Abstract: Existing approaches to combining multiple forecasts usually offer
     either theoretical richness or empirical robustness, but rarely both.  A
     new method for combining forecasts, which attempts to overcome this
     imbalance, is proposed.  The proposed Odds-Matrix (OM) method allows easy
     inclusion of relevant subjective and empirical information about the
     forecasts, while providing weights that are: 1.  intuitively meaningful,
     and 2.  not dependent upon large numbers of observations of prior forecast
     accuracy.  A simulation study was conducted to test the proposed approach
     against currently available methods.  Results indicated that the approach
     performed at least as well as existing methods when large quantities of
     data were available and significantly more accurately when data were
     sparse.  The new method also was found to be computationally simpler than
     traditional methods.