>why bother with a benchmark if the pick is pre-ordained

Validating assumptions

Curiosity/learning

Enraging a bunch of HN readers who were apparently born with deep knowledge of PG and Redis tuning