Use for a robust, outlier-resistant central value with weights ($0.01): send observations {value, weight}; get the value where cumulative weight reaches half the total. Deterministic definition, exact ties take the midpoint; works on any numeric data you supply.