Monster Classes: Reworking Types/Subtypes, HD, etc from the Ground Up

[Durzan]

As was touched on in this thread, having monster type and HD inextricably linked creates problems. So lets fix this. While we're at it, lets collab on reworking monsters from the ground up.

How do we go about creating classes exclusively for monsters? How do we separate monster type from HD and all other similarly dependent stats? What needs to be consolidated and simplified? What needs to remain the same?

I'll use the OP and the next post to keep track of a consensus.

[Durzan]

Alright, lets discuss.

[noob]

It is simple: you make 3 kinds of hit dice: Brute, skilled and Defensive.
Brute gets full bab and big hit dice as well as a strong save.
Skilled gets a strong save, a lot of skill points and medium bab and medium hit dice(probably D8)
Defensive gets 3 strong saves(the rest of the specifics is less important)
Most outsiders would have a mix of defensive and of skilled(and a bit of brute for some specific kinds of outsiders) while stuff like dragons or gelatinous cubes or most animals would get mostly brute hit dice.
The varied kinds of hit dice are "classes" just like racial hit dice and each hit dice increase the cr of 1/3

[Durzan]

It is simple: you make 3 kinds of hit dice: Brute, skilled and Defensive.
Brute gets full bab and big hit dice as well as a strong save.
Skilled gets a strong save, a lot of skill points and medium bab and medium hit dice(probably D8)
Defensive gets 3 strong saves(the rest of the specifics is less important)
Most outsiders would have a mix of defensive and of skilled(and a bit of brute for some specific kinds of outsiders) while stuff like dragons or gelatinous cubes or most animals would get mostly brute hit dice.
The varied kinds of hit dice are "classes" just like racial hit dice and each hit dice increase the cr of 1/3

Not a bad starting point.

Edit: Updated the title and content of the OP to better reflect the purpose behind this thread.

[ideasmith]

Simplest Answer: Use the Augmented subtypes. So that a fomorian might be Fey (Augmented Giant).

Simple Answer: Give a name to each of the existing types of creature hit dice. Mix and match as desired.

Neither of these require rewriting existing monsters. (Though both allow such.) This makes life a lot easier for the DM than a whole new set of monster classes which require rewriting every single monster in the game.

[Durzan]

Simplest Answer: Use the Augmented subtypes. So that a fomorian might be Fey (Augmented Giant).

Simple Answer: Give a name to each of the existing types of creature hit dice. Mix and match as desired.

Neither of these require rewriting existing monsters. (Though both allow such.) This makes life a lot easier for the DM than a whole new set of monster classes which require rewriting every single monster in the game.

Rewriting existing monsters and simplifying the system is part of the point. That includes cutting and simplifying the number of types and subtypes. Heck, that was the entire point of the thread linked in the OP. Setting up monster classes was half the discussion in that thread, so I thought it might be smart to start up a separate thread with that particular purpose in mind.

[Ursus Spelaeus]

There are a lot of cases where a monster would seem to straddle two types. Should a modron be a construct or an outsider?
To handle this situation, you might consider gestalt rules for monster classes.

[PairO'Dice Lost]

There are a lot of cases where a monster would seem to straddle two types. Should a modron be a construct or an outsider?
To handle this situation, you might consider gestalt rules for monster classes.

If you're rewriting the list of types anyway, you don't need to keep every old type around as a full type. Construct could easily become a subtype, and so can Undead, so for instance a modron is an Outsider (Construct), a stone golem is an Elemental (Construct), a dracolich is a Dragon (Undead), a banshee is a Fey (Undead), and so forth.


Of course, some sort of pseudo-gestalt is a good option for other reasons, namely all the monsters that are basically a PC class stapled to a monster. Instead of giving various monsters arbitrary progressions that generally end up far too strong or far too weak for their CR, it's much better and easier if you can say a solar is an Outsider//Cleric, a true dragon is a Dragon//Sorcerer, a pixie is a Fey//Ranger, and so on. And it solves the RHD problem and the associated-class-levels hack, too; a frost giant (LA 4/RHD 14) isn't worth anywhere near ECL 18, but a frost giant Giant//Fighter or Giant//Barbarian 18 might be, and a frost giant shaman is over-statted and under-magicked at RHD 14/Blackguard 8 but would be much better as a Giant//Cleric 14 or the like. Just figure out what the cost for that is (ECL +2 makes sense, given 2 CR X creatures = 1 CR X+2 creature) and you're golden.

[rferries]

If you're rewriting the list of types anyway, you don't need to keep every old type around as a full type. Construct could easily become a subtype, and so can Undead, so for instance a modron is an Outsider (Construct), a stone golem is an Elemental (Construct), a dracolich is a Dragon (Undead), a banshee is a Fey (Undead), and so forth.

Very much agreed with this!

[ideasmith]

Rewriting existing monsters and simplifying the system is part of the point. That includes cutting and simplifying the number of types and subtypes. Heck, that was the entire point of the thread linked in the OP. Setting up monster classes was half the discussion in that thread, so I thought it might be smart to start up a separate thread with that particular purpose in mind.

My apologies. When I read the original post, simplification looked like a possible means rather than the intended goal.

A useful step in revising something is examining what it is already like. I am therefore providing a list of the existing ‘monster classes/hit dice’ for reference. I have added nick names for ease of communication (since ‘8-sided hit dice, medium BAB, good Fortitude saves, and 2 skill points per level’ is rather wordy).

Nickname	Hit Dice	BAB	Good Saving Throws	Skill Points	What Currently Has
Glamorous	6	0.5	rw	6	Fey
Heavy	8	0.75	f	2	Elementals (Earth or Water); Giants; Plants, Vermin, Some Humanoids
Bestial	8	0.75	fr	2	Most Animals
Dire	8	0.75	frw	2	Some Animals
Light	8	0.75	r	2	Elementals (Air or Fire); Some Humanoids
Firm	8	0.75	w	2	Aberrations; Some Humanoids
Wondrous	8	1	frw	8	Outsiders
Uncanny	8	1	rw	2	Monstrous Humanoids
Empty	10	0.75	-	2	Constructs; Oozes
Enchanted	10	1	fr	2	Magical Beasts
Stiff	12	0.5	w	4	Undeads
Awesome	12	1	frw	6	Dragons

There are a lot of cases where a monster would seem to straddle two types. Should a modron be a construct or an outsider?
To handle this situation, you might consider gestalt rules for monster classes.

One solution is to add new subtypes such as outsider-like or construct-like. I don’t see gestalting hit dice as helpful for this. Gestalting would mainly affect the features, and what you’d want to meld is the traits.

[Just to Browse]

I do think you should do monsters from the perspective of a creature's job, but I don't think the classification of brute / skilled / defensive is all that useful. I am particularly against the concept of the "skilled" monster being anything other than an exception.

Monsters are foes for PCs. It is their primary role, and in many campaigns it's their only role. It's important to recognize this as your starting point, because it tells you which characteristics are valuable and which ones are an annoyance to track. In general, tactical guidance like "don't eat elves or dwarves" shouldn't show up in PC options, and large quantities of skill points shouldn't show up in monster options (though there are exceptions to both of those heuristics).

