Rotationsfairness
Rotationsfairness is a fairness criterion in algorithmic decision making and process design that requires treatment to be invariant under cyclic rotations of positions or options. In systems where items are displayed or assigned in a round-robin or circular order, Rotationsfairness aims to prevent positional advantages by ensuring that no option benefits solely from its fixed position; over time, each option should be exposed to each position with equal frequency.
Formally, consider n positions arranged in a cycle and a policy that assigns outcomes based on position.
Applications of Rotationsfairness appear in various domains. In user interfaces and recommender systems, it helps reduce
Evaluation typically involves exposure and outcome metrics, rotation-invariance tests, and permutation-based analyses to detect deviations from