I agree that there's no real right or wrong answer. Whatever works with your workflow but keep enough of an open mind if a different process might streamlines things or improve your process a bit
For me, if I'm chimping trying to get the lighting perfect, I'll delete those right out, but when I'm shooting I tend to just leave the images as-is. Keep in mind, I shoot RAW.
When i get back to a "base camp" (home, hotel room, etc) and download the images (and rename them to the overall shooting project and time/date shot sequence), I'll clear out any I deem unsalvageable or not worthy of keeping: a misfocus, the unideal brackets, shots in a machine-gunned sequence (action shots), etc. that are less than ideal for the composition for the idea of the shot. Depending on what I'm shooting, this could be either 2/3rds worthy of post work (usually a staged shot) to 1 in 10 worthy of post (BIF/action/brackets)
From there, I'll edit in post, and rank them based on the result. If the result is good enough as a snapshot/low res web image naturally they'll be ranked lower. A large-format print-worthy shot will be ranked the highest. A print-worthy shot that had to be cropped will be slightly less. etc. All of them are stored on a RAID array for immediate editing with about 2TB of local access space, and then mirrored onto my NAS.
As I learn a bit more post tricks, I'll go back and retry some stuff. Usually about once every 6 months or so, if I have nothing better to do I'll try and re-work an image ranked lower, enhance an image ranked higher, and then re-decide if I want to keep or cull. In general, higher ranking shots are kept, lower-ranked shots are tossed, and occasionally one moves up or down the ladder.
The only exception is a "sentimental" shot that involves something that a family member really likes, even if it's not a good technical shot and I'd usually toss it, I'll keep it.