I think of monsters along 3 dividing lines:

• Defense: How much effort does it take to handle the monster?
• Offense: How dangerous is the monster?
• Access: How easily can the monster affect any given PC?

Some of the common monster roles in my mind are:

• Artillery: This creature's job is to present a threat that must be handled at range or by closing the gap. Manticores are artillery monsters. Weak defense, medium offense, strong access.
• Boss: This creature's job is to be a boss monster. Dragons are boss monsters. Strong defense, offense, and access.
• Closet (troll): This creature's job is to murder you for having the audacity to be in full attack range. I hate this kind of design, but it's all over D&D. Hydras are closet monsters. Medium defense, strong offense, weak access.
• Filler: This creature's job is to take up space in a fight. Brown bears are filler monsters, as are most animal companions. Medium defense, offense, access.
• Fodder: This creature's job is to get kicked around. Your average NPC-classed CR 1/2 races are fodder monsters. Weak defense, offense, and access.
• Meatbag: This creature's job is to present a resilient threat that usually requires consistent attention. Zombies are meatbag monsters. Strong defense, medium offense, weak access.
• Puzzle: This creature's job is to create an encounter that players must "solve". Hellcats are puzzle monsters. Weak defense, medium offense, medium access.
• Trap: This creature's job is to surprise the least suspecting players, then die quickly. Mimics are trap monsters. Weak defense, strong offense, strong access.

From there, some of the basic chassis stuff seems easy. Artillery probably need 2 good saves but a low HD so they're resistent to BFC and weak to rogues. Meatbag probably need 1 or 0 good saves but a high HD and bonus HP or something. Puzzles and Traps should have terrible defense. Closet monsters should be booted off the face of the earth.

There is also room for hybrids in this model. I wrote the kanzzadit as a puzzle monster the first time the PCs encounter it, and a filler monster once the players know it's laundry list of arbitrary abilities.

[noob]

I do think you should do monsters from the perspective of a creature's job, but I don't think the classification of brute / skilled / defensive is all that useful. I am particularly against the concept of the "skilled" monster being anything other than an exception.

Monsters are foes for PCs. It is their primary role, and in many campaigns it's their only role. It's important to recognize this as your starting point, because it tells you which characteristics are valuable and which ones are an annoyance to track. In general, tactical guidance like "don't eat elves or dwarves" shouldn't show up in PC options, and large quantities of skill points shouldn't show up in monster options (though there are exceptions to both of those heuristics).

I think of monsters along 3 dividing lines:

• Defense: How much effort does it take to handle the monster?
• Offense: How dangerous is the monster?
• Access: How easily can the monster affect any given PC?

Some of the common monster roles in my mind are:

• Artillery: This creature's job is to present a threat that must be handled at range or by closing the gap. Manticores are artillery monsters. Weak defense, medium offense, strong access.
• Boss: This creature's job is to be a boss monster. Dragons are boss monsters. Strong defense, offense, and access.
• Closet (troll): This creature's job is to murder you for having the audacity to be in full attack range. I hate this kind of design, but it's all over D&D. Hydras are closet monsters. Medium defense, strong offense, weak access.
• Filler: This creature's job is to take up space in a fight. A hound archon is a filler monster, as are animal companions. Medium defense, offense, access.
• Fodder: This creature's job is to get kicked around. Your average NPC-classed CR 1/2 races are fodder monsters. Weak defense, offense, and access.
• Meatbag: This creature's job is to present a resilient threat that usually requires consistent attention. Zombies are meatbag monsters. Strong defense, medium offense, weak access.
• Puzzle: This creature's job is to create an encounter that players must "solve". Hellcats are puzzle monsters. Weak defense, medium offense, medium access.
• Trap: This creature's job is to surprise the least suspecting players, then die quickly. Mimics are trap monsters. Weak defense, strong offense, strong access.

From there, some of the basic chassis stuff seems easy. Artillery probably need 2 good saves but a low HD so they're resistent to BFC and weak to rogues. Meatbag probably need 1 or 0 good saves but a high HD and bonus HP or something. Puzzles and Traps should have terrible defense. Closet monsters should be booted off the face of the earth.

There is also room for hybrids in this model. I wrote the kanzzadit as a puzzle monster the first time the PCs encounter it, and a filler monster once the players know it's laundry list of arbitrary abilities.

the truth is that you classified hound archons as "filler" while it works perfectly fine as a boss monster: it is CR4 and at that cr it somehow got perfect mobility(greater teleport at will) , an aura of menace which is a quite strong debuff, relatively powerful attacks for its level that have high attack bonuses and an impressive 10 dr and 16 spell resistance just in case.
It would give trouble to many level 4 adventuring teams especially since stocking up on evil aligned weapons is rare.
CR4 Dragons on the other hand are usually way less dangerous than hound archons.(except if you go and start reading "manual of bad cr association number 5" and the like)

In fact the truth is that dragons are just like the majority of the outsiders: boss if you take one of a much higher level than the party and filler if at a lower level
Filler and boss are not distinct functions.

(unless you consider that greater teleport at will(+flight on most high level outsiders) is good access and not average access in which case you should make a "teleporting polyvalent outsider" category with "good access, above average offence, above average defence")

[Just to Browse]

I keep forgetting about the archon list of abilities. Hound archons are definitely pushing the idea of medium access with greater teleport. They're also pushing the idea of medium defense with magic circle, given that PCs fighting hound archons will probably lean towards evil. I think something like a brown bear is a much better example of CR 4 filler, and I'll update the post to reflect that.

I am going to strongly disagree with the idea that boss monsters & filler monsters are not distinct roles. Good boss monsters have abilities that support different kinds of tactics (on their side & on the PC side). They need a spectacular array of defenses & powerful tricks to save them from dying to kiting, various SoLs, ubercharge attacks, and a myriad of other common PC "I win" tactics. But filler monsters should not have those things, because then they're no longer filler.

If anything, filler & fodder monsters are the most closely related, but at higher levels their jobs diverge more. For example, filler monsters need simple abilities that get them past the "you must be this tall to fight" bar like fear immunity or mobility effects, whereas fodder monsters probably shouldn't.

[noob]

I keep forgetting about the archon list of abilities. Hound archons are definitely pushing the idea of medium access with greater teleport. They're also pushing the idea of medium defense with magic circle, given that PCs fighting hound archons will probably lean towards evil. I think something like a brown bear is a much better example of CR 4 filler, and I'll update the post to reflect that.

I am going to strongly disagree with the idea that boss monsters & filler monsters are not distinct roles. Good boss monsters have abilities that support different kinds of tactics (on their side & on the PC side). They need a spectacular array of defenses & powerful tricks to save them from dying to kiting, various SoLs, ubercharge attacks, and a myriad of other common PC "I win" tactics. But filler monsters should not have those things, because then they're no longer filler.

If anything, filler & fodder monsters are the most closely related, but at higher levels their jobs diverge more. For example, filler monsters need simple abilities that get them past the "you must be this tall to fight" bar like fear immunity or mobility effects, whereas fodder monsters probably shouldn't.

