Subject: Re: Question about RRS126-2sd
I would appreciate if you would comment on which is the correct way in my calculations:

Method 1: when I calculate RRS126, in Excel I am calculating the Slope of N stocks taking the Log Prices for 126 days. To calculate 2SD, I am using the Excel function of STDEV.S for Log Returns of N stocks for 126 days as the first step. In the next step I am using STDEV. S *(252)^(1/2), which gives me “Annualized” Volatility of these stocks. As a final step, 2 times the annualized volatility (2sd) is subtracted form the slope (RRS 126).

Method 2: The difference from the first Method is that Multiplying the Slope calculated over 126 days with 252 to annualize it. For 2SD, it is the same as above which gives me “Annualized “volatility. The final step is the same as above, except the Slope has been multiplied with 252 to annualize it.

As I wrote before, I think 2 is wrong because by "annualizing" you are doubling the slope. It has no effect on the ranking of stocks by RRS, but when you subtract the volatility measure it has half the effect on the doubled slope.

So I think method 2 is correct. There is one small difference from what I do. While I calculate RRS126, I calculate volatility (STDEV of returns) on a whole year's worth of price data. It should be over 252 days, but due to practical spreadsheet limitations I've shortened it to 248 days.

Elan