Optimization and Tiers: The Tier System Expanded.

[bekeleven]

The tier system holds up across equal levels of optimization, and until someone makes a comprehensive list of all possible degrees of optimization, that's the best we can do.

Introduction
The existing Tier system doesn’t take into account what happens when Emperor Tippy builds a wizard in the same party as a 13 year-old builds a fighter. In fact, it doesn’t even take into account what happens when CaptnQ builds a monk in the same party as the 13 year-old builds a wizard. Perhaps less so. The mantra repeated on these forums is a simple one:

Player > Build > Class

Which make it hard to catalog the power of a class. JaronK’s Tier System nonetheless did so, and most agree (at least partially) that he was onto something. It states that Tippy’s Wizard is more powerful than his fighter, and CaptnQ’s Artificer can solve more problems than his warblade. It states that the 13 year-old, new to the game, will build a psion more powerful than his binder.

Not so, say I.

I don’t know about you, but if I built a binder at age 13, I would’ve tried a bunch of vestiges until I finally found one that seemed to work for me, then use that all of the time. And when I leveled up, I’d try one or two of the unlocked vestiges, under the assumption that they would be more powerful than the ones I’d left behind. Sure, I might be taking nonproficiency penalties for that magical greataxe I found last dungeon, but by chance I’d eventually cobble together a working build capable of solving problems.
My Psion, on the other hand? Blasting spells all. Day. I probably wouldn’t even use more than one or two of them (namely, the highest level ones). In other words, my binder is probably better.
JaronK made the point that the power of tiers can be relative to optimization. In other words, Biffoniacus_Furiou’s commoner would be equal to young me’s Warblade: “Also note that with enough optimization, it's generally possible to go up a tier, and if played poorly you can easily drop a few tiers, but this is a general averaging.”
My position is that this view is incomplete. A wizard played by someone just starting the game wouldn’t be able to solve every situation as easily as a druid – druids are near impossible to screw up in any capacity, both build-wise and play-wise. The bare minimum someone will play a druid past level 5 is still leagues ahead of most other classes, even tier 1 classes. Below, you’ll see that I rate Druid as a (1-1-1): The only class that remains tier 1 through all 3 optimization grades. Yes, of course Eggynack’s Druid would be more powerful than mine were a decade ago. I still blundered through my similarly masterful friends’ campaigns and made a beautiful mess of things.
Below, I enumerate each grade as well as the grades above and below my range. Please be familiar with JaronK’s Tier System before you continue, as this is a derivative work.

The Five Levels of Optimization
These are, by necessity, subjective. I use my own experience and opinions to create them. You can define your own if you like, but this is what I used to create the table below. I didn’t base these entirely on balance decisions. Most of the division is based on system mastery, especially at the low end, and the remaining is mostly based on what I would build and bring to a table in good faith.

To reiterate:

Player > Build > Class.

If the player plays poorly, nothing will save him.

What Separates the Grades
I primarily used the following factors to divide specific builds and optimization tricks into their grades.

1. System Mastery. This is the most important one. It's the only one that affects play (and player > etc.), and also informs many factors. As a player reads the books more (or goes on the internet in search of advice), they will necessarily tire of the material they have access to (or see more recommended). Conversely, if a player gets a new book, they will generally either read it and become more enfranchised, or skim it and take a few tricks, but not necessarily use them to their fullest. In other words, more material access generally correlates to knowing how to use stuff, and since the entire system is placing things in general buckets, I'm running with it.
   
   • As stated, Book Access is a consideration. A DM can ban books from the game, which minimizes the effects of book access.
   
   
2. "Spirit Of The Game." This is a rough one to explain, but basically: Grade C is where the designers intended the game take place. Midlevel Grade C D&D feels like sword and sorcery fantasy. It feels like lord of the rings. Eventually, someone playing the game read the trap rules and asked, "Can't I make a create food trap?" Then they solved world hunger. "Can't I create a teleportation circle network?" Then they solved travel. "Can't I create Mage's Lucubration..." and there go limited spell slots. This person didn't know it yet, but they were on the road to the tippyverse. The first and most fundamental cornerstones of the Tippyverse aren't things that require high system mastery to do, merely the ability, and willingness, to redefine the genre of Dungeons and Dragons for yourself and your playgroup. And the more you move from "Intended Play" to consequential mechanical use, the more you tend to rise in grade.
   
   I don't say this as a judgement call. Both Grade C and Grade A games have a lot going for them, and I've played in both. Both can be fun. I think most tend to play in grade B, not just because they can't optimize higher, but (also or instead) for other reasons - It's easier for the DM to build encounters, it's easier for the DM to build a world closer to fantasy convention, and because most people get into D&D because they like sword and sorcery fantasy.

S: Theoretical Optimization (TO)
"Stay within the rules."

This grade of optimization consists of anything that I would never bring to a table, or even suggest at a table, unless I knew the GM was aware of how it worked and made an informed decision to allow it. Here is a very incomplete list of the types of things I consider TO:

• Ancestral Relic Custom Runestaves
• Drown Healing
• Most epic skill use
• Diplomacy optimization, basically at all
• Item Familiar (including XP boost)
• Dark Chaos Feat Shuffle (Bonus feats/Otyugh Hole, not just retraining)
• Infinite Loops/Destroying the World
• Cost-free metamagic reduction
• Starting games with a dozen partially-charged wands
• Anything involving free wishes (Ice Assassin, Shapechange into Zodar, Gate, etc.)
• Stacking Nightsticks, Stacking Theurges, and other “questionable” interpretations
• Most things involving Kobolds

I don’t tier things in TO. TO is useless to grade. First, because at and around this level of brokenness, you can build every class as a kobold with cross-class ranks in Knowledge (Religion). The class isn’t as important as the ability to hide from GM smacks. Second, I would never build these and try to avoid playing near them as well, so why would I care what tier your Incantatrix hits?

There are different levels of build within this optimization grade. But it's very hard to find a dividing line within the group between Pun-Pun and the rest. Simply, there are no meaningful numbers if you're so close to infinity.

A: Top Practical Optimization (Top PO)
"Win the game without breaking it."