You just did describe hound archons as something that was not in your table of roles at all.
And I doubt a brown bear counts as having medium access: between large size which restrains where it can go and also no mobility power beyond a walk speed of 40 they look quite lacking also they quite lacks defence against spells too much for counting as medium defence.

[Just to Browse]

Large size is generally not an impediment to mobility, and the strength boost from being large can sometimes improve access when it comes to making jump / climb / swim checks. A 40' speed and some simple skills are an appropriate level of access at CR 4 when PCs are, at best, using 2nd-level spells. Contrast that to a creature with legitimately weak access, like an Otyugh outside it's native environment or a Minotaur zombie pretty much anywhere.

I'm also surprised you think the brown bear has weak defenses, given their decent saves (sans will which is an okay thing -- filler monsters are allowed to have weaknesses like all monsters), decent HD, decent AC, and decent HP. Remember that brown bears aren't supposed have to have great or even good defenses. They're meant to take up space and require PCs to spend effort. The brown bear is a stat ball that accomplishes this job.

Also, I'm not sure the a hound archon is outside the description of these jobs, given than monsters can be hybrids of roles. If a hound archon starts as a boss and becomes a filler when PCs get high level, that's perfectly fine. If at-will greater teleport means that archons can never be good filler monsters should instead just be considered CR4 boss monsters, that's fine too.

[PairO'Dice Lost]

I'd add a couple more common roles to round out the list:

• Mastermind: This creature is primarily a non-combat threat, and generally only engages in combat after a great deal of preparation, otherwise it goes down like a chump. Succubi and dopplegangers are masterminds. Strong defense, offense, and access, but in limited areas.
• Controller: This creature isn't much of a threat on its own and primarily acts through minions, whether summoned, animated, mind-controlled, or whatever. Yellow musk creepers and tsochar are controllers. Weak everything alone, varies by minion.

I am going to strongly disagree with the idea that boss monsters & filler monsters are not distinct roles. Good boss monsters have abilities that support different kinds of tactics (on their side & on the PC side). They need a spectacular array of defenses & powerful tricks to save them from dying to kiting, various SoLs, ubercharge attacks, and a myriad of other common PC "I win" tactics. But filler monsters should not have those things, because then they're no longer filler.

Good boss monsters do need to have more capabilities than the average monster, but not every boss monster has to have a huge array of mandatory things, and "Boss" isn't really a distinct role compared to the ones listed here. Beholders are artillery bosses, and survive by shutting down a lot of attacks with antimagic and flight; aboleths are puzzle bosses, and survive because you have to find them first through all the illusions; mind flayers are mastermind bosses, and survive by dividing-and-conquering the party; dragons are closet bosses, and the ones without good spellcasting survive by having good all-around defenses and being able to stay out of reach and strafe with breath weapons; and so forth.

Likewise, fodder come in lots of different varieties as well--goblin archers are artillery fodder, zombies are meatbag fodder, quasits are puzzle fodder, and so forth. Any build-monster-by-role system you come up with is really going to struggle to make all boss monsters and all mooks fit in the same bucket, since boss fights are supposed to be unique and memorable so you're going to want any two boss monsters to be more different from one another than any two monsters of other roles, and mooks all contribute differently to fights (zombies hordes are mostly there as obstacles, massed humanoid archers are mostly there to plink away at HP and force the party to either burn actions and resources to deal with them or be more vulnerable to other monsters, stirges are mostly there as long-term debuffs in monster form, and so on).

Roles like "boss" and "fodder" make more sense as different points on the capabilities-by-CR scale, like the warmage and barbarian having fewer capabilities and choices at any given level than the wizard or warblade despite filling the same thematic niches, and how dragons are generally under-CRed to make them very potent multi-threats. Except obviously you'd make explicit statements to the effect of "these monsters are particularly well-rounded and complex for their HD, their CR is a few points higher and they're good as boss monsters" and "these monsters are particularly narrow and simple for their HD, their CR is a few points lower and they're good as mooks" and have specific ranges for those things in your build-a-monster-from-scratch rules.

[Just to Browse]

I like the two other categories. I think an aboleth is a good candidate for a controller, honestly.

Beholders are artillery bosses, and survive by shutting down a lot of attacks with antimagic and flight; aboleths are puzzle bosses, and survive because you have to find them first through all the illusions; mind flayers are mastermind bosses, and survive by dividing-and-conquering the party; dragons are closet bosses, and the ones without good spellcasting survive by having good all-around defenses and being able to stay out of reach and strafe with breath weapons; and so forth.

I will gladly repeat my claim that hybrid monsters are a totally fine part of this descriptive system, and I fully embrace the idea of Beholders acting as artillery bosses just as much as I embrace kobold archers being artillery fodder. I will also concede that bosses don't absolutely positively need a bunch of defenses. But I do believe that boss and fodder creatures are distinct roles, and I strongly disagree with the idea that you can designate bosses are well-rounded / complex monsters.

To argue this, I want to return to my initial point:

Monsters are foes for PCs. It is their primary role, and in many campaigns it's their only role. It's important to recognize this as your starting point, because it tells you which characteristics are valuable and which ones are an annoyance to track.

Boss monsters should be defined by "can this creature be a good boss fight?" While it's usually easier to write a good boss fight when it has a lot of explicit options, a boss doesn't need to be complex. Bosses do need to support different kinds of tactics so players can get creative, and there's a very good chance that bosses must be designed around SoLs, kiting, uberchargers, etc because those are the tactics PCs use to end fights quickly and a "good boss fight" should not end before everyone has acted at least once.

Not all complexity or broadess is inherently boss-worthy. For example, on a Hamatula, the long list of skill points, 3/day usage of fireball / lightning bolt SLAs, and Infernal Wound are not meaningful complexities. They're technically options for the monster, but they don't complement what it does and they're all fairly weak for a CR 16 monster. It doesn't matter at all that a Hamatula has +8 when using Disguise to act, but it does matter that a Hamatula has SR 28, regeneration, teleportation, illusions, and a solid full attack routine with a high chance of stunlocking. If the Hamatula didn't have illusions, teleportation, or SR, it wouldn't matter how many 3rd-level evocations it could cast 3 times per day, it would still be a pretty crappy boss fight. Those options do not inherently make a creature worthy of boss monster status.

This is in contrast to the jobs of most other creatures. If an ubercharger can find the right spot, they should be rewarded with the ability to quickly dispatch a non-boss artillery creature, whereas a beholder should actively disintegrate terrain to prevent charging and/or aim its antimagic field at souped-up PCs.

On a similar note, fodder doesn't inherently benefit from being narrow and simple. For example, a hydra's fast healing makes it a bad fodder creature even though it is extremely narrow and simple, especially at higher CRs.

I want to note something important that I agree with you on: At the end of the day, the best way to build a boss / fodder monster may be to just designate CR ranges in which certain creatures are good bosses / fodder. But that doesn't mean that boss & fodder aren't distinct roles. Dragons and beholders are simply designed for a different purpose from their fellow closet / artillery monsters, and that's something we should recognize when building D&D Fix #299238304.

