|
PureBytes Links
Trading Reference Links
|
It seems that real problem involves the higher moments
of the distribution, right? In comparing a sample of
equity curves, if they are all normally distributed
then there's not much of a problem with the sharpe
ratio. However, if the curves happen to exhibit
varying degrees of skew and kurtosis then there is a
problem in that you aren't comparing apples to apples,
so to speak. An equity curve with a high SR and
negative skew might not be preferable to an equity
curve with a lower SR and positive skew.
Trey
--- DH <catapult@xxxxxxxxxxxxxxxxxx> wrote:
> > Another example: System A returns 0.001% greater
> than the risk-free
> > interest rate with zero-to-few drawdowns, and
> almost-perfect
> > consistency. System B returns 60% per year on your
> account with
> > modest 10% drawdowns. System A will have the
> higher SR due to
> > near-zero standard deviation, but I'd still prefer
> system B.
>
> All these hypothetical system with near-zero SD,
> making the divisor
> approach zero, are fun mental exercises showing how
> Sharpe sucks in
> theory but, in the real world of real systems, it's
> pretty rare to see a
> system with a Sharpe > 1. If you see one with Sharpe
> > 3, it's one you'd
> better pay attention to. ;-)
>
> --
> Dennis
>
>
|