Top PO is the grade of character that I build when I’m told to optimize hard and I know the other players have decent system mastery. I almost never build full casters at this tier; I prefer to tweak a martial build all to hell instead. Top PO takes advantage of a lot of tricks, but the tiers are actually visible because you’re somewhat limited by your chassis and class features.
Here’s an idea of what I mean by top PO:

• Custom skill-increasing items
• Item Familiars (minus XP boost)
• Metamagic Reduction (besides DMM abuse, Incantatrix’s Metamagic Effect, and similar)
• Enhance Wild Shape, Polymorph, and some Shapechange
• A Knowstone or two or 46
• 7th+ level Chameleon spell slots
• A stat above level+25, particularly mental stats

As you can see, most of the things on my watchlists are spells or spellcaster-related. I’ll add: Duh. Of course that’s what you watch.
Games can be a bit silly at this grade, and they certainly hit rocket tag much faster than lower op levels, but nobody is explicitly god and nobody needs a PhD to compete. To me, this type of game is good in small doses.

B: Standard PO
"Be good the game."

This is the grade to which I optimize a character if I’m not explicitly told to go above it. This level of optimization isn’t that hard to achieve, as it only requires basic system mastery and the desire to build with synergy.

This tier is likely the widest. I’d say I jumped from grade C to B sometime around when I joined the Wizards forums. Nowadays you can point to the moment a person reads JaronK’s Tier System for the first time as the time they cross into PO range. The thing is, I generally build PO characters whether I play with experts or noobs, varying only the base class I start with. If I play with a party of noobs, I’ll build a Superman Monk or Two-Weapon Fighter. If I build for a game on these boards I’ll make a factotum or Wildshape Ranger.

Here are sample shenanigans you can find in PO:

• Most Uberchargers
• Intelligent Divination use (basic Scry + Die)
• Stockpiling Scrolls
• More than one ACF on a character
• More than one prestige class on a character
• Use of X Stat to Y Bonus
• Dragonborn Warforged

Grade B covers the majority of games I see, but its lack of real milestones prevents me from easily breaking it down further. You can say some builds are B- and others are B+, but that gets beyond the scope of this guide.
C: Newbie Op
"Trust in the Designers."

I built grade C characters from 2003 until about 2007, when I signed up for the wizards boards out of curiosity.
Newbie Op is the optimization level of the Vow of Poverty Monk. This isn’t to say that all Newbie-Op characters are that bad, but rather that all newbie op characters trend that far downwards from their power level in other optimization grades. A Newbie Op Druid is about 4 tiers more powerful than said monk, after all. Newbie Op is built around the assumption that the makers of the game carefully balanced all character options against each other, meaning that building the character you think is the coolest won’t put you behind others – and may put you ahead. Here are some optimizations I tried during my first years with the game:

• Dual bastard sword fighter
• Evocation-spec Sorcerer, banning illusion and abjuration
• Vow of Povery Fighter (TWENTY NINE FEATS! @( FEATS!)
• Solving every problem with silent image (not as a shadowcraft mage)
• A wizard with only his 2/level free spells
• Healbot Clerics
• Never summoning familiars, because they’re a liability
• The same, but with animal companions, because I didn’t read that part very closely
• Playing a Kobold (because they’re funny!)

The games I played in 2003 were D&D, and in some ways they were more fun than games I play now. Don’t knock lack of system mastery! Sometimes the guy playing the monk got depressed because he couldn’t hit the gnoll and didn’t know why, but in general, I got together with friends, and we all hung out and had fun. Isn’t that what D&D is all about?

In one sense, this guide is self-defeating. If anyone legitimately playing at this level reads this guide and JaronK’s Tier System, the process will likely move them past it. It can be hard to put that Djinn back into a bottle.

One thing all of my early D&D groups had in common was a focus on damage. You win a fight when the enemy's health hits 0. If you fireball the enemy, you are that much closer to winning. Why would I use solid fog, when it doesn't help me win? I don't know if this attitude was universal, but it was certainly rather widespread at the time.

Arguably, even grade C can be higher-optimized than the designers intended, due to the extremely limited nature of playtesting. Druids were so powerful because the original Vadania playtester focused more on a Scimitar throwing build than wild shape, so the designers never realized how potent being a bear riding a bear summoning bears could be. Those so inclined could refer to this category as C- or D grade, as it's a bit under my assumed newbie-op level.

Still, I think that by 2003 and the release of 3.5, even WotC realized the power of druids, at least to the extent that they could do more than prepare Summon Nature's Ally as one of 3 seventh-level spells, buy a +2 Keen Throwing Returning Scimitar, and spend feats on Point Blank Shot, Far Shot, Improved Critical (Scimitar), and Weapon Focus (Scimitar). And Track. Scribe Scroll was all right.

F: Deliberate Failure

Frankly, a Wizard can suck even more than a Fighter could ever dream of sucking. A Fighter can stab himself to death, but only a Wizard could Plane Shift to some horrible far realm to be tortured for an eternity of insanity.

If you’re building a character in bad faith, you lower your power level without limit. I don’t address that in this guide, as it’s both pointless and futile.

Mixing Grades
I don't recommend mixing disparate optimization grades in a campaign, but if you must, it would be good to know how they relate. As a very rough guide, Grade A = Grade B+2 = Grade C+4. This works better on lower-versatility characters.

For instance, an ubercharging barbarian is frequently a Grade B Tier 4, because it can do one thing really well, but that thing doesn't always help. In many Grade-C campaigns, that thing will help more often, since enemies that are grade C won't make use of inclines, rubble, concealment, and other common anti-charge tricks. In addition, the lower levels of ambient optimization mean he's 1shotting a much larger subset of enemies. He has now (approximately) moved from tier 4 to tier 2, equal with a wizard - fewer options, but higher numbers. The wizard uses his utility spells (fly and invisibility) to God his way around encounters on bad terrain or with spread-out enemies, and the barbarian slices and dices the rest. A rogue with optimized UMD and solid spells and scrolls would be able to keep up with or outshine these party members (Grade A Tier 5), but be inherently more limited and easy to break, since she's sort of still a tier 5 class, just with all the dials turned past the max.

There are plenty of inherent issues bringing higher-grade characters into lower-grade campaigns, even if the grade isn't S. Basically, a lot of optimization takes some groundwork to understand. I built a diplomancer fooling around once around '06 and ended up with +20 to diplomancy at level 2. In our first session, I met my party members in a tavern, then said, "Hold on, I'll get us some drinks" and asked for a round on the house.

