An analysis of world class performance improvement over chronological time was used to develop mathematical curves of performance for each speedskating event. The equations were calculated using an unconstrained non-linear least squares iterative curve-fitting technique. A non-linear model was selected to satisfy the principle of diminishing returns, that is, it is more difficult to achieve a unit of performance improvement as performance approaches the theoretical limits of man. Minimum criteria of acceptance for each curve were: 1) the coefficient of determination must be 0.90; 2) the year 2000 predicted value must be less than the current world record; and 3) the curve must be progressive, i.e. satisfy the principle of diminishing returns. Satisfactory curves were calculated for seven of the eight male and female events studied. A four year Olympic cycle was noted and a definite change in performance trend was identified as occurring in the mid 1960's. The calculated mathematical curves can be used in three applications: 1) setting objective individual goals and evaluating these goals; 2) evaluating and comparing training programs on successive years; and 3) evaluating total programs.