[PairO'Dice Lost]

I like the two other categories. I think an aboleth is a good candidate for a controller, honestly.

It's sort of a hybrid controller/mastermind with a side order of puzzle, really. A delicious, delicious side order. Tastes like PC.

I will gladly repeat my claim that hybrid monsters are a totally fine part of this descriptive system, and I fully embrace the idea of Beholders acting as artillery bosses just as much as I embrace kobold archers being artillery fodder. I will also concede that bosses don't absolutely positively need a bunch of defenses. But I do believe that boss and fodder creatures are distinct roles, and I strongly disagree with the idea that you can designate bosses are well-rounded / complex monsters.

The difference between boss/mook roles and the other ones is a matter of scope. Every other role is defined both by what it is and what is isn't. Artillery critters are strong at range and often good at keeping out of range of enemies, and they aren't good at attacking or defending in close quarters and generally can't take hits even at range. Meatbag monsters do have good all-around defenses and often have some sort of fast healing/DoT/etc. ability that punishes you for leaving them alone for more than a few rounds, and they don't have any good ways to deal with fast or far-away foes and generally aren't impressive on their own. You can take a role and take a monster concept and figure out what abilities would and wouldn't fit.

However, bosses don't really have an "isn't" list, and mooks don't really have an "is" list. A boss is (or should be) good at dealing with a whole party, can resist or ignore many combat-ending effects, has several varied tactics it can take in combat, and so forth, but there's not really anything you can point to and say "This isn't a thing a boss should have" (even things like "dies in one hit from certain weapon" or "dies on contact with a certain materian" have their place; rakshasas and Bowser say hi), and likewise a mook isn't (or shouldn't be) particularly complicated to run because the DM might have to use a couple dozen at once, can't punch above its CR with any niche SLAs or the like, doesn't make for a hard or challenging encounter on its own, and so forth, but there's not really anything specific that a creature must have to qualify as a mook.

If you say something is an artillery/mastermind you have some idea of what that would mean for its combat style and what synergies might be present, and you can look at a hypothetical artillery/closet monster and say that doesn't make sense because the two roles are thematically opposed. Saying that something is an artillery/boss or meatbag/mook doesn't give you any distinct information from the boss or mook side, it just tells you that they should be artillery++ and meatbag––.

If anything, bosses and mooks should be templates, not roles. Getting a bunch more actions/free ability activations/natural weapons/handy SLAs/immunities/etc. in exchange for a CR bump--and those would all be distinct templates, not one "Boss Monster" template; dragons have the Bunch O' Natural Weapons template, beholders have the Do Ten Things Per Round template, and so on--would work well to explicitly distinguish them from "normal" monsters and lets you cost different aspects of being a boss monster appropriately. The reverse would also work, of course, for giving minions the Glowing Weak Point template or Consistency Of Wet Cardboard template or Freaks Out And Runs After Getting Hit template or whatever else.

Not all complexity or broadess is inherently boss-worthy. [...] Those options do not inherently make a creature worthy of boss monster status.
[...]
On a similar note, fodder doesn't inherently benefit from being narrow and simple. For example, a hydra's fast healing makes it a bad fodder creature even though it is extremely narrow and simple, especially at higher CRs.

No arguments there. Complexity is necessary for good bosses and simplicity for good mooks, but not sufficient for them.

[Just to Browse]

Pair, I'm not fully convinced by your arguments, but I will admit that I see your point on the boss / fodder designations feeling like ++ / -- (but is it post- or pre-increment? ). I am also more in favor of turning [role] creatures into bosses with templates as opposed to consciously crafting a boss monster out of [role] class features, so our opinions may be more aligned than I initially thought they were. Given that, I would like to move ahead with more of the brainstorming process, and let the dice fall where they may (pun fully intended) with respect to how boss & fodder monsters get designed later on.

And on the topic of more brainstorming, I have another opinion which I think may be controversial. I would love to get some feedback on it.

Regarding Generic Classes

I do not like the idea of all, or even most, monsters being built primarily of a generic class. I have some reasons with examples for each:

Iconic Abilities
Some monsters have iconic abilities that should always be present in their stat blocks, but aren't compatible with certain combats. A Medusa is an example of this. While theoretically I can imagine a snake-hair archer at level 1, with a slow attack at level 3, poison at level 5, and a death gaze at level 7... I just don't see the value in labeling that creature a "Medusa". The same is true of the Bodak, Basilisk, Gorgon, etc. I think players that face medusae want to fight a creature that turns them to stone, so it's better to set the medusa at a minimum CR of 5-6 and give her class-style advancement options from there onwards.

Level-Based Life Spans
This qualm with a level-based system goes in the opposite direction. I think that some monster archetypes just have a limited range of levels in which they are interesting. For example, Vargouilles and Stirges are great low-level threats, but they don't handle well into the CR 10 range on their own, let alone the CR 20 range. I am comfortable with some creatures pokevolving, but not every concept goes to 20, and I believe D&D should gently encourage DMs to let go of low-level encounter designs when PCs get strong enough.

Optimization Weirdness
I actually tried something similar to this several years ago when I attempted to consolidate animals. When I was working on them, I realized that the way animals actually work and the way an optimized PC would work are very different. A CR 6 cat creature should be some kind of tiger or lion, and thus be large. But an actual high-dexterity pounce build would probably benefit more from being tiny and getting Dex to damage. It creates this weird Stormwind-esque situation where players have to choose between being a good lion or being a good pounce-attacker.

A generic monster progression with Roles / Type Abilities is at a very high risk of creating this kind of unfortunate decision point. I think players would be better off if their Lion class had mandatory class features up through level 6, and they weren't allowed to stay tiny to reap the benefits of size bonuses.

Solution?
So while I'm against the idea of large generic monster class tables, I'm not against the idea of monster class tables existing. I'm also kind of excited about the idea that some DMs can choose to have their Darkmantles evolve into Cloakers if they want.

What I would personally like to see is a 20-CR table of design benchmarks. People can pick a subset of levels for their monster, and use those to build out advancements.

For example, let's say you're making an Earth Elemental as a Meatbag class.

• The Meatbag CR table has a column labeled Average HP. It's 25 at level 1, 35 at level 2, 50 at level 3, and so on (not sure if those numbers are right or not)
• The description for Average HP has a few guides for how to get your meatbag monster to the proper HP. It has d12 HD, so you can do (2.5*CR) HD with Con -, or HD and DR/- equal to your CR, and some complicated math for Constitution.
• You choose the 2nd option for your elemental.
• By default, the guidelines suggest that your Meatbag has a 20' land speed so you add that to your elemental.
• As you build the monster's progression, you check each Meatbag level for suggestions. At level 1 the table just recommends something flavorful without a ton of power budget, so you put Earth Mastery and a burrow speed here.
• At level 2, the table recommends a mechanical distinction that keeps the monster's plodding nature, so you improve burrow to earth glide but add a stipulation that prevents the earth elemental from being a stealth master.
• At level 3, the table recommends a mild anti-PC defensive ability or an interesting tactic in combat. You add Push at this level, and improve it to medium size.
• At level 4, the table recommends whatever you didn't choose at level 3. You realize that SRD earth elementals are pretty boring to fight so you give it the ability to make difficult terrain within its reach.
• At levels 5 and 6, the table recommends upgrading existing features into higher-power ones. You give the elemental some stat boosts and make it large at level 5. You improve earth glide to the SRD version at level 6.
• At levels 7 and 8, the table recommends giving meatbags a way to intercept PCs and/or a defensive ability. You add some kind of super-gravity effect to stop fliers and give the elemental more DR specifically against ranged attacks.
• At levels 9 and 10, the table recommends keeping your number of racial features low and tailoring the creature to take up more space. You give the elemental more stat boosts (including +Speed) and make it huge at level 9
• But at level 10 you ignore the guidelines and give your elemental a capstone called Elder Elemental that grants some spell-like abilities. You want it to play sort of like a Mastermind at this level, so you give it summoning and battlefield control SLAs.

