|| This report assesses the relative performance and utility of some basic procedures for forecasting adult returns to four salmon stocks in northern British Columbia. These include sockeye runs to Long Lake, Owikeno Lake, and the Skeena River, and even-year pink salmon returns to the Bella Coola system. These stocks were selected because adequate data were available, enabling comparison of the performance of a variety of simple biological and non-biological models. The report also documents preseason 1997 run size forecasts for nine sockeye and five pink salmon stocks or stock groupings in northern British Columbia. Algorithms for each forecasting model provided forecasts of stock size one year in advance and were evaluated by hindcasting stock size in previous years using only the data that would have been available at the time. This retrospective analysis provides a robust measure of how well the various forecasting algorithms would have worked had they actually been used.