Crazy Jerome
First Post
I think of this issue as one of immersiveness -- not in Monte's slightly strange use of the term, but in the sense of asking how much of the player's time is spent thinking about what is happening inside the fiction as opposed to thinking about the rules mechanics. The more the players have to step "outside the game", either to justify the mechanics or simply to select their next action, the less the players are thinking about the fiction itself.
Of course, this isn't the same thing as realism, but it comes from a similar motivation. The primary issue with a realism failure is that it yanks the players out of the fiction -- either because the mechanic is too complicated or because the results/process is unbelievable.
The inherent problem here though is that "immersiveness" and "hit points" only go together as a learned skill. That is, "hit points" themselves are anti-immersive. But immersiveness is not mere simulation, but rather overcoming the need to step outside the rules. But it also depends upon a rule that can "fade into the background"--which presumes a certain limit on handling times or the type of mechanics.
Rolling d20+mod vs DC is counter to immersion. Rolling percentage dice on a roll under versus threshold is also counter to immersion. If you naturally think more in percentages (regardless of how "accurate" your thoughts are), then the second system, all else being equal, will be more immersive to you--and likely, feel more "realistic" as a bonus. OTOH, if you get sufficient practice with the d20+mod vs DC option and not the percentage version, the way the former fades into the background will gradually overcome this natural preference.
So I think the game is going to step between the players and the simulation, from an immersive standpoint. The question becomes about payoff down the road. If hit points can fade into the background and thus become a mechanic that intrudes not at all, then those first few times when they stepped in between the players and the simulation is a good deal. And compared to a lot of mechanical options for recording wounds, the evidence is that they can for many people. For other people, not so much.
Moreover, in practice some substantial subset of players are going to get that "realism failure," one way or the other, from a given mechanic. Even simple mechanics can be too complicated for some. And any mechanic that tries to model both process and result is bound to be overly complicated--and likely fail at both to boot--again, at least for some people. It is not an achievable goal to make mechanics that will seldom have such failures in a single, small group--let alone the wider audience of the game.