It’s not really about what ”works best” for the average recipient, because these emails are not sent out of the goodness of the senders’ hearts. If data show that fancy HTML emails translate into more paying customers, then that’s what drives these decisions.
It’s of course possible that this is not the case, however to be convinced of that I’d like to see the stats rather than speculation based on individual people’s subjective experience. And naturally it can depend on the product and target demographics, too.