this is logical, do yo agree?
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
I neither agree nor disagree. I can write a formula for .4 times the 20-Period Price Percent Change plus .4 times the 40-Period Price Percent Change minus the Standard Deviation expressed as a percentage of the Moving Average of Price over the same period as the Standard Deviation and let you decide for yourself however.
40 * (C / C20 - 1) + 40 * (C / C40 - 1) - 100 * SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 - 20 * AVGC20 ^ 2) / 20) / AVGC20
You can adjust the weightings or express the SD as a percentage of Price instead of the 20-Period Simple Moving Average of Price if you would like.
Modelling Bollinger Bands (& Standard Deviation) in a TC PCF
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 3/29/2011
Posts: 10
|
Do you know the pcf for the annualized standard deviation of daily returns?Thanks, Nate
|
|
Registered User
Joined: 3/29/2011
Posts: 10
|
20 day that is
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
We can annualize the daily returns, but not the standard deviation. So while your original request is not possible, you can write a formula for the 20-Period Standard Deviation of Annualized Daily Returns:
100 * SQR(ABS(((C / C1) ^ 253 - 1) ^ 2 + ((C1 / C2) ^ 253 - 1) ^ 2 + ((C2 / C3) ^ 253 - 1) ^ 2 + ((C3 / C4) ^ 253 - 1) ^ 2 + ((C4 / C5) ^ 253 - 1) ^ 2 + ((C5 / C6) ^ 253 - 1) ^ 2 + ((C6 / C7) ^ 253 - 1) ^ 2 + ((C7 / C8) ^ 253 - 1) ^ 2 + ((C8 / C9) ^ 253 - 1) ^ 2 + ((C9 / C10) ^ 253 - 1) ^ 2 + ((C10 / C11) ^ 253 - 1) ^ 2 + ((C11 / C12) ^ 253 - 1) ^ 2 + ((C12 / C13) ^ 253 - 1) ^ 2 + ((C13 / C14) ^ 253 - 1) ^ 2 + ((C14 / C15) ^ 253 - 1) ^ 2 + ((C15 / C16) ^ 253 - 1) ^ 2 + ((C16 / C17) ^ 253 - 1) ^ 2 + ((C17 / C18) ^ 253 - 1) ^ 2 + ((C18 / C19) ^ 253 - 1) ^ 2 + ((C19 / C20) ^ 253 - 1) ^ 2 - 20 * (((C / C1) ^ 253 + (C1 / C2) ^ 253 + (C2 / C3) ^ 253 + (C3 / C4) ^ 253 + (C4 / C5) ^ 253 + (C5 / C6) ^ 253 + (C6 / C7) ^ 253 + (C7 / C8) ^ 253 + (C8 / C9) ^ 253 + (C9 / C10) ^ 253 + (C10 / C11) ^ 253 + (C11 / C12) ^ 253 + (C12 / C13) ^ 253 + (C13 / C14) ^ 253 + (C14 / C15) ^ 253 + (C15 / C16) ^ 253 + (C16 / C17) ^ 253 + (C17 / C18) ^ 253 + (C18 / C19) ^ 253 + (C19 / C20) ^ 253 - 20) / 20) ^ 2) / 20)
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
|
Guest-1 |