A blended forecast is a scenario member that is a combination of actuals and plan. Most financial cubes contain a scenario dimension which contains Actual, Plan and then variances. But oftentimes companies want a "blended forecast" which provides actual values in the historical months and then plan in the future months. This is a "rolling" forecast because each month as actuals become "historical" they replace the plan numbers.
Implementing this in a cube
A blended forecast is easy to populate in a MOLAP cube. Rather than "duplicate the data", in the fact table, this can easily be accomplished by using the combination of some named sets and some MDX scripting.
This is something that can be easily accomplished on a remote call.