[tempguy9999]

Thank you for a helpful comment. I can see the initial comment wasn't sufficient now you've explained how it looks to you.

The 2nd post is harder to justify in a way. Some of it comes from experience, and understanding the behaviour of a merge join vs a loop join for example, and how much slower the latter can be if applied wrongly, which points to why the optimiser is so vital.

Edit: there's also the cardinality estimation (which can so easily go wrong), the dynamic programming for joins and the combinatorial explosion when it comes to ordering/reordering of inner joins, heuristic costs of using indexes vs not, etc.

Some of it was hearsay, but I stated it as such. Anyway, point taken, and thanks again.