34
This post is in reaction to the most recent Legends & Lore post by Monte Cook. This will be a serious test case for my rule against fulmination in this blog - but never mind that, go read Monte's post and come back.
I've run two D&D campaigns up to 13th level (one 3.0, one 4e) and I'm currently playing in AE at 14th level, and I can say with great certainty that the ideas and goals of high-end play don't enter into anyone's thinking when they say that gameplay breaks down at high level. Now, players and/or DMs may not prefer the power level in which they "create their own planes of existence and lay waste to planets," but that's just a matter of taste - the system doesn't push you to tell stories about those kinds of things at high level. Our AE game has discussed the concept of planar travel, but not as anything we'd ever for any reason want to do. It'll probably come up in that context a few more levels, and that's fine. The plot has gone from local to continental; we're solidly in the system's concept of "paragon" tier, if you will.
No, it's not the angels and devils that cause the problems of high-level play. No one would talk about "the game breaking down" because they're just uncomfortable with the change in style. When people say "the game breaks down," what they're talking about is that the rules aren't working. The rules break down for different reasons in each edition; it's not like designers haven't made a valiant attempt at improvement every decade-or-so. Not working may need more definition also: if odd corner-case tactics are your best bet, even though their chance of success is slim, the game is not working as designed, and characters aren't behaving according to any kind of archetypal guideline. Players can still have fun in these kinds of games, but I think everyone at the table is aware that things have gotten cheesy and walk away a little more dissatisfied.
Advanced Dungeons & Dragons, Second Edition
This is the furthest back I can go with any experience of my own. My highest-level 2e campaigns each saw one character reach 9th level. Notable problems include:
Linear fighter, exponential wizard: The game was designed so that wizards would be hard to play at low levels, and would get a payoff for their survival at higher levels. The other side of that is that fighters, who theoretically start off strongest, stop really increasing in power. To be more precise: everyone's hit points hit a soft-cap, while a wizard's damage (and pretty much only a wizard's) keeps scaling up. Otiluke's freezing sphere is a great example of this; it scales up at 1d4+2/caster level, with no cap. By the time they can cast it, that's 12d4+24 damage on a failed save (save vs. spells negates); fail that save and an average damage roll means you're also saving against massive damage.
On the other hand, most of the new spells wizards are picking up at these levels don't deal direct damage; they're either wonderful and/or bizarre new utility effects, or they're save-or-die effects. The funny thing about save-or-die spells in 2e is that as the target becomes higher level, all save-or-die effects are decreasingly effective against it, since the target's saving throw difficulty ignores the potency of the effect in question (unless there's a specific modifier called out in the effect). I want to talk about spell scaling in more detail; see below.
Thieves reach a skill cap at 95%, and there's really no discussion in the core rules of what happens with their skills beyond that point - it's entirely a matter of houserule. This is a place where it was good to start off as a demihuman, but later on it's fine to be human.
But it doesn't matter, because at these levels, nonhuman characters can't advance any further anyway, unless the DM has opted for a rule to let them keep advancing more slowly. It's not really worth getting into how terrible this is; I know the OSR guys think it's fine because "we never played to that high of level anyway," but given that this is a discussion of high-level play, that just supports my argument.
A lot of high-end monsters have Magic Resistance, which in this edition is a flat and basically unmodifiable percentage chance of failure for absolutely any spell cast on that monster. This is necessary because the dragon needs an additional chance to not die instantly on the wizard's turn; on the other hand, there's no solution but to pick the one or two spells that target terrain, and thus get around Magic Resistance completely.
At 21st level and above, things change around a lot more, but I don't really want to re-familiarize myself with Dungeon Master Option: High-Level Campaigns nearly enough to speak about it in detail. I'll just note that the first half of the book is the book on how to be a good DM that the DMG really needed to be. Not all of the advice is really in keeping with best-practices DMing, but it's mostly excellent. Also, it introduces enough new class abilities and such that... yeah, the title of this book should have been AD&D 2.5 Edition. Someday if I feel really motivated, I might give it some deep analysis in this blog.
