In-Sample and Out-Of-Sample:

We again emphasize that all optimization - wherever applicable - was done in an in-sample period $t\in[-1249,0]$ and all performance evaluation was done in the out-of-sample period $t\in[1,1001]$.