What followed was 15 minutes of argument culminating in me making a new character. My Grade C group wasn't ready for a Grade A (or higher) optimization, even stapled to a Bard useless in combat. So diplomacy got essentially houseruled out.*With his rules niche bulldozed, he was homeless. That which can be asserted without system mastery can be dismissed without system mastery. Using a delay death loop to build the Omniscifier would result in a similar houserule, if used in one of my games around the same time.

A partial solution is to avoid large numbers. Numbers are incredibly easy to compare, making bigger ones boogiemen. Instead of the Mailman, build a God Wizard. The incremental action economy advantage of Haste is harder to see - and if it is seen, it will be in the form of your party members contributing more. That was Treantmonk's thesis when he made it (well, one of them).

The point of most games of D&D is not to win. In most cases, playing to the optimization grade of your table will lead to a better experience for all. Build Pun-Pun if you must get it out of your system, but come with a real character after you've proven that you can "win."
*With the benefit of hindsight, I support this move. Diplomacy is the nuclear weapon of D&D

[bekeleven]

Case Studies
Barbarian: Barbarians are remarkably unchanged as you optimize. Mainly, you start with a higher-strength race and your power attack returns get larger. The build mostly scales down with the power level of the group, although towards the low end it may lose the ability to reliably hit/kill encounters. But hey, so does everybody else. Barbarian maintains a constant tier, although they're low in tier 4 in grade A.

Bard: A high-op bard is adding a billion D6 to everyone’s damage dice, but it took me forever to see that. My first thought on seeing Bard: I have the chassis of rogue but without sneak attack. I have spells like a wizard if the wizard sucked at his job. And my unique class feature requires my standard action every round to give everyone a bonus that only hits +4 at level twenty? Jack of all trades, master at jack.

Cleric: Grade A and B clerics are CoDZillas. Grade C clerics are healbots/buffbots who occasionally venture into blasting spells. In my experience, cleric felt bad to play; A group I DMed in 2006 rolled to see who would take the “cleric bullet.” However, as I look back, the buff spells did a lot to help the group – even if I didn’t see it at the time.

Druid: Once you get past the “It’s a cleric with some useless class features in place of the good healing spells” you begin to see the power of druids. Even if wild shape is limited to wolves (or maybe a bear) it still boosts stats and grants attacks. With fewer blasting spells on the list a druid player will be forced to get creative, and may stumble on some of the real power pieces available (yes, I am trying to spin that as a positive for some reason). Finally, although they’ll likely make multiple mistakes with the numbers (unless the GM is more experienced), a noob’s companion can’t really be screwed up – even an advanced riding dog can dish out damage at low op-levels. And if my experience is any indication, the newer the group, the more leeway the GM gives when controlling the companion.

Fighter: My very first character was a level 1 human fighter. I looked at toughness and saw that it granted 3 HP and went, “1 HP is worth 33% of a feat.” Then I saw Endurance and Diehard and said, “For two feats I can get 10 HP. I’m ahead of the curve!” This is a 100% true story.

A Grade-C fighter takes diehard. A Grade-B fighter takes Zhentarim, Imperious Command and Never Outnumbered. A Grade A fighter shuffles his feats and springboards into a prestige class. There are worse classes.

Monk: I’ve written about monks before. If I end up in an S-grade game, I like to build a monk and use subdermal plating and other dragon mag tricks to make a character that might be able to equal an A-grade Tier 4, a B-grade tier 3, and would dominate a C-grade game with Tier 2 gamebreaking power.

But short of TO tricks?

When I was making the above guide I was searching google aimlessly for information on monks. Besides hundreds of spirited defenses of the best class wizards ever printed, I found a single forum thread from around 2004. It was by a player discussing a challenging encounter his last session where he paid close attention and finally realized that his monk was dealing less damage per round than the wizard’s summoned monster. The responses included others telling him he built wrong, some saying monks are tanks first and foremost and shouldn’t get worried about DPS, and still others saying he must have rolled poorly.

Paladin: The three most prominent paladin archetypes are A-Game buffer, Ubermount and Shining Armor. A-Game shines in grade B, making the paladin a sort of weird Bard archetype. Ubermount also shines in grade B, making the paladin most comparable to a barbarian. "Shining Armor" is a generic term for the C-grade attempt to play the paladin as the designers may have intended. It's not as good, even relative to the playing field.

In Grade-A, none of these three archetypes pulls its weight, but the paladin can still contribute meaningfully to encounters.

Ranger: I am explicitly assuming grade A and B rangers are using Wild Shape, perhaps as a springboard into the Master of Many Forms, the best prestige class. (Heard it here first – and last, folks.) Other differences include wider spell access, partly in the sense of looking through more books, but specifically in that while you can take spells to make you a good archer, it's not really what you rely on.

Rogue: The main difference between C and B can be summed up in 3 words: Use Magic Device. There are other differences, ACFs, Penetrating strike, and Craven being among them. But access to spells via UMD turns the rogue into a more powerful, more versatile performer.

The difference between B and A, however, is rather different. The Rogue's main highlights are skills and somewhat-fiddly damage. At A-grade optimization, neither of those things shines, as both are fairly trivial, and many other classes outcompete the rogue, even by accident (optimization level A is the one where you look through web articles and find divine insight). Rogue drops a tier, having hit its maximum use in grade B. This isn't so bad, as grade B is the one I see most played.

Sorcerer: Complete Mage has a section on how to play spellcasters. p.23: "Summon Monster (I through IX): A creature conjured to fight in your stead offers no guarantee that it will deal even a single point of damage."

This is the only mention of summon monster spells outside of the Summoner archetype. The archetype which says this, under weaknesses: "Casting spells that send summoned creatures into battle might prevent some of the more powerful area spells from being cast at the same time. Although it's certainly not out of the question to cast such spells regardless, one has to question the purpose of even summoning the creatures in the first place."

This chapter lists sample spell lists for each archetype. Imagine a sorcerer following them.

Grade C sorcerers, however, are tricky to tier. Absent optimizations like runestaves, the primary determiner of a sorcerer build's power and utility are the spells known (and how it's played, but that's obviously harder to quantify). Just as low levels are swingy due to the dice being the largest part of every equation (I know what will happen when I roll 1D20+18 to hit and 1D8+20 for damage; 1D20+3 to hit can be off by a factor of ~6 and 1D8+2 changes by a factor of ~3), low grade sorcerers are swingy due to the lack of meaningful obstacles to a single good trick. A sorcerer learning Glitterdust at level 4 can force a Grade C DM to roll over and play dead, start an arms race, or throw exclusively Grimlocks at the party until the sorcerer lets go of his shiny new toy. I rated the sorcerer where I've observed him most, but even at low grades it's possible to stumble onto a really neat trick. Hennet presumably took Spell Focus (Evocation) because he liked his Lightning Bolt, but his third level spells also include Dispel Magic and Haste.

