A measure of how stable and reliable algorithm rankings remain across different datasets or evaluation conditions.