Dungeons &Dragons, Third Edition
The problems of high-level play in 3.x are considerably better documented than those of 2e, for the simple reason that XP tables were rescaled to make high levels attainable within just a few years of regular play.
The issue of scaling between classes was considerably improved, but I don't know that I think it was solved except within certain narrow bands of playstyles. There's a lot more room to argue about which class is strongest at any given level, though, and that at least is as it should be.
The scaling of higher-level spells is still a big problem, though. As in 2e, spells that deal plain ol' hit point damage become rarer, though meteor swarm is still out there. Again, see below for a more detailed discussion of spell scaling. The function of saving throws is a huge improvement, because not all spells have an equal chance of success against a target.
Death by massive damage becomes a chief threat of the campaign, because 50 hit points of damage (especially in post-power-creep 3.x) becomes pretty trivial and common; when it's less than a quarter of a fighter's hit points, why is he worrying about it? On the other hand, it's not difficult for any critical hit, a rogue's sneak attack, or a polar ray to deal that much damage. The game becomes focused on who rolls a 1 on their massive damage save (since they only fail on a 1 anyway).
Where 2e was really quite stingy with immunities for PCs (at least in its core rules - its splatbooks hide shamefacedly in the corner during this discussion), functional or actual immunities are relatively easy to come by in 3.x - against energy types, death magic, divinations, critical hits and sneak attacks... the list goes on. Immunities are also very common among high-level monsters, particularly immunity to sneak attacks and critical hits. Spells and abilities became problems, so the designers introduced things that simply negated those spells and abilities. This is bad for the game in a whole other list of ways; the game becomes focused on guessing which buffs your target is running, and/or a game of who rolls better on their dispel magic or greater dispelling caster level checks.
Spell resistance is a big improvement on Magic Resistance, but it's still problematic when it gets out of hand. It's another chance for one of those save-or-die spells to fail, so high-level creatures still need spell resistance or a similar mechanic.
A monster's Hit Dice, which at low levels are almost the same as reading its Challenge Rating, balloon out of control. This is necessary to give monsters the hit points, attack bonus, saving throws, or whatever else that the Monster Manual and/or the DM want them to have. (The hit point thing is especially problematic with undead, as they do not have a Con score.) The problem is that a number of rules (Turn Undead, holy word, and so on) are tied to Hit Dice or level rather than CR. Signature abilities that should be fun to throw around are irrelevant against a lot of monsters, but the only way the player could make that kind of judgment ahead of time is to know the monster's stat block - Hit Dice has even less of an in-character marker than CR does.
Ability score damage and drain, and negative levels, enter the field of gameplay and in many cases become common, as Monte notes. This is a major problem, though, because they stop the game for math with effects that cascade throughout the system.
Resurrection from death becomes readily available. This part is fine, even though it doesn't fit all campaigns. On the other hand, there's that level loss; short of taking away a favorite piece of gear, this is the best way I've ever seen the make players leave a session angry. Level loss as a consequence of a fight that the rest of the party goes on to win does some really odd things to the game, and the games I ran scrapped that penalty for death and replaced it with a random-draw system to keep a sense of risk attached to death.
Death effects are more common; the feel of the fantasy has changed so that it's appropriate for a character to die a few times and expect to come back. It's just, you know, how things are done. The problem that I have with this is that death kicks the player out of the game until the rest of the party achieves his resurrection. Don't get me wrong, I don't think death should be removed from the game; I just don't favor high-level play putting the Grim Reaper on speed-dial.
Dungeons & Dragons, Fourth Edition
Once again, they made a lot of improvements. Problem adjusted!
The linear fighter/exponential wizard problem is actually pretty much put to bed in this edition. Congratulations! The only problem is that this homogenized gameplay too much, but that's a totally separate issue.
