scientiae: For instance, hitpoints are, realistically, the only effective way to introduce random outcomes (otherwise you would have predetermined outcomes, as for
Amber) but they are also very silly at high levels. (As any victor will admit, candidly, luck plays a factor in all success —— martial or otherwise.) Instead, perhaps, a better process would be to increase damage resistance, rather than total hitpoints, for higher skill / more experienced player characters. After all, ever human has a very similar capacity to absorb damage, but better experience in martial pursuits grants some people the ability to avoid the worst damage in an encounter. (Certainly add some small amount to the hitpoint total: this seems trivially sensible; but multiples of a base die implies higher level characters have geometric increases in fighting skills, which seems less sensible.)
You could have random outcomes without hit points; just make each attack have a random chance of instant kill. D&D feels that way at low levels (except maybe 4e, because I hear you start with more HP but get it more slowly), but it really isn't fun gameplay when attacks keep missing long enough for you to get unlucky and killed in one hit. (And then, of course, Baldur's Gate doesn't implement the death's door rule, which would otherwise give you a chance of survival in that situation, so level 1 gameplay is even worse than it hould be in that game.)
A better way is just to have damage and healing effects scale nicely along with hit points. Perhaps have attacks between equally matched opponents take away similar portions of hit points throughout the game, or perhaps make attacks take a larger portion of player HP away in order to sustain the game's difficulty. (One common problem with CRPGs and other games with growth systems is that often the game fails to sustain its difficulty in the later stages, as the player gets stronger faster than the enemies do.)
Another thing: If it's possible to reduce damage received from physical attacks, armor really should have that function; it feels rather ridiculous that heavily armored characters are harder to hit in D&D (and many CRPGs that trace their routes back to D&D; the earliest CRPG I know of that broke away from that is the original Dragon Quest, which is actually very non-D&D-like in its mechanics). (Of course, the reverse situation, heavy armor hurting evasion, is easy to get wrong; Final Fantasy 2 is a big offender here, to the point where heavy armor can make you lose a fight that you could have othewise won without taking any damage.)
scientiae: For instance, the
Shadowrun methodology heals the last wound when a fight ends. If hitpoints are some sort of measure of the fatigue and duress of a particular encounter, then this seems a good compromise, also considering there exist medical items that heal a set amount (varying by the skill of the character who applies it).
I don't like this sort of thing. In particular:
* It becomes necessary to keep track of the last wound received in addition to keeping track of HP.
* It means that a hit for 1 point of damage can result in tons of HP lost (which translates into loss of healing resources) after the battle, which can be rather frustrating.
* Also, consider that a mechanic like this favors multi-hit attacks. (It's not the only mechanic I've seen that does this; damage limits (where a single attack is capped at 9999 points of damage (in a game where bosses have more than that amount)) are another example.)
scientiae: You make a good case for a separate domain for the 4th edition. I see why they did what they did, since all the corrections are aimed at weaknesses in the game (opportunities for players to metagame instead of roleplay) and some of the solutions [attempt to] fix some of what I found irritating in the earlier releases.
Funny thing, just the other day I was talking to someone who has played table top RPGs and she doesn't like 4e because it's too mechanics heavy.
(There's also the fact that, what some call "metagaming", I see as just playing the game, and finding strategies that work well.)
