Work

Lower tail intuition fails: maximum outperforms minimum in replicated ranked unequal–set sampling for weighted exponential models