I thought I would expand on an almost incidental idea in Falling off the RNG.
I think D&D 3.x has too many armor class values. I believe this is mostly to model certain types of play and to encourage player characters (and other monsters) to act in specific ways. I think this is somewhat heavy-handed and while it does work after a fashion, is awkward.
Touch Attacks in D&D 3.x
The more I think about it, the more I believe I have found a better option than touch attacks.
Touch attacks are largely a mechanism to give poor-BAB characters an opportunity to hit high-AC targets, by removing the benefits of armor. This chafes me something fierce; it is bad enough that heavy armor has no better Armor Class than light armor and high Dexterity, despite costing money, armor proficiency (cheap, true, lots of classes grant it at first level), Armor Check Penalty, and movement speed, without having it easily bypassed entirely. Note that as a weapon quality brilliant energy is a +4 equivalent weapon quality that isn’t as good as touch attacks (natural armor still applies against brilliant energy weapons, so another armor class value that admittedly doesn’t come up much).
I think there is a better and more equitable way of handling this.
I am aware of three places touch attacks are used: spell attacks, incorporeal monsters, and grappling.
Spell Attacks
Spell attacks that do not target saving throws mostly use touch attacks. Instead of having them bypass armor entirely, grant a bonus to the attack equal to the level of the spell used. Rather than totally negating armor bonus and shield bonus (and any enhancements on armor and shield), and natural armor bonus, a wizard attacking with his highest-level spells has an attack bonus roughly on par with a fighter of the same level (wizard BAB + highest spell level is the same as a fighter’s BAB).
The fighter probably has higher Strength than the Wizard has Dexterity, can have more Weapon Focus, and can always have his highest attack bonus once a round (plus lower-bonus attacks if he gets to use his iterative attacks). On the other hand, the wizard has the option of bending reality on a whim and the fighter doesn’t. The fighter should be allowed to be better at something.
Also, in most cases the spells that use touch attacks seem like they could reasonably be affected by armor. For instance, acid arrow is a touch attack that armor should defend against. Armor applies against a thrown flask of acid — the acid arrow gets +2 to hit that the flask doesn’t, so there is still specific benefit beyond not having to carry around corrosive materials that could cause damage when the flask gets broken in your pouch — so it seems reasonable armor would apply here.
If you want, this also leaves room for implements to improve the spell attack bonus. A frost wand could quite reasonably add to the spell attack bonus when using spells with the (Cold) descriptor, for example.
Incorporeal Monsters
Incorporeal monsters in the Revised System Reference Document (RSRD) are expected to be undead. I expand this in my campaign to include certain elementals and the like, but that’s not terribly important.
What if we treat armor in a manner consistent with weapons? Just as mundane weapons cannot hit incorporeal creatures, and magic weapons can only hit half the time, have mundane armor useless against incorporeal creatures and let magical armor work half the time. I think natural armor would continue to work normally, just for simplicity (as living tissue it is naturally resistant to intrusion by incorporeal spirits, or something).
Incorporeal monsters are still pretty scary against people without magic armor and/or magic weapons, and worrisome for those that do, without having CR 3 monsters (allip) able to render high-level armor and shield and amulet totally useless. It also makes ghost touch a reasonable +1 equivalent armor quality instead of an overpriced +3 equivalent armor quality that might need to be put on multiple pieces of equipment — putting it on roughly on par with the ghost touch weapon quality.
I’m not entirely happy with having to determine on each attack whether the armor bonus applies or not. I considered just giving incorporeal creatures a better attack bonus (make them the same as the good attack bonus, because they are harder to defend against). In the end I decided that treating weapons and armor consistently, and the thematic appropriateness of being unable to directly affect the incorporeal creatures, were worth the effort.
Grappling
Grappling as a whole needs to be reworked from the rules presented in the RSRD. I see no reason to use a complex and often ignored rules subsystem as a reason to not make this change.
I’ll be looking at a few places (Pathfinder rules address it, for one, and I can think of a few other places). Failing all else, just having Armor Check Penalty apply to your grappling check might be enough — you lose some flexibility when wearing armor, but I don’t know that you’re all that much easier to grab when wearing a breastplate than you are when you are not wearing a breastplate.
Flatfooted Condition in D&D 3.x
I’m not really fond of the flatfooted rules either. They don’t seem to come up often enough to make them worth the effort because the people who would be significantly affected by them (i.e. ‘have really high Dexterity bonuses’) are the least likely to be affected by them.
The ‘combat advantage’ mechanism from D&D 4e looks like a much more workable mechanism here. Between the bonus to hit when you have advantage and the opportunity to make good use of it (combat options — offensive or defensive — that work only when you have advantage) I really can’t argue that flatfooted armor class is worth keeping any more.
Closing Comments
I think the game can be reasonably reduced to a single Armor Class value. Touch attacks can be replaced with something that better scales with level (getting past the differences in base attack bonuses between classes, at least some of the time) or limited to uncommon and thematically appropriate monsters. Flatfooted attacks can simply go away in favor of a more flexible and applicable mechanism.