[R-SIG-Finance] performance evaluation and sharpe ratio

[Bastian Offermann]

Hello all,

i am currently doing some portfolio resampling experiments and wonder 
how to best evaluate different investment strategies based on the sharpe 
ratio.

i do have a time series for 2 equity indices from jan 2002 till dec 2007 
of daily log returns (1548 observations) and perform an unconstrained 
markowitz optimization to obtain both portfolio weights for a given 
return level.

my experiment is based on deMiguel (2007) performing further markowitz 
optimizations on subsamples of 1300 observations each, i.e. the first 
sample is r_1, r_2, ..., r_1300, the 2nd sample is r_2, r_3, ..., r_1301 
and so on. i finally obtain 249 subsamples and as many portfolio weight 
vectors.

how do i best examine each strategy using the sharpe ratio? 
in-sample-test? out-of-sample test?

any suggestion is highly appreciated. thanks in advance.

kind regards

b