After all that, you write out a table of important stats, features, and feats by CR, and then release your Earth Elemental class into the wild.

Looking back on this fictional design process, hopefully it's easy to see that "class" features can be added, upgraded, swapped, or completely ignored. That level of freedom as a monster-homebrewer is what I really want. Instead of having Meatbag monsters get their Meatbag Ability I at level 1, I'd like a set of guidelines to let designers build experiences that excite the players at their table.

[RedWarlock]

You guys realize you are basically recreating the whole digression of 4e monster roles. Boss vs elite vs standard vs minion, especially with the 2nd-generation concept that minions (“filler” under your terminology) have standard role (brute/soldier/controller) in addition to BEING a minion.

4e may have lopped off outlier mechanics, but a good 70-80% of it is conceptually the same as it’s 3e forebears, and makes logical sense for what it tries to do.

[Just to Browse]

Quibble: the 4e equivalent of a minion is fodder in our particular usage.

I agree with you, it's definitely very close to 4e. The reason I kept the word "artillery" is because I picked it up from 4e and the name is great, and Pair actually used the word "Brute" as a class too (and "Harrier", which is close to Skirmisher). This is also reminiscent of Dungeonscape 3e's descriptive roles and encounter design template.

However, there's a distinct difference between the 3e and 4e version. 3e's approach was DEscriptive, in that the phase spider stat block was written in 2000, but it wasn't classified as Ambushers until 7 years later. 4e's approach was PREscriptive. So far we have kind of hybrid approach. Thus there are no Lurkers, because in D&D 3e stealth monsters don't usually pop out to harass PCs and then disappear, because the first time it happens they just... get killed.

[RedWarlock]

Whoops on fodder/filler. What's the distinction between fodder and filler, conceptually? 4e minions exist at all levels, whereas the fodder definition seems to be limited to stuff at CR 1/2 regardless of party level.

As for descriptive vs prescriptive, once you begin iteration on that analysis, what's the difference?

[PairO'Dice Lost]

Pair, I'm not fully convinced by your arguments, but I will admit that I see your point on the boss / fodder designations feeling like ++ / -- (but is it post- or pre-increment? ). I am also more in favor of turning [role] creatures into bosses with templates as opposed to consciously crafting a boss monster out of [role] class features, so our opinions may be more aligned than I initially thought they were. Given that, I would like to move ahead with more of the brainstorming process, and let the dice fall where they may (pun fully intended) with respect to how boss & fodder monsters get designed later on.

Yeah, it seems we have similar mental models on how monsters should look in the book and during play, and differences can probably be chalked up to 'brewing and DMing styles. Getting some concrete examples to cogitate on would certainly help.

I do not like the idea of all, or even most, monsters being built primarily of a generic class. I have some reasons with examples for each:
[...]
I think players that face medusae want to fight a creature that turns them to stone, so it's better to set the medusa at a minimum CR of 5-6 and give her class-style advancement options from there onwards.

Having certain CR minima for monsters isn't actually incompatible with using monster classes for advancement from the start.

Assuming a Medusa counts as a Puzzle monster for this example (the puzzle being how to deal with her with your eyes closed), you could certaily work up a hand-crafted 6-HD Medusa and then have her advance by taking levels in Puzzle Monster thereafter. However, you could also have "Debuffing Gaze Attack" as one of the available Puzzle Monster abilities that's only available to take around level 5 or 6, of which Petrification is one of the options for the gaze attack effect, then build a Medusa as a Puzzle Monster 6 and simply declare that you can't fight 2-HD or 4-HD Medusas unless you want to fight juvenile medusae before their gaze attacks come online, you child-killing monster.

The former approach definitely has the benefit of being able to customize any monster as desired without having to stick within certain parameters; on the other hand, monsters being thrown together without regard to stat ranges, numbers and strengths of abilities, and so forth is one of the reasons the "let's make monster classes" projects keep on popping up, so that's not necessarily a good thing. In contrast, the latter approach has a nice incidental side benefit: if every monster is built with classes, then you don't need separate PCs-as-monsters rules like level adjustment or savage progressions, players can just follow the actual MM progression for those monsters, detouring early or progressing past it as desired.

This qualm with a level-based system goes in the opposite direction. I think that some monster archetypes just have a limited range of levels in which they are interesting. For example, Vargouilles and Stirges are great low-level threats, but they don't handle well into the CR 10 range on their own, let alone the CR 20 range. I am comfortable with some creatures pokevolving, but not every concept goes to 20, and I believe D&D should gently encourage DMs to let go of low-level encounter designs when PCs get strong enough.

I agree that Stirges and Vargouilles shouldn't necessarily be assumed to pokevolve into Dire Stirges and Vargouille Lords at mid- to high levels, but I don't think that issue applies to the monster classes themselves. I can certainly imagine a mid-level monster that has a paralyzing roar, poison bite, and a transforming touch attack among its abilities such that the same monster class could be used to build a Vargouille and an Insert Mid-Level Monster Name Here, even if a literal Vargouille turning into a Vargouille Lord with the same stats as an IM-LMNH would be thematically incoherent.

A generic monster progression with Roles / Type Abilities is at a very high risk of creating this kind of unfortunate decision point. I think players would be better off if their Lion class had mandatory class features up through level 6, and they weren't allowed to stay tiny to reap the benefits of size bonuses.

Considering that existing monsters of most types get larger with more HD, I see no problem with requiring animals to be of certain sizes. But I also don't see a problem with allowing a DM to build a hypothetical 7- or 8-HD pouncing critter that's Small or smaller, if that's what he's going for.

There's definitely a tension between statting a given monster (which has certain mechanical and flavor constraints to "really be" or "really feel like" that monster) and creating monster classes (which should have few to no constraints to allow building all sorts of monsters out of them). It's much like writing up an arcane caster class: an abstract "Arcane Caster" class has basically no constraints aside from "at some point, must be able to cast some arcane spells," but writing up "Duskblade" vs. "Dread Necromancer" vs. "Wizard" all come with very different conceptual requirements and a given class feature could be useless for one and broken-overpowered for another based on its other class features, casting progression, spell list, and so forth.

