Forums  > Pricing & Modelling  > Alternative method of computing standard deviation of returns  
Page 1 of 1
Display using:  


Total Posts: 1
Joined: Jan 2023
Posted: 2023-01-19 02:19
I came across a more complex model and as part of it they used a different version of Sharpe ratio where the denominator was computed differently than normal. Normally, we take the series of returns and compute the standard deviation, but this different method takes the standard deviation of the percent difference between the actual prices and the prices of the zero variance returns that gives the same total return.

Attached explains why we would want to make the change. The grey curve is the zero variance returns, ie returns are constant for each period. Both other lines have the same returns, just in a different order, so their standard deviations of returns are exactly the same and yet we can clearly see that the orange line is more risky. The bordered box shows that this metric correctly identifies the firm with better volatility, where standard deviation of returns fails to distinguish.

Has anyone seen this metric out in the wild before? Does it have a name?

Attached File: stand_dev.xls

leaf worm

Total Posts: 1
Joined: Jan 2023
Posted: 2023-01-19 22:07
I've been working in something like that for the price of gold with no useful result But I am a worm only...

I am trying to attach a pic, with the price above and the indicator bellow Please note the grey line, to compare the upside...

Time is in the x axis in seconds from 1970 GMT - 3

If you need more details and share your findings with me you can contact me directly via /*mcabrera (at) protonmail (dot) com*/

Previous Thread :: Next Thread 
Page 1 of 1