A Grade A sorcerer, meanwhile, knows and can cast as many spells spontaneously as a wizard has prepared. (Grade S sorcerers can cast all the spells forever.)

Wizard: A Grade A or B wizard is probably playing god or something similar. A grade C wizard (in my experience) spends 90% of the time acting like a grade C sorcerer. They stay tier 2, in my estimation, because the few utility spells they select will be enough to bypass many grade C encounters (Fly and Teleport being some of my main picks in my early years).

[bekeleven]

Tier By Optimization Grade

Class	A	B	C
Druid	1	1	1
Wizard	1	1	2
Archivist	1	1	3
Artificer	1	1	4
Cleric	1	1	4
Sorcerer	1	2	4
Warmage	1	4	4
Favored Soul	2	2	3
Wilder	2	3	5
Swordsage	3	3	2
Ranger	3	3	4
Bard	3	3	4
Warblade	3	3	2
Paladin	4	3	4
Barbarian	4	4	4
Fighter	4	4	5
Monk	4	4	5
Duskblade	4	3	3
Rogue	5	4	5
Healer	5	5	5
Warrior	6	6	5
Commoner	6	6	6

In your glowing reviews of my class tiering ability, please, spare my ancestors.

Updates to follow.

[bekeleven]

Incoming responses!

To head off the flood: No, I don't know all of these classes equally well. No, I haven't even played every single one firsthand, much less at every op level. A lot of this is educated guessing.

Even if you disagree with every call I make, please understand the intentions behind it. Simply, tiers are mutable.

I've been ruminating on this for a while. I'm sure everyone will tell me my numbers are all 2 or 3 off, and that's fair. I just wanted to get this somewhere public so people could help me brainstorm what more correct numbers could be. And, of course, add more classes.

I can't promise I'll get to everything immediately, but I plan I'll read every response suggesting more accurate tiers.

[rrwoods]

I like this idea, and I'd like to see it expanded and refined. However I have to disagree with the idea that some of the things in your 'S' grade are things that shouldn't be brought to the table at all.

Diplomacy optimization is useful, when the player isn't purposefully using it on every encounter but rather only the social ones. Maybe this seems arbitrary, but it's really no more arbitrary than the Wizard not solving every encounter with game-breaking spells. To me at least.

A dozen partially-charged wands also seems not-game-breaking to me, but maybe that's lack of experience talking? I could be convinced.

I could see maybe making an 'S' grade for "questionable" things, where most DMs will at least raise an eyebrow and some will outright say no, and then an 'X' grade (which you wouldn't tier the classes in) for the really dumb stuff.

[Endarire]

rrwoods, thank you for explaining the tiers in more detail! It's good to know (and have in a fancy-looking guide post) that optimization ranks (S through F) do matter. A lot. They determine whether someone is 'meh' or shines - at least to a large degree.

As I discovered by playing D&D 5E, being a theoretically powerful character (in the 5E example, a Half Elf Bard focused on crowd control, stealth, and social skills) is powerful if your party plays that way! Some of the most powerful characters are ones that fit within their group and don't try to replace the group.

[LudicSavant]

Diplomacy optimization is useful, when the player isn't purposefully using it on every encounter but rather only the social ones. Maybe this seems arbitrary, but it's really no more arbitrary than the Wizard not solving every encounter with game-breaking spells. To me at least.

"You use the Diplomacy skill basically at all" was on the CharOp campaign smashers thread next to Pun Pun back on the WotC boards, IIRC.

...Just sayin'. YMMV. I, at least, am of the general opinion that the skill is badly designed to the point that it should just be avoided or replaced entirely.

Simply, tiers are mutable.

I agree with this; insofar as obviously the power level of characters varies based on optimization, sources, table standards, player skill, etc.

A few concerns, though.

For one, I question the ability to make practical categories that would cover the wide ranged of player skills, table standards, sources, optimization levels, etc. Too many variables.

Even if categories are limited, your categories seem rather ambiguously defined. Moreso than JaronK's tier list, it's not clear what a given numbered tier means (it seems to mean something distinct from JaronK's definitions). Even if it was, there would be the issue of how you would test your assertions. You could just say "well, a player of X skill wouldn't use that tool." In this sense, how is your opinion subject to falsifiability?

For instance, how exactly do you decide if Tool X for Class X is an A or B level tool? Choosing whether to consider a given tool for a class to be "category A" or "category B" can easily shift their tiers, and it seems like that's pretty much purely a judgment call on your part. How exactly do you decide if Tool X for Class X is an A or B level tool?

Also, does a Tashalatora count as a Monk, a Psychic Warrior, or just something not covered by the thread because the tiers only reflect single class 20 characters? (That last bit doesn't sound right, since prestige classes and such are mentioned)

And of course, if you're doing full scale theoretical optimization, all characters approach the same power level, because everyone starts getting infinite power. I can make Pun Pun with a Commoner.

I guess if your only point is to get across "tiers are mutable" that's cool, but then I just worry that you're putting a lot of work into a chart which could be taken the wrong way.

[Vhaidara]

Diplomacy optimization is useful, when the player isn't purposefully using it on every encounter but rather only the social ones. Maybe this seems arbitrary, but it's really no more arbitrary than the Wizard not solving every encounter with game-breaking spells. To me at least.

Warlock 1/ Binder 1/Marshal 1. I can automatically turn a hostile creature friendly in a single round, and am only about 6 points off of reaching Helpful (which can easily be managed through Item Familiar or Custom +skill items, both legal at that tier of play). Done.

Spoiler: Math

Show

Naberius for Taking 10 and ignore the Rush penalty, making the magic number +40.
6 Ranks
+4 Charisma
+6 Beguiling Influence
+2 Half Elf
+4 Motivate Charisma
+3 Skill Focus
+2 Negotiator
+6 Synergy
+1 Honest

For a total of +34

[rrwoods]

Hm. I see what you're saying on Diplomacy. But, if we're talking about practical optimization levels, we're talking about real campaigns, and how many DMs actually run Diplomacy by-the-book? In my experience, none. Maybe that's not the case on a more global scale, and I know this forum is a big stickler for talking about RAW by default (and with good reason!), but in this case does it make sense to assume a deviation, for practical purposes?

