Because of the alpha creation that results from the use of Ensemble Methods, we consider EAM more than simply a standard wisdom of the crowd approach. However, there is no question that a multi-predictive engine platform has been proven to outperform single predictive engines over longer periods of time.

Regarding your question relating to defining the process (i.e., embedded predictive engine), the easiest approach is to base it upon each fund's holdings data, especially looking at holdings and weights relative to the benchmark. This can be done by outsiders assuming that they have access to timely holdings data. Therefore, there is no need to build a multi-variate model nor have access to the actual managers' decision matrix. It also can be used for managers that are heavily quantitative, heavily qualitative, or any mix of the two.