Specifically, I think the "Javascript the Good Parts" approach applies to Vim as well : if 90% of the vim shortcuts I use only 5% of the time, it's both not worth my time to invest in learning them, it's also making invest less time really mastering the best shortcuts.
So I feel like Crockford's mindset in that JS book totally applies to vim: if you can do some motion with two keys instead of one, but using a smaller set of motions that you get to know REALLY well, then trying to remember the extra shortcuts just so you can save a key is not a good investment - on the other hand it is a great way to continue to mythologize some kind of "super efficient" developer/writer whose work somehow is made better by typing faster.
Vim is complex enough as it is. I'd love to see a really opiniated "Vim : The Good Parts¨ manual - where the author collects the most useful motions, and bluntly says "these are NOT worth learning".