I don’t think it can be quite so rule based. Here’s a quote from Um Bongo from awhile back.
Somewhere back in the mists of time I put together a fairly authoritative answer on this particular question as it had come up for serious debate - I tried searching for it, but I think the terms were too common.
Put simply you have two routes - Occluded and Open, though frequency compression gives you some more room at the margins, the overall approach is the same.
Occluded: essentially stuff the biggest receiver on there, block the canal with a mould and drive enough sound to allow the 6-8Khz to be audible. Yes it can be done, but is it a good idea - possibly not for two reasons. Namely: the sensation of your ear being blocked and the subsequent overproduction of LF sounds AND the effect called the downward spread of masking, which prevails over your better hearing in the lower pitches.
Open, smaller receiver, much more open fit and accept you just can’t hit the (probably dead) regions of the cochlear, used with a very little bit of frequency compression to bring some of the higher sounds back into play. You end up with a much kinder fitting to wear every day and perhaps miss an octave of sound that you would be otherwise overloading your entire cochlear with.
Nearly all my fittings go down the #2 route - less is more etc.