We have developed a method for optimizing chemical composition of steel to provide maximal compressive residual stresses at the surface of steel parts after intensive quenching. Depending on size and form of steel part, we can answer a question what kind of steel will provide the best results in increasing service life. It is based on calculation of hardenability using Grossmann equation and multiplying factors depending on chemical composition of steel. However, we have met a problem. Within the same grade of steel, hardenability sometimes differs more than three times if chemical composition is at its lower and then at its upper range. How can we manage this problem?