3e has quite a few mixed-success stories — great ideas that didn’t quite work out as intended, and tend to feed into the system’s big problems. The best example of which is probably 3e’s buffet-style multiclassing.
Buffet-Style Multiclassing: As I mentioned in The Love, 3e’s multiclassing rules are markedly better than the strange subsystems that preceded it. But they’re also weighed down by inconsistent class design and ill-conceived XP penalties.
Level Adjustments: LAs are 3e’s attempt at making monstrous races playable. The problem is that the designers either didn’t quite now how to assign LAs, or they really didn’t want players playing races from the monster books. In either case, the result is that LAs are usually much too high, especially considering the racial hit dice that monstrous races are often saddled with. In other words, most monstrous PCs are glass-jawed buffoons.
Iterative Attacks: IAs are 3e’s evolution of the 2e fighter’s multiple attacks per turn. (Anyone else remember wondering whether the fighter got his two attacks on even rounds, or on odd rounds?) Overall, IAs are probably an improvement — but they come with their own problems. The first problem is a matter of tracking — ever had to roll four different attacks, each with a different modifier, and often with various circumstantial modifiers on top of everything else? If you have, you know how much of a chore it is even with color-coded dice.
The second problem is that to use your iterative attacks, you can’t do anything more than shuffle sluggishly through the battlefield between sword swings. (Except if you happen to be a lion totem barbarian.) Not only does this create a boring combat dynamic for characters and monsters who rely on iterative attacks, it also creates a very bizarre and counterintuitive side effect. The side effect is this: If your BAB is +6 or higher, your best first move against any melee opponent is to…not move. Instead, wisdom dictates that you wait for your enemy to charge and swing with its one attack, and then retaliate with your full attack routine. (Except if you have some kind of ubercharger build.) Which makes no fucking sense.
Spontaneous Cure Spells: This is D&D’s first attempt to alleviate its traditional healbot phenomenon. And it does help; no more filling every spell slot with a cure spell “…just in case!” Except of course if you’re anything but a good-aligned cleric. While spontaneous curing does make the cleric class more appealing, it doesn’t eliminate the healbot phenomenon; it still reinforces the expectation that clerics exist to heal, heal, heal! And shame on Joe for wanting to command the enemy rather than heal Bob’s barbarian.
Skills: 3e skills are a definite improvement on 2e’s combination of thief percentage-based skills and ability-based nonweapon proficiencies. Unfortunately they’re new player-unfriendly, they’re mired in system mastery, they pigeonhole characters into cliche tropes, they create auto-failure/success at later levels, and they all too often create the sensation that one is filling out a tax form rather than playing an enjoyable game.
Essentially, the skills chapter is a microcosm of everything that’s fiddly, opaque, and frustrating about 3e.
Feats: Feats are one of 3e’s greatest innovations. Unfortunately, their wildly inconsistent quality holds them back. In fact, Monte Cook actually used Toughness to define the problem of system mastery that plagues 3e. Some feats grant minor bonuses, such as Weapon Focus and Spell Penetration. Others grant dramatic new benefits, such as Leadership and Natural Spell. Still others grant benefits that the game assumes the PCs have, such as Weapon Finesse and Power Attack.
Wealth-by-Level: WBL guidelines are good because they essentially light up a big neon sign in the DMG that reads YES, D&D EXPECTS PCS TO GET LOTS OF STUFF. Unfortunately, this advice doesn’t tell new DMs what particular things the game expects PCs to have, and the gold-piece values themselves are based on the very swingy random treasure tables.
What the DMG should advise DMs of is the Big Six which D&D expects the PCs to have, because without the bonuses they grant, the game goes sideways as the PCs gain levels. Even better would be a chart of bonuses-by-level guidelines, to give DMs an idea of when the PCs are expected to have masterwork armor, and when they’re assumed to have +5 armor and two other +5 AC boosters.
As an added bonus, a bonuses-by-level chart makes it easy to grant those bonuses as innate bonuses so that magical items can be more special. Innate bonuses also keep things from going pear-shaped every time someone fires off a dispel magic.
Magical Item Crafting: Most days, I believe that standardized crafting rules are an improvement on 2e’s vague DMG advice, which includes turning every craft attempt into a fetch quest. It’s just a shame that 3e crafting rules suck.
Basically, 3e magical item crafting fails because it costs XP. Some players hold onto XP like a miser hangs onto gold, but the fact is that the XP rules turn this cost into an insignificant inconvenience. The lower level the crafter is, the more XP he earns, the faster he levels up again. Hell, he might even get lucky and sling-shot past his miserly allies due to this quirk!
And with the gp-cost of item crafting being 50% of market value, savvy crafters can outfit themselves and their allies far beyond what the game or their DM expects.