What I would personally like to see is a 20-CR table of design benchmarks. People can pick a subset of levels for their monster, and use those to build out advancements.
[...]
Looking back on this fictional design process, hopefully it's easy to see that "class" features can be added, upgraded, swapped, or completely ignored. That level of freedom as a monster-homebrewer is what I really want. Instead of having Meatbag monsters get their Meatbag Ability I at level 1, I'd like a set of guidelines to let designers build experiences that excite the players at their table.

Agreed, flexibility is key. Again, though, I don't think that advice and benchmarks are actually incompatible with more strict Lego-block-style monster classes. Everything in the books is basically written as "Hey DM, here's a bunch of pre-written stuff you can use for convenience and standardization, but feel free to make your own!" and this can work the same way.

If you provide an Artillery Monster class that basically works like an Astral Construct ("at level X, add one Menu C ability to your monster" and Menu C is a fixed list of stuff that can be traded down for Menu B items), that lets a DM inexperienced with monster 'brewing following the instructions and come out with something roughly balanced. If you then provide some benchmarks and some sidebar commentary on why Menu C has the abilities that it does, why Menu C choices are available at level Y through level Z, and so forth, someone who's built a few Artillery monsters using that class can start to experiment with a few abilities and be assured that their choices fit with the general Artillery design philosophy, and eventually just throw together a monster without reference to that specific ability list because they can intuit what a Level X Artillery monster should look like.

You guys realize you are basically recreating the whole digression of 4e monster roles. Boss vs elite vs standard vs minion, especially with the 2nd-generation concept that minions (“filler” under your terminology) have standard role (brute/soldier/controller) in addition to BEING a minion.

4e may have lopped off outlier mechanics, but a good 70-80% of it is conceptually the same as it’s 3e forebears, and makes logical sense for what it tries to do.

Honestly, 3e monster class projects like this one and 4e monster roles have essentially nothing in common except the particular division of monster roles, and even those don't necessarily line up closely (see: Lurkers and Soldiers).

1) 3e monster classes try to determine what ranges of capabilities a monster will have at a given level, while 4e monster roles have very strict formulas for numerical stats an a "" for monster powers. A Kobold Wyrmpriest is a Level 3 Artillery (Leader) and a Spitting Drake is a Level 3 Artillery. Both have one basic melee attack and one Range 10 single-target energy attack, but nothing in the Artillery description tells you what the range/attack/damage/etc on those should be (which is why they differ for no particular reason), how many abilities are acceptable for a Level 3 Artillery monster (the Spitting Drake has just those two, the Wyrmpriest has 3 others, 1 obviously-Leader-based ability, 2 obviously-Kobold-based ability, and a dragon breath thrown in for flavor), or anything about base stats since their speed, AC, and other stats are all different.

The Hobgoblin Archer is also a Level 3 Artillery, but its artillery attack has a much longer range, attacks AC instead of Ref (with enough of a bonus that basically the same natural d20 roll is required to hit, of course), and has a kicker to boost an allied Hobgoblin Archer's attack. It also has a Hobgoblin-based ability, which is a defensive power where the Wyrmpriest's Kobold-based ability is a mobility power. Nowhere in the 4e books will you find any rationale or advice on why those differences occur or how you would build one of those from scratch with a hypothetical 4e Artillery monster class, while those things are the entire point of a 3e monster class system, because the design philosophies are so different.

2) 4e Solos and Minions are locked into the same numerical ranges as their normal counterparts, but again there's no rhyme or reason to their powers. A Level 1 Kobold Minion can stab someone with a javelin (or throw it at them) and shift 1 square as a minor action, while a Level 14 Kuo-Toa Guard can...stab someone with a spear and shift 1 square as a move action as long as it stays next to its target, and a Level 16 Ogre Bludgeoneer has no movement power and doesn't even have something like the Kuo-Toa's swim speed to add variety. Power-wise, those minions have actually gotten weaker as their level increases.

You're right that MM2 and onward addressed that glaring oversight, but again there's no logic or standardization by primary role. A Level 23 Minion Skirmisher Angel of Light has a death throes that damages enemies and buffs allies, the Level 5 Minion Soldier Rupture Demon has a death throes that gives a stacking buff to a single ally (and can trigger that power when it grabs an enemy, yet the power does nothing specifically to that enemy for some reason). What makes "buffing death throes" a Skirmisher or Soldier power, and why does a Skirmisher harm its enemies but a Soldier doesn't? The answer is, once again, "".

A 3e equivalent of minion-type monsters should absolutely figure out how its minion-ness relates to its primary role, and then minion-izing it should simplify it, make it better in large numbers, and so forth without diluting the concept or making it feel tacked-on.

Whoops on fodder/filler. What's the distinction between fodder and filler, conceptually? 4e minions exist at all levels, whereas the fodder definition seems to be limited to stuff at CR 1/2 regardless of party level.

Fodder is equivalent to the 4e Minion, where they come in large numbers and are more of an annoyance than a threat. A Filler monster is one that actually is a threat, but doesn't have a very specific role, just plays well with other monsters of any role and balances out the action economy--the Mario of monster roles, if you will.

As for descriptive vs prescriptive, once you begin iteration on that analysis, what's the difference?

A descriptive approach starts with a bunch of monsters made out of arbitrarium and tries to retrofit a system onto them to advise DMs on how to use them. A prescriptive system starts with a bunch of roles and concepts and builds monsters to those specifications with the intention that the end result gives a specific set of desired outcomes. This attempt is a hybrid approach, as JtB mentioned, because it's trying to build a prescriptive system to recreate existing monsters and is thus constrained by the existing descriptive system.

[Just to Browse]

Assuming a Medusa counts as a Puzzle monster for this example (the puzzle being how to deal with her with your eyes closed), you could certaily work up a hand-crafted 6-HD Medusa and then have her advance by taking levels in Puzzle Monster thereafter. However, you could also have "Debuffing Gaze Attack" as one of the available Puzzle Monster abilities that's only available to take around level 5 or 6, of which Petrification is one of the options for the gaze attack effect, then build a Medusa as a Puzzle Monster 6 and simply declare that you can't fight 2-HD or 4-HD Medusas unless you want to fight juvenile medusae before their gaze attacks come online, you child-killing monster.

The problem here is that the gaze attack is supposed to be central to what a Medusa is (and what a CR 4 Basilisk is, and a CR 8 Bodak, etc). If you want a good puzzle monster class, it needs to make good puzzle monsters at CRs 1-20. This means the Medusa "Build" is going to have level 1-7 features from that Puzzle Monster class. If a Death Gaze is a Puzzle III class feature, then a Medusa is going to have Puzzle I and Puzzle II, as well as Racial I/II/III class features. But those extra features don't make the Medusa a better puzzle encounter. I'd argue they go the opposite direction, because they add a bunch of noise.

There are ways around this problem, like writing Puzzle I: Sickening Gaze, Puzzle II: Slow Gaze, Puzzle III: Stone Gaze, all of which require the previous feature. But now you can't write a CR 4 Basilisk which trades mobility & range for -2 CR unless you create a new oddball class feature like Puzzle II: Stone Gaze (No Ranged Attacks Allowed).