Also, I didn't really explain the tiers... that was bekeleven :-P

[rrwoods]

Pardon the double-post, but I thought of something and figured this was worth the bump. Apologies if this is out-of-line.

Someone mentioned the optimization grades being "ambiguously defined", which I think is a valid criticism. What might we be able to do to fix that, and make this system more meaningful in discussions?

Right now, the grades are defined by their *goals*, which I'll loosely summarize:
S - "win" the game, or break it
A - optimize as much as possible, without choosing options that fall under S
B - avoid options that fall under C
C - naively choose options, assuming that all options are written in a way to help make my character relevant
F - deliberately choose poor options

Give or take a few things here and there, that's my understanding of the grades. As mentioned, for practical purposes, it's safe to ignore S and F.

Where the ambiguity lies is in what types of characters these optimization grades actually produce. For meaningful discussion, we need to define the grades not by *goals*, but by some other easy-to-measure metric. We can't use *effectiveness*, since that's what the tier system gives us (and the whole point of the grade system is to make the class + grade = tier mapping). So what do we use? I'm at a loss here, but I'm extremely interested in this topic.

EDIT: Maybe we can get somewhere if we pick a couple case studies for options other than just classes.

Where does Spirit Lion Totem Barbarian fall, and why?
Where does Time Stands Still fall, and why?
Where does Whisper Gnome fall, and why?
Where does Lesser Aasimar fall, and why?
Where does Leadership fall, and why?
Where does Ocular Spell fall, and why?
Etc.

I reiterate "and why" because that's the real question here, and it may lead us to a pattern that gives us a real metric by which to define the grades. I have my ideas about each of those things, but some of them may be more controversial. The OP has a couple case studies as well but doesn't address them in detail (diplo optimization, for example) and I think addressing each case in detail can give us clues.

[Ger. Bessa]

I might not see something obvious, but about the suggestions for Optimization Tiers

* What is TO Theurge stacking ? (And when does a theurge gets more power than a caster with a dedicated PrC ?)
* Where does the "free metamagic" of Anima Mage and Recaster goes ? The obvious thought was for DMM, Incantatrix and maybe Circle Magic, but what of definitely limited /day uses ? Is Ultimate Magus that efficient at the suggested Grade A optimisation ?
* What of multi-pouncing (with teleportation and Shadow Pounce) ? (Guess it goes into "stacking" and questionnable interpretations)
*Isn't a stat of 25+ normal with Elite array + WBL ? (15 + 5(lvl) + 6(item) ). The wish or tome boost reaches 30 but 25+ at lv20 seems normal (unless you're talking of stats at lv 1).

Also, many "rules" are not to be taken blindly, (no more than 1 PrC ? what about fluff ACFs ?) and I might add that 7th level chameleon spell slots taste more like TO cheese than basic ancestral relic runestaff. (AR staff of Wish recharged with Wall of Salt otho, is reblochon)

[LudicSavant]

Hm. I see what you're saying on Diplomacy. But, if we're talking about practical optimization levels, we're talking about real campaigns, and how many DMs actually run Diplomacy by-the-book? In my experience, none.

If all DMs are houseruling Diplomacy, then they by definition are not using the RAW Diplomacy rules which Bekeleven is suggesting that they should not use.

[GilesTheCleric]

I like having a grading system that doesn't focus on the flexibility/ power dichotomy. This one seems to better fit a table manners/ system mastery/ ease of building axis.

However, it doesn't fit (wasn't designed to fit, I think) my favourite type of "optimization": focused building. When building a character, I like to build to a theme or to some specific ability. This means that I search far and wide for build resources that help emphasize my particular goal. Some of the options are worse picks than something else that could have occupied the same resource space, and thus end up affecting the tier rating of the final build. Now, some of the methods of completing such a build can be torturous, sometimes even requiring calls to tricks this guide calls grade S. I am certain that such builds would fall under grades B or A in this guide, and tiers 2-4 when looking at their output. Think Iron Chef.

How can this guide account for builds like this? The tier system does so by judging output, which is a fairly defined value. I'm grateful that this guide instead looks at inputs (also decently well-defined), but that method as it stands doesn't account for synergies or for system mastery applied to the resources of lower grades. Is this something you would like this treatise to account for, or is it beyond the scope? The ease of building aspect of this guide seems like it should apply, but doesn't quite.

Edit: I think that the reason this guide doesn't fit quite so cleanly is because the inputs being evaluated are more varied: System mastery and resources at the same time. The tier system evaluates only two outputs, but its method of structuring them turns them into single outputs: First flexibility, then power, where T1 and T2 are effectively separate from T3+.

[LudicSavant]

Someone mentioned the optimization grades being "ambiguously defined", which I think is a valid criticism. What might we be able to do to fix that, and make this system more meaningful in discussions?

Right now, the grades are defined by their *goals*, which I'll loosely summarize:
S - "win" the game, or break it
A - optimize as much as possible, without choosing options that fall under S
B - avoid options that fall under C
C - naively choose options, assuming that all options are written in a way to help make my character relevant
F - deliberately choose poor options

If we're determining tiers by which options you're allowed to use, ambiguity stems from the fact that it's not clear how one is deciding if an individual component option is S, A, B, or C. Moving even one option up or down a category can meaningfully change the relative power level of classes in those categories.

A much less ambiguous way to do things would be to compare the power levels of actual specific builds.

Maybe we can get somewhere if we pick a couple case studies for options other than just classes.

Where does Spirit Lion Totem Barbarian fall, and why?
Where does Time Stands Still fall, and why?
Where does Whisper Gnome fall, and why?
Where does Lesser Aasimar fall, and why?
Where does Leadership fall, and why?
Where does Ocular Spell fall, and why?
Etc.

I reiterate "and why" because that's the real question here

Pretty much.

