Hi, thanks for the video, I'm just not convinced by the 30 denominator at the end to compute the sortino. I think it should be the number of days the return falls short of the MAR. For example let's say you used 255 daily returns, and you had just one day below the MAR, but that day would be -100%, (-100%)^2 = 1. 1/255 = 0.004 or 0.4% downside deviation, which annualized would be 1.36%. A really good number... despite having a day where you basically lost 100%... that doesn't sound right to me
Can you please clarify upon the value of n we take for the calculation of downside deviation. ( Is it the number of observations below expected value or total no of observations)
Basically with your way of computing the sortino, it does not matter if the portfolio manager had one really bad day or just many small bad days. that's not right, it should be much better if the PM had many small bad days.
@cheznikos Thanks. 1. It's not my way, it's per GIPS and the correct technical definition; 2. I agree that you have an excellent argument, I halfway come to a similar conclusion, but 3. the weakness in your remedy is: you would treat similarly the manager who underperfomed (e.g) 1% one day, and the manager who underperformed every day
At the point where you're calculating the Monthly Variance, you seem to divide by 160, not 30 (you quickly show to contents of that cell being "=SUM(J16:J45)/160"). Why..?
Classic BT video - clear, concise and accurate. Thanks a lot.
Performance ratio ranking==> Generalized Rachev Ratio> Farinelli-Tibiletti ratio> Sortino-Satchell ratio> Sortino ratio>Sharpe Ratio.
Hi, thanks for the video, I'm just not convinced by the 30 denominator at the end to compute the sortino. I think it should be the number of days the return falls short of the MAR.
For example let's say you used 255 daily returns, and you had just one day below the MAR, but that day would be -100%, (-100%)^2 = 1. 1/255 = 0.004 or 0.4% downside deviation, which annualized would be 1.36%. A really good number... despite having a day where you basically lost 100%... that doesn't sound right to me
Can you please clarify upon the value of n we take for the calculation of downside deviation. ( Is it the number of observations below expected value or total no of observations)
Basically with your way of computing the sortino, it does not matter if the portfolio manager had one really bad day or just many small bad days. that's not right, it should be much better if the PM had many small bad days.
Hi! Can the Sortino Ratio be applied on a single stock or is it just for portfolios?
Hi Bionic Turtle, where can download the Sortino ratio (versus Sharpe ratio) spreadsheet?
hello, I have seen the Sharpe use the port return - risk free (u are using "average" in the initial narrative)
can you share the link please for spreadsheet
@cheznikos Thanks. 1. It's not my way, it's per GIPS and the correct technical definition; 2. I agree that you have an excellent argument, I halfway come to a similar conclusion, but 3. the weakness in your remedy is: you would treat similarly the manager who underperfomed (e.g) 1% one day, and the manager who underperformed every day
How did you get the downside deviation?Thank you
Would be useful to explain what a number of 0.39 actually means or implies instead of just a number.
At the point where you're calculating the Monthly Variance, you seem to divide by 160, not 30 (you quickly show to contents of that cell being "=SUM(J16:J45)/160").
Why..?
I60* (green cell) not 160
yes this is correct; the bionicturtle is wrong
sorry annualized would be 0.04%*sqrt(255) = 6.26%, still a too good number
The denominator in standard deviation is (N-1) not N
Pouf...it is going over my head