Could you please explain How do we do that?
Well, first, the practical answer:
Just a speculation, but I think if you sent Elann a very very polite private message he might send you the spreadsheet he uses.
I think such a thing exists, not sure.
Here is a summary of my flawed understanding.
The RRS is not too bad, but I have never implemented the -2sd adjustment, so that needs checking.
You need six months of daily split-adjusted price history for each of the ~400 stocks that are Timeliness 1 and 2.
With that in hand, it’s just a matter of arithmetic.
First, calculate the five day return for each of those ~400 stocks.
You don’t want anything that just popped a lot, as those pops tend to unwind.
So you do the DH10, i.e. drop the 10% of stocks, 40 stocks, that had the highest 5 day return.
For each of the remaining 160 stocks:
Take the prices for the last 6 months for the first stock, let’s say in column A.
Take the log of all those prices =ln(a1) in column B.
Calculate the slope, =SLOPE(B1:B126,A1:A126)*240
Repeat for each of the ~400 stocks that are T1 or T2.
That’s the RRS, the Regression slope we will used to sort for Relative Strength.
Also, calculate the Standard deviation of one day returns. (the -2s or -2sd in the screen name).
Calculate the 126 days of one day returns from the price series.
From a very old post: " volatility as of the close of TradeDayNo d is computed by applying Excel’s STDEV function to the Natural Logarithms of TR1 for TradeDayNos d - 239, d - 238, …, d - 1, d, and multiplying the result by (240)^(1/2).
With that in hand, for each stock calculate [slope minus two times that standard deviation].*
This [slope-2sd] finds those that are not just up a lot, but trending smoothly upwards a lot, eliminating those that are up a lot but jagged.
Then, sort the ~160 stocks based on that metric rrs-2s
The stocks you want are the ones with the highest slope measured that way.
Go down the list looking for the highest ranked picks that have suitable call options available.
With the search engine, you can find m any fine old posts on the subject that will be more correct than my precis here.
Search engine: http://www.datahelper.com/mi/search.phtml
Example find: https://discussion.fool.com/at-the-top-of-this-thread-vizcacha-c…
Jim
- I’m not 100% sure about units at this step. Needs checking.
I always thought it should really be the RMS of the daily stock price deviations from the trend line, but I digress.
** you may have spotted that odd 240 in the formula above.
See this ancient post
https://discussion.fool.com/new-backtest-method-on-sux-data-1437…