[xyianth]

I might not see something obvious, but about the suggestions for Optimization Tiers

* What is TO Theurge stacking ? (And when does a theurge gets more power than a caster with a dedicated PrC ?)
* Where does the "free metamagic" of Anima Mage and Recaster goes ? The obvious thought was for DMM, Incantatrix and maybe Circle Magic, but what of definitely limited /day uses ? Is Ultimate Magus that efficient at the suggested Grade A optimisation ?
* What of multi-pouncing (with teleportation and Shadow Pounce) ? (Guess it goes into "stacking" and questionnable interpretations)
*Isn't a stat of 25+ normal with Elite array + WBL ? (15 + 5(lvl) + 6(item) ). The wish or tome boost reaches 30 but 25+ at lv20 seems normal (unless you're talking of stats at lv 1).

Also, many "rules" are not to be taken blindly, (no more than 1 PrC ? what about fluff ACFs ?) and I might add that 7th level chameleon spell slots taste more like TO cheese than basic ancestral relic runestaff. (AR staff of Wish recharged with Wall of Salt otho, is reblochon)

TO theurge stacking is the use of both sides of a theurge on the same base class, netting you two effective levels of casting per level in the theurge class. (This can be done a variety of ways, but one of the easiest is to play a sha'ir 4/mystic theurge 8; which lets you cast as a 20th level sha'ir at ECL 12)

Any free metamagic is potentially gamebreaking. It depends on how it is used. I'd personally put free metamagic in the high-end of PO, (tier A in this thread) but arguments could be made for it being TO level. (tier S) Ultimate magus has the potential of combining TO theurge stacking with metamagic reduction. I'd place it in tier A as well, though any use of it to theurge stack is strictly tier S territory.

Multi-pouncing is not much better than an ubercharger in my opinion, even when executed on an idiot crusader type build. I'd place it in tier B, just like most uberchargers. If you can somehow find a way to do it more than 4 times per round, it might enter tier A territory; but I'd think that was more a result of whatever you are using to break the action economy instead of the multi-pouncing.

I think the stat thing was saying any stat equal to 25+character level (i.e 45+ at ECL 20) is tier A. A score of 25 is not unusual at all, a score of 45 requires several concurrent optimization tricks.

As for not taking rules blindly, I believe that was covered in the whole Player>Build>Class idea. I'd say that any tricks from tier A (for example) used to optimize a less than optimal tactic is not tier A anymore. But when rating tricks, we should assume that the trick is being used to its fullest when determining what tier it occupies.

[Troacctid]

If all DMs are houseruling Diplomacy, then they by definition are not using the RAW Diplomacy rules which Bekeleven is suggesting that they should not use.

Let's be fair, the Diplomacy rules give DMs a lot of wiggle room as written.

[bekeleven]

For one, I question the ability to make practical categories that would cover the wide ranged of player skills, table standards, sources, optimization levels, etc. Too many variables.

Trust me, my initial mental model had several more grades. I eventually squished it into 3 grades, with a ++ up top and a -- on the bottom. Most of the grades I initially thought of ended up under grade B (the divisions between them were muddier than the rest), which is why it's rather broad. In my opinion, the vast majority of players here build optimization grade B.

Even if categories are limited, your categories seem rather ambiguously defined. Moreso than JaronK's tier list, it's not clear what a given numbered tier means (it seems to mean something distinct from JaronK's definitions). Even if it was, there would be the issue of how you would test your assertions. You could just say "well, a player of X skill wouldn't use that tool." In this sense, how is your opinion subject to falsifiability?

The tiers are defined the same as JaronK's tiers, with an addendum, mainly: "This is how this class performs relative to other classes at this optimization grade, assuming the campaign is being run at this optimization grade." This is, for instance, one reason that warblade and wizard converge on tier 2 at grade C. Both tend to completely overpower by-the-book early 3.0 and 3.5 modules, but both require just a bit more mastery than grade C to break all aspects of a campaign in half. I suspect that in plenty of grade C campaigns, the warblade will overpower the wizard. It certainly looks more impressive in all the metrics I used to grade characters in 2003.

For instance, how exactly do you decide if Tool X for Class X is an A or B level tool? Choosing whether to consider a given tool for a class to be "category A" or "category B" can easily shift their tiers, and it seems like that's pretty much purely a judgment call on your part. How exactly do you decide if Tool X for Class X is an A or B level tool?

You're right. It's 100% a judgement call on my part. I think the broad strokes, the shape of the curves, are straightforward enough that most would agree. But when it comes down to the minutiae, the questions are

• If I see other characters in this campaign using custom skill-boosting items and metamagic reducers, what do I feel is fair for me to build?
• What requires similar amounts of system mastery (book access and system knowledge)?
• What feels similarly in the "spirit of the game"? For instance, Grade C is the game as the designers intended it to be played. While Grade A is the tippyverse and beyond. Create food traps breaking the economy is the simplest example of something that doesn't require much system mastery, doesn't feel "unfair," but still moves the game away from the spirit in which D&D was intended to be played. (From there it's just a hop to the Mage's Lucubration trap.) I'm not saying this as a judgement call, that grade C is better. I'm using it as a metric on what feels more Gygax at one end and more Tippy/CaptnQ/Biffoniacus_Furiou/Rubik on the other end. As I said, I tend to play somewhere in the middle.

It's not about power and versatility. That's what the tiers explain. Grades are a separate axis from those. That's why grade A and B low-tiered characters can party with grade B and C high-tiered characters. For a very, very rough equivalency, Grade A = Grade B+2 = Grade C+4. So a newbie-op Druid can party with Rubik's rogue/healer.

Remember: The original Tier system states, "Given equal system mastery and book access, here's the relative power and versatility of each class." I'm breaking it down and saying, here's how that relationship changes as system mastery and book access change. In general, I view system mastery as a larger factor than book access, since it determines playstyle, and here, let me center some text again, it's fun.

Player > Build > Class

Also, does a Tashalatora count as a Monk, a Psychic Warrior, or just something not covered by the thread because the tiers only reflect single class 20 characters? (That last bit doesn't sound right, since prestige classes and such are mentioned)

Tashalatora is a psywar option in my view. Just like a diplomancer with a marshal build doesn't elevate the marshal to tier 1-2, Tash doesn't bring up Monk. This is a convention from the JaronK's original tier system. Go here and read, "But what about dips?"

And of course, if you're doing full scale theoretical optimization, all characters approach the same power level, because everyone starts getting infinite power. I can make Pun Pun with a Commoner.

That's the example I used to explain why I didn't tier Grade S.