The main problem that I saw in 4e was tracking all of the different stacking/non-stacking bonuses and penalties to attack and damage. Every die roll was a small conversation in itself; when a lot of attacks include if/then situations that you can't determine until after you're several die rolls in, it becomes a real headache. I don't remember what it was, but the wizard in the game I ran had some truly godawful-to-track feat that dealt +2 damage if he hit more than one target... so I couldn't start recording damage until after he'd made most of his attack rolls. This was only a problem because with AoEs, you always roll damage first and attacks second.
The secondary problem was the huge variety of increasingly complex minor and immediate actions. These are super cool when each character has up to two of them; past that, every character's turn gets slower, because the character with all the interrupts needs a few moments to decide whether or not to use her ability.
The increasingly prolific use of the stunned and dazed conditions by both sides is just frustrating, and mechanics intended to patch that frustration feel obtrusively artificial - and they're just more numbers to track.
Tracking auras is fine when there's one on the field, but when monsters meant to be used in multiples are given auras, the DM has his work cut out for him.
It's not a primary concern by any means, but PCs can pile up a lot of minor-to-middling magic items that do weird things the DM needs to keep in mind. This is a problem in any game where the PCs accumulate Stuff, but 4e's magic items are especially specific and obscure in their applications, so it's more of a problem here.
The Spell Scaling Problem
As advertised, I want to explain what I see as the central problem of 2e and 3.x, and any game with a certain approach to scaling.
A first-level wizard can cast a first-level spell to moderately injure a level-appropriate opponent (magic missile).
A third-level wizard can injure a level-appropriate opponent a little bit more, or completely freaking murder a first-level opponent (I'm thinking of scorching ray here).
A fifth-level wizard can injure a large group of level-appropriate opponents, or completely freaking murder a large group of first-level opponents (fireball, lightning bolt). (Rarely used side note: hold person is great for single-target murder at this level.)
A seventh-level wizard can inconvenience a large number of level-appropriate opponents (ice storm, wall of fire, whatevah) or slay outright one level-appropriate opponent who fails two saves. (phantasmal killer).
A ninth-level wizard can significantly injure a large group of level-appropriate opponents (cone of cold), instantly kill a huge swath of low-level opponents (cloudkill), or slay outright one level-appropriate opponent who fails one save (hold monster - okay, maybe technically two, depending on how the coup de grace works out for you). Baleful polymorph also gets a mention here - while they're not dead, a beached carp is not much of a threat, and you can whittle down their unaltered hit point total at your leisure.
An eleventh-level wizard can kill a level-appropriate target with one failed save (flesh to stone), and instantly kill a potentially very large number of almost-level-appropriate targets (circle of death). Note that we've left hit point damage behind. I regard this as bad, because a fighter and a wizard no longer really contribute to the other's attempt to kill a target. The wizard could keep slinging big evocations, of course, and these will continue to be a big problem for level-appropriate targets, but never really enough to kill an uninjured opponent outright.
A thirteenth-level wizard can kill a level-appropriate target in a variety of unusual ways, mostly based on spells specific to the creature's Type (banishment, mass hold person) or... kind of whatever. Reverse gravity is always good for a laugh. So is finger of death - it's like disintegrate, but without that pesky ranged touch attack.
Wizards of fifteenth and seventeenth levels do commensurately more awful things to their opponents. Almost nothing in the ninth-level spell list deals hit point damage, other than meteor swarm,Bigby's crushing hand, and sometimes prismatic wall.
The tl;dr version is that wizards are only mildly useful against opponents at the start of play, but they had to improve on that relative status eight times. By the third step of improvement (fourth-level spells) things are starting to get out of hand, but at least you get multiple saves. After that point, though, the game gradually pares down the target's possible defenses, because how else do you improve upon "one spell, one kill"?
Monster abilities scale upward to keep up with the things wizards can do. Maybe you see that the other way around, but I think monsters could stay about the same and just use bigger numbers if all they had to worry about were characters attacking their hit point totals.
And that is how, and why, D&D (and every other game) breaks down at high level. Judging by other games, we should just be glad D&D takes as many levels as it does to have these problems. The solution to all of this, as I see it, is to completely change the game's concept of spell levels six and up, if you're going with 3.x-style Vancian magic.