Does SLOPE outperform bridge regression?