On the other hand, a set of guidelines allows designers to ignore hardcoded class features in favor of clean designs. The puzzle monster table can have a line like "Puzzle monsters generally should not have more than 3 features that PCs would need to be aware of." Guidelines also allow designers to put their foot down on monster options that would be inappropriate for PCs of the same level, like a Level 6 PC with Petrifying Gaze.

I agree that Stirges and Vargouilles shouldn't necessarily be assumed to pokevolve into Dire Stirges and Vargouille Lords at mid- to high levels, but I don't think that issue applies to the monster classes themselves. I can certainly imagine a mid-level monster that has a paralyzing roar, poison bite, and a transforming touch attack among its abilities such that the same monster class could be used to build a Vargouille and an Insert Mid-Level Monster Name Here, even if a literal Vargouille turning into a Vargouille Lord with the same stats as an IM-LMNH would be thematically incoherent.

I'm not so much questioning whether it could be done, but whether it should be done. If Vargouilles and Stirges got a dedicated 1-8 CR guideline table, it would be clear to designers that those concepts are meant to lose their value at or before CR 8. Alternatively, showing designers how Vargouilles can be built from a 20-CR table falsely indicates that the Vargouille model of encounter design makes for interesting combats at CR 10, whereas our IM-LMNH shouldn't need to chain SoLs together or get into a grapple.

Considering that existing monsters of most types get larger with more HD, I see no problem with requiring animals to be of certain sizes. But I also don't see a problem with allowing a DM to build a hypothetical 7- or 8-HD pouncing critter that's Small or smaller, if that's what he's going for.

There's definitely a tension between statting a given monster (which has certain mechanical and flavor constraints to "really be" or "really feel like" that monster) and creating monster classes (which should have few to no constraints to allow building all sorts of monsters out of them). It's much like writing up an arcane caster class: an abstract "Arcane Caster" class has basically no constraints aside from "at some point, must be able to cast some arcane spells," but writing up "Duskblade" vs. "Dread Necromancer" vs. "Wizard" all come with very different conceptual requirements and a given class feature could be useless for one and broken-overpowered for another based on its other class features, casting progression, spell list, and so forth.

Note that I am not talking about DMs, but players. DMs can struggle with this too, but players are strongly encouraged to maximize the value of their few character resources. If a PC must choose between Large + Pounce or Small + Pounce, they are choosing between being a good lion or being a good contributor to combat. I don't like the idea of asking players to make those choices if I can help it, and I think class guidelines will force players to make that choice a lot, since the number of class features for each level can be enormous and thus weirdly-flavored synergies will abound.

Having guidelines for monster design instead of a rigid set of class features solves this problem because it asks you to use good judgment. To use the arcane caster example, it's better to design a Dread Necro and let players use it, than to design a Sorcerer and ask players choose between flavorful spells (command undead) or useful spells (glitterdust).

If you provide an Artillery Monster class that basically works like an Astral Construct ("at level X, add one Menu C ability to your monster" and Menu C is a fixed list of stuff that can be traded down for Menu B items), that lets a DM inexperienced with monster 'brewing following the instructions and come out with something roughly balanced. If you then provide some benchmarks and some sidebar commentary on why Menu C has the abilities that it does, why Menu C choices are available at level Y through level Z, and so forth, someone who's built a few Artillery monsters using that class can start to experiment with a few abilities and be assured that their choices fit with the general Artillery design philosophy, and eventually just throw together a monster without reference to that specific ability list because they can intuit what a Level X Artillery monster should look like.

Guidelines can do this too. If you have a document that says "At level X, your monster should usually have features A, B, and C" and then there's a list of possible features, an inexperienced designer can follow those instructions. Monster design may even be easier for an inexperienced DM this way, because guideline tables can have rows like "Recommended Total HP" and "Recommended Save Bonus" which novices can copy/paste into a stat block instead of picking a Constitution score. Guidelines also make it easier to experiment, because now there is no mandate that Stone Gaze is a Menu C (level 6+ only) ability when a Basilisk with stone gaze is fine at CR 4.

[PairO'Dice Lost]

There are ways around this problem, like writing Puzzle I: Sickening Gaze, Puzzle II: Slow Gaze, Puzzle III: Stone Gaze, all of which require the previous feature. But now you can't write a CR 4 Basilisk which trades mobility & range for -2 CR unless you create a new oddball class feature like Puzzle II: Stone Gaze (No Ranged Attacks Allowed).

Should that be a goal, though, to allow any monster ability at any HD range? Petrification is definitely a "must be this tall to play" kind of monster ability, and there's not much space to squeeze in a variation on petrifying gaze at CR 4 when the Cockatrice at CR 3 only has a petrifying bite and even that's pretty strong against expected Fort modifiers at that level. I see no problem with saying that a petrifying gaze is a 6-HD-and-up ability and no amount of concessions would lower that bar, because no matter how slow/short-ranged/etc. the critter is it is still a walking AoE petrification effect with at least a 1-in-20 chance of character removal per sighted PC per round.

Now, that doesn't mean that this hypothetical Mini-Basilisk couldn't take Slow Gaze and flavor it as a Partial-Petrification Gaze or whatever, or that there couldn't be a weaker 5e-style Slow And Also Eventual Petrification If You Fail A Bunch Of Saves Gaze available at low HD, or whatever. But I think that giving monsters abilities that punch far above their CR and then trying to artificially lower their CR by giving them a bunch of related drawbacks is like trying to weaken the Wizard class by removing its familiar: yeah, it's a noticeable drawback in situations where having a familiar would be really handy, but it doesn't affect the power of its spells one bit and those are what really matter. That kind of design didn't work for That Damn Crab or the Adamantine Clockwork Horror, and it wouldn't work for a Mini-Basilisk or Slightly Lighter Shadow.

On the other hand, a set of guidelines allows designers to ignore hardcoded class features in favor of clean designs. The puzzle monster table can have a line like "Puzzle monsters generally should not have more than 3 features that PCs would need to be aware of." Guidelines also allow designers to put their foot down on monster options that would be inappropriate for PCs of the same level, like a Level 6 PC with Petrifying Gaze.

Well, firstly, the medusa is literally a 6-HD creature with a Petrifying Gaze, so you can't both keep that and also let PCs just take monster classes as if they were PC classes.

But secondly, those guidelines are things you can put right into the class progression. If you think Puzzle monsters should have no more than 3 signature abilities, you can make a class table that looks something like this:

Level	Feature
1st	Ability A I
2nd
3rd	Ability B I
4th
5th	Ability C I
6th
7th	Ability A II
8th
9th	Ability B II
10th

...and so forth, requiring those monsters to upgrade their signature abilities as they level instead of acquiring new ones. And if a DM wants one particular Puzzle monster to have more, well, the same "DMs, feel free to ignore these hard rules and customize things" advice applies in both cases.

Basically, I feel that if you're going to all the trouble of changing all the monster types, HD base stats, and so forth to the point that you can't use any published monsters as written, there's no point in just writing up rough guidelines about how "Here's how we intend Mastermind monsters to be built, stick with these guidelines and the end result will probably match those intentions" and leave it at that when you can just make a Mastermind monster class and say "These are the rules encoding all of our assumptions about Mastermind monsters, build a monster using this class and you'll definitely follow them."

