Miller sees the emergence of the RPG format as the creative addition of open system world building to the closed systems of hobbyist wargaming. Role playing proper appeared when wargamers started using the relatively inflexible rule sets of wargames as the basis for collaboratively imagining adventures in fantasy settings.
I've already seen some pushback to Miller's account, and honestly, I'm not versed enough in the history of the scene to adjudicate, but I find it striking that the theoretical division he presents — open vs. closed — maps pretty well to the way system and setting are discussed in the OSR scene. Miller writes:
I can throw down the GURPS, D6 System, Savage Worlds or any other generic system and that will be inadequate. Without the setting in which the game takes place, it’s just rules and die rolls.
But we need not think purely in terms of generic systems. Discussion about taking the system from one game and applying it to the setting from another is common. Free Kriegsspiel Revolution goes a step further, demonstrating to what extent the open system component of Miller's equation can be made to stand on its own.
The implication, whether we like it or not, is that the units we think of as role playing games are chimera, patched together from parts which, more often than not, have no integral connection to one another. The phenomenon we called a role playing game may, in fact, be something occurring at the intersection of other games.
Maybe Miller's account is wrong. I've already seen it argued that early iterations of Dungeons & Dragons hewed close to a simulationist model — wargaming with goblins, in effect — and that may be so. At the very least, it's possible to imagine a fantasy dungeon crawl game constrained by a much more tightly closed system than those used in modern OSR/NSR games, and I'm sure some people would both play and enjoy that. But would we call it a role playing game? If not, then we may still have to reckon with the hybrid nature of the beast.