Right now, the grades are defined by their *goals*, which I'll loosely summarize:
S - "win" the game, or break it
A - optimize as much as possible, without choosing options that fall under S
B - avoid options that fall under C
C - naively choose options, assuming that all options are written in a way to help make my character relevant
F - deliberately choose poor options

This is roughly what I was shooting for. I'll add, again, that Grade B is the widest and tends to include most play from enfranchised players (the type that visit these forums and post in threads like this).

I might not see something obvious, but about the suggestions for Optimization Tiers

* What is TO Theurge stacking ? (And when does a theurge gets more power than a caster with a dedicated PrC ?)
* Where does the "free metamagic" of Anima Mage and Recaster goes ? The obvious thought was for DMM, Incantatrix and maybe Circle Magic, but what of definitely limited /day uses ? Is Ultimate Magus that efficient at the suggested Grade A optimisation ?
* What of multi-pouncing (with teleportation and Shadow Pounce) ? (Guess it goes into "stacking" and questionnable interpretations)
*Isn't a stat of 25+ normal with Elite array + WBL ? (15 + 5(lvl) + 6(item) ). The wish or tome boost reaches 30 but 25+ at lv20 seems normal (unless you're talking of stats at lv 1).

Also, many "rules" are not to be taken blindly, (no more than 1 PrC ? what about fluff ACFs ?) and I might add that 7th level chameleon spell slots taste more like TO cheese than basic ancestral relic runestaff. (AR staff of Wish recharged with Wall of Salt otho, is reblochon)

A few notes -

• TO Theurge stacking is using a theurge to advance the casting of two classes, at least one of which is a theurge. I am of the opinion that it doesn't work by RAW, but there's an argument.
• Using metamagic to reduce a spell below its level or to persist a pile of round/level buffs (DMM, as you said, being the easiest method). Using Easy Metamagic to cast Rapid spells at +0 3 times a day is not. The system is holistic. There's a big sliding scale between those extremes.
• Level+25 stats (think 26 int at level 1, 35 int at level 10, 45 int at level 20).
• Ancestral Relic Custom Runestaff refers to a build liked by some people on these forums where, basically, you can cast any spell spontaneously for a single feat investment. The mechanics are more complicated, of course.

I'm not too familiar with multi-pouncing. I'd have to see the specific build.

Edit: Turns out there's a bunch of responses since I started writing this. I will get to them soonish.

[Malroth]

Warlock
Tier D easy to enter by accident by believing the class is a blaster Runs about a 3 compared to other D's or a 6 if trying to be a C and failing
Tier C Slightly Better options at this level than a Bard but not by much 4
Tier B Worst place for a warlock, can be an ok DPS guy with eldritch claws but lack of versitility hurts here 4
Tier A Can actually serve as a DPS blaster or a HIPS stealth monster or Decent artificer replacement or a Powerful Minion master but is unilkely to be more than 2 of these 2
Tier S Works surprising well in this tier, Enhanced Diplomacy, Ignores item creation prereqs, NAD, and a few infinite loops that can easily be started with self crafted scrolls or by darkspeeching yourself a few hiveminds Low2

[LudicSavant]

Remember: The original Tier system states, "Given equal system mastery and book access, here's the relative power and versatility of each class." I'm breaking it down and saying, here's how that relationship changes as system mastery and book access change. In general, I view system mastery as a larger factor than book access, since it determines playstyle, and here, let me center some text again, it's fun.

I get what you're going for, but I feel there are a few issues that you may want to look into. You aren't evaluating system mastery and book access as separate categories, for example. Also, player ability can not only take many sizes, but also many shapes. What one guy finds easy to optimize, another may find difficult, and vice versa. What is high effort / obscure / complex to one person may be low effort / obvious / simple to another, and vice versa. Different players have an easier time with different things.

For example, a lot of people talk about Fighters being easier for them, but in my newbie days I seem to recall having had an easier time optimizing casters, simply because I had the freedom to change out elements of my build more easily and didn't feel as much of a need to plan build elements in advance. I could just try things out, and if I didn't like them, I could try something else without building a whole new character. I didn't need to have any feat trees or expensive equipment or anything to make a Web spell do its thing.

Player > Build > Class

I generally agree with that. I can even provide a lot of examples from my own experience (e.g I've seen a smart player with a purposely gimped character dominate a game, and I've seen a player who took a ridiculous fleshraker druid build straight off the forums have no idea how to use it and be absolutely useless).

However, your chart still is tiering classes, and leaving it rather vague what tools, builds, etc they are using to be evaluated as such. What if you tried tiering specific build archetypes? That seems like it would remove a lot of ambiguity regarding what tools are being evaluated, and would make the chart more open to falsifiability (e.g. someone could reasonably come in and say "hey, this build is better than that build")

[MukkTB]

I like the idea, but I have some objections to the categorization of what behavior makes S, A, B, C, and F. Particularly I disagree on some specifics, but I also disagree with only listing a few examples. I think rrwoods' statement of goals is accurate, but it is also not accurate enough to be useful for specifics. We either need to be looking at a build by build classification or go item by item classifying each with regard to its place on the chart.

Going item by item I immediately see disagreement coming into play. Particularly class S is going to be a point of contention. We can probably all agree about infinite loops and drown healing. On the other hand partially charged wands does not seem to belong in the same class of things as infinite loops. We feel strongly about some of these matters, and the classification of a single little thing has spawned mega threads full of flame.

I would feel better about addressing individual builds. However even that is a daunting challenge. And at many levels of play, builds are not straight jackets. You can combine builds. You can adapt them to your tastes. Or you can build strong and weak generic setups that are not defined by a small set of choices or abilities but instead represent a well rounded approach to the class.

Spoiler: Partially Charged Wands

Show

For example with regard to partially charged wands. These items can facilitate cost efficient cheese. However there does not seem to be any reason in game why some wands should not come onto the market with variable number of charges left. A DM attempting to simulate a living breathing world would probably be able to respond to players searching for partially charged wands. However I would expect these items to not always (rarely?) come in with the exact number of charges the player desires. I would not even expect such items to exist on the market for all spells. Buying a partially charged wand, like buying or building any custom item, even if it is allowed by the rules, requires a bit of roleplay time between the player and the DM.

[rrwoods]

I think rrwoods' statement of goals is accurate, but it is also not accurate enough to be useful for specifics.

That's kinda my point: They're classified by goals, but they *want* to be classified by something else (that maps closely to those goals in practice, but is also easy to specify).

