Fairness and randomness boundaries
Random selection can make equal-probability choices from the entered pool, but it cannot prove that the input list itself is appropriate or unbiased. Visual wheel/dice animations are presentation only. Seeded tools remain reproducible pseudo-random workflows rather than cryptographic randomness, and rating-balanced teams use a transparent heuristic rather than an optimal or human-judgment model.