Well, the multiple-characters-at-once thing was often very location-specific, and was a habit picked up because first to third level OD&D characters (and in a different way, early RQ characters) were so damn brittle.
The rotating pool thing seemed a lot more common though, but this was also in environments where people moved characters from game to game all the time. If someone had a really tight closed campaign it probably wouldn't have happened, but I wasn't seeing a lot of that in 1976.
Perhaps it's my sensibility from being an "older gamer" but in my most recent 5e campaign, with three players new to the game, during Session Zero I asked each of them to run two PCs. I did so because:
1. We were starting at first-level and I know first-level PCs have low hit points and are very subject to swings of the dice - the DM can spoonfeed easy encounters but if you put PCs in truly dangerous situations, it is quite likely one or more will die at low levels.
2. Three PCs (instead of 4) means I either have to
really put on the kiddie gloves at lower levels or accept that there is a significant chance of death.
3. My DMing style is both something of a sandbox (and I tell my players outright "I won't tailor encounters to you; you could easily get into something over your head at low level, retreat should always be considered an option") and has old-school sensibilities (longer, resource-draining dungeon excursions between trips to a safe haven so short and long rests will be few and far between) which I knew these players would not know at first. So expected there to be some "bad decisions" early that led to death (though to be fair, I always tried to telegraph "this is not a good idea").
4. I am also a "if the result of a roll would be obvious to the characters, the roll is made in the open" type of DM (combat rolls, for instance, are always made in front of the characters; something like an Insight/Deception check will not be)... meaning I can't/won't fudge rolls to keep PCs alive.
Given all of that, I expected (1) there would be lethality and (2) the party would need more than three characters to succeed. Having each PC double up solved both 1 (you're not out of the game when one of your characters drops) and 2 (a party of six survives much more easily than a party of three). In my experience, when players know there is real risk involved the rewards for success are sweeter - you are acutely aware you actually COULD have failed and weren't just brought to victory by DM fiat/storytelling. And I articulated all of this philosophy in Session Zero.
Yes, all three of the players mostly inhabited a "primary" character and had a "secondary" character and yes, there were a few deaths along the way (one of the players still has a fear of gelatinous cubes), but after a couple of levels, the swinginess of low-level combat went away, the players got better at reading situations, and got better at "equally inhabiting" both characters.
Not for every situation, but I think in this case it was definitely the right call, and it was a call I could foresee before the campaign started. These players won't always want to play multiple characters, but I think for small groups doing old-school-type dungeon crawls at low levels, I would definitely recommend.