[Beheld]

Honestly, to me the biggest problem with the Tier system is not that is doesn't tell you who is the strongest at what level of optimization, it is that it doesn't tell you the things you might actually want to know to find out if you would enjoy playing a character.

To me the questions you actually want to know about a class before you consider playing it are:

1) How easy is it to "build" a strong character of that class?

2) How easy is it to play the character after it has been built?

So for example: It is really easy to build a competent Beguiler, you write Beguiler on your character sheet, and then you have a competent beguiler, and you could also do better by looking up good domains to arcane disciple or whatever, but that isn't necessary. So if someone wants to make a character for the first time, I can recommend Beguiler over Wizard, because Beguiler will be easier. Wizard on the other hand is very hard to make, because you have to make individual decisions as to which spells you include, and you have to evaluate a lot of spells to do it. And Barbarian is even harder, because the only viable Barbarian builds involve tons of source dipping for that one thing, and cheesing to get a competent character, and if you make any mistakes, it isn't like a Wizard, who can just fall back on a great different spell, it just means you are bad.

On the other hand, if ten years ago I was making a character for my kid brother, or for someone knew who isn't a regular RPG nerd, and is just getting into things, I want to know that Sorcerers, even though they are harder to build, are really easy to play once built. And so are Barbarians, but Druids are actually super hard, because you have to pick each day from every single druid spell in existence, and then you have to decide on good wildshape forms based on all sorts of factors, and then you have to choose in combat whether you cast a spell you prepared, wade into melee, or spontaneously summon one of a whole bunch of monsters. That's a lot going on, so a new player is more likely to make mistakes or throw up their hands in frustration.

One thing I don't care about even a little bit, is how easy or hard, or even worse, how many different way a class can break the game, which JaronK's tiers loveee to count. If a character can break the game in one way, then it can break the game in every way that matters, because every table gets to break the game only one time before the game is broken. If you already have an infinite shadow army and Glabrezu army, and Wished for a staff of wishes, then you are done, there is no more game to play. And it doesn't matter if you could also use schism to BS yourself to 13 standard actions a turn, because the guy with access to infinite wishes can also just wish to get items that allow him to have lots more standard actions through a variety of means. No one cares how many ways you can break the game (except JaronK).

[GilesTheCleric]

Bekeleven, thank you for thinking of this and posting it. I'm glad to have a different codified way to think about classes in relation to each other.

This thread seems to have gotten a bit "constructive", so maybe I can help balance that out a bit.

[Killer Angel]

Introduction
The existing Tier system doesn’t take into account what happens when Emperor Tippy builds a wizard in the same party as a 13 year-old builds a fighter.

Doesn't the Tier System suppose the same level of optimization in the party? Am I remembering it wrong?

[xyianth]

Doesn't the Tier System suppose the same level of optimization in the party? Am I remembering it wrong?

Yes, you remember correctly. The point of the tier system was to provide a quick and easy reference for which classes are better played together assuming all players had the same amount of optimization applied. (whether that was none or tippyverse) As a side bonus, it made it somewhat easy to make allowances for players with better optimization skills. You could limit the better optimizers to lower tiers and help achieve a better result overall. I think this thread is just trying to codify those allowances at a finer grain.

[rrwoods]

... Sort of.

The tier list assumes a specific but unspecified level of optimization. It's useful only if you know what that level of optimization is.

Saying that wizard is tier 1 is going to seriously confuse anyone in optimization grade C, especially completely new players, because their power is so not-apparent when reading only their class description. I'd almost argue that at grade C wizard is tier 4.

Also: the tier system isn't designed to measure how enjoyable a concept is to play; that kind of thing isn't measurable. That doesn't make it useless. That conversation is offtopic for this thread, but I'd argue that knowing what tiers your party mates plan to play CAN increase the enjoyment factor of the game as a whole.

[LudicSavant]

Saying that wizard is tier 1 is going to seriously confuse anyone in optimization grade C, especially completely new players, because their power is so not-apparent when reading only their class description. I'd almost argue that at grade C wizard is tier 4.

I think this whole system makes the mistake of assuming that newbies, in general, will make the same mistakes and assumptions as each other. This is simply not true in my experience. Not every new player looks at a Cleric and instinctively assumes that their job is to be a healbot and that other things in their toolbox aren't important. Not every new player looks at a Wizard and then doesn't immediately start looking over what these spells he's being granted are. It doesn't require a "build" for someone to just grab Stinking Cloud and start being awesome.

By the same note, players can easily make mistakes with non-casters, too. There was just a thread recently where there was a DM talking about a level 5 character in their party who was only doing a low attack bonus ~1d8 damage hit every round, for example. That's not going to be doing anything in a manticore fight.

[Fizban]

I also disagree with your ranking of some game elements*, but the idea is good. Rather than focusing on a list of examples which is likely to cause immediate disagreements, I'd lean more towards rrwoods's rephrasing with the examples shifted down or spoilered or something to emphasize a lack of finality. Rank S is exploiting the rules to the max, A is avoiding the worst of S, and B is trying but not super hard. C rank is what I admittedly look down on but seems to be the most common in real life: just doing the obvious. Having a table of those interactions is quite handy, makes it easy to go down the list and see which classes don't change and which do, just needs moar classes.

*Ancestral Relic Runestaff, I don't know what other people are doing with it but I just see it as the sorcerer's answer to the wizard's ability to buy any spell for dirt cheap, and then either make their own custom Runestaff or use some other spontaneous casting trick. If Knowstones exist then sure it looks cheap, but those are dragon mag, and Runestaves have a cast limit that cuts into the sorcerer's "I cast until it works" shtick.

[Killer Angel]

Saying that wizard is tier 1 is going to seriously confuse anyone in optimization grade C, especially completely new players, because their power is so not-apparent when reading only their class description. I'd almost argue that at grade C wizard is tier 4.

Yeah, makes sense.
And anyway, in the end, the Tier System is more what you'd call guidelines...

[Melcar]

Introduction

I built grade C characters from 2003 until about 2007, when I signed up for the wizards boards out of curiosity.

This is me as well. I have got to say though, that I really dont play A or B characters all the time. Sometimes I get more joy out of C characters, because they seem more "realistic" which is kind of stupid in a high fantasy game, bot there is a more LoTR feeling to C grade characters... Maybe its just me.