0

There are two methods.

  1. Only the numerator the average realized return will be rolling (10 years) while the downside deviation is calculated from full time period.

  2. Both numerator and denominator are rolling.

Which one would be correct?

Anon9001
  • 91
  • 8
  • 4
    Although 1. might be interesting in some ways, it would not be a "rolling Sortino Ratio". Generally a "rolling" statistic means the same statistic is independently re-computed for each of many partly overlapping sub-periods of fixed length. – nbbo2 Jul 27 '22 at 07:39
  • @nbbo2 So this means the second method is correct? – Anon9001 Jul 27 '22 at 07:47
  • 1
    Personally I would use the second method, yes. – nbbo2 Jul 27 '22 at 15:05
  • @nbbo2 Sorry for the late reply but why the second method? Downside Deviation only punishes downside risk . If you are doing rolling DD it would tend to underestimate risk of asset when the time period selected for the asset contains no bear market. I think its better to use the DD calculated for entire time period for every rolling data point and only change the numerator for every rolling data point due to this. Rolling Volatility makes much more sense to me as standard deviation tends to punish both upside and downside risk. – Anon9001 Aug 28 '22 at 15:24

0 Answers0