Note that I am not talking about DMs, but players. DMs can struggle with this too, but players are strongly encouraged to maximize the value of their few character resources. If a PC must choose between Large + Pounce or Small + Pounce, they are choosing between being a good lion or being a good contributor to combat. I don't like the idea of asking players to make those choices if I can help it, and I think class guidelines will force players to make that choice a lot, since the number of class features for each level can be enormous and thus weirdly-flavored synergies will abound.

Having guidelines for monster design instead of a rigid set of class features solves this problem because it asks you to use good judgment. To use the arcane caster example, it's better to design a Dread Necro and let players use it, than to design a Sorcerer and ask players choose between flavorful spells (command undead) or useful spells (glitterdust).

The first paragraph here implies that you're expecting "PCs can be monsters using monster classes" to mean that a player would ask to play a generic "monster" and the DM would have them assemble a custom monster from scratch with all the dumpster-diving that implies.

But the second paragraph, and the way I've been assuming this would be used, implies that a player would ask to play a lion or a succubus or whatever, the DM would point them to the little block in the MM entry (or whatever the DM creates for a homebrew monster) that says "Here's a [Blah]'s base stats at 1st level and here's which monster abilities they take in which order," and the player would perhaps be able to make a few ACF-style substitutions but would otherwise follow that individual monster progression just like they'd follow a Rogue or Warmage progression (with multiclassing, going past the progression, and so forth obviously still being options).

So I'm not really seeing the conflict here. A player wouldn't be agonizing over how to synergize things with pounce any more than they'd agonize over how to build a Lion Totem barbarian; they can pick feats and grab items and so forth to help out, but the progression is largely fixed and they were well aware of that before they decided to play one.

Guidelines can do this too. If you have a document that says "At level X, your monster should usually have features A, B, and C" and then there's a list of possible features, an inexperienced designer can follow those instructions. Monster design may even be easier for an inexperienced DM this way, because guideline tables can have rows like "Recommended Total HP" and "Recommended Save Bonus" which novices can copy/paste into a stat block instead of picking a Constitution score. Guidelines also make it easier to experiment, because now there is no mandate that Stone Gaze is a Menu C (level 6+ only) ability when a Basilisk with stone gaze is fine at CR 4.

I'm very much against a "monster base numbers are completely fixed by level and ability scores/size mods/etc. have basically no effect on anything" setup, because I played through AD&D where monsters were built like that and putting a belt of cloud giant strength on an ogre or troll gave you a Divide By Cucumber error because their ability scores and the way they interacted with other stats were largely undefined. 4e got away with that only because all of those spells and items were removed and monsters were basically cardboard cutouts that popped into existence for combat and didn't want you to think too hard about how they work, but that won't work for 3e.

Much better, I think, to have standard ability score progressions for different classes (e.g. "Given the standard array, a Meatbag monster puts its highest score in Con and increases that at every opportunity") and make it explicit that any average-whatever-by-level numbers assume those scores and the DM is free to play around with those as desired. It avoids the problem of having totally different mechanics for newbies that don't match the actual rules you expect them to use when they gain more experience, since then they have to make a singular and explicit jump as opposed to gradually getting more comfortable with the rules as-is and modifying them as they feel comfortable doing so.

[Just to Browse]

Should that be a goal, though, to allow any monster ability at any HD range? Petrification is definitely a "must be this tall to play" kind of monster ability, and there's not much space to squeeze in a variation on petrifying gaze at CR 4 when the Cockatrice at CR 3 only has a petrifying bite and even that's pretty strong against expected Fort modifiers at that level. I see no problem with saying that a petrifying gaze is a 6-HD-and-up ability and no amount of concessions would lower that bar, because no matter how slow/short-ranged/etc. the critter is it is still a walking AoE petrification effect with at least a 1-in-20 chance of character removal per sighted PC per round.

Not any HD range, but lower than would be appropriate for most monsters. Like we've been mentioning, the basilisk and other death gaze creatures are generally considered puzzle monsters. A basilisk is appropriate for its CR because it has -1 Init, a single unimpressive melee attack, a 20' move speed, and tactics that include giving "a half-hearted pursuit" to any PCs that run away / regroup when their strategy fails. In contrast, a Medusa is +2 CR because it has a shortbow, 30' move speed, and supposedly a high enough Bluff check to get it into gazing range. This sort of design should be done regularly. If your monster-generation method doesn't allow designers to trade abusable weaknesses for early access to scary monster powers, I would argue it fails to capture an important element of monster design.

Well, firstly, the medusa is literally a 6-HD creature with a Petrifying Gaze, so you can't both keep that and also let PCs just take monster classes as if they were PC classes.

But secondly, those guidelines are things you can put right into the class progression. If you think Puzzle monsters should have no more than 3 signature abilities, you can make a class table that looks something like this:

[snip]

...and so forth, requiring those monsters to upgrade their signature abilities as they level instead of acquiring new ones. And if a DM wants one particular Puzzle monster to have more, well, the same "DMs, feel free to ignore these hard rules and customize things" advice applies in both cases.

I think it's good that PCs can't just take monster classes all the time, and I am confident that a playable level 6 Medusa would trivialize many low- and mid-level encounters. Even if you do create a set of menu abilities, some of them just aren't going to be appropriate for PCs (like level 6 death gaze).

To the table: if there are only class features every other level, then you create a weird system where puzzle monsters at CR 1 & 2 / 3 & 4 / 5 & 6 are close in power level, whereas 2 & 3 / 4 & 5 / etc see enormous power jumps. If level 4's feature is bonus HD and a stat boost, while level 5's feature is "death gaze", your designers and players are going to have a bad time. If we need to craft our design tool so that it creates wonky outputs for 50% of all CRs, then I think our design tool isn't useful.

But the second paragraph, and the way I've been assuming this would be used, implies that a player would ask to play a lion or a succubus or whatever, the DM would point them to the little block in the MM entry (or whatever the DM creates for a homebrew monster) that says "Here's a [Blah]'s base stats at 1st level and here's which monster abilities they take in which order," and the player would perhaps be able to make a few ACF-style substitutions but would otherwise follow that individual monster progression just like they'd follow a Rogue or Warmage progression (with multiclassing, going past the progression, and so forth obviously still being options).

A monster class, like all classes, is part of the social contract at the table. That social contract has long established that class features are balanced and interchangeable when offered at equal levels. That is true (or supposedly true) for all classes with select-able class features... except monster classes now. This is unlike the warmage, because the features of the warmage are all available to you from the moment you crack open Complete Arcane.

Attempting to thread the needle here gives you a kind of pseudo-class. Players can't just use it to build their character, the only people who get free reign over it are DMs / designers, and the class needs cautionary warnings so that you don't combine 2 synergistic features to build an overpowered monster. That doesn't sound like a monster class to me. It sounds like a set of design guidelines. I'd much rather lean into these being guidelines and do away with the rigidity of pseudo-class-features.

I'm very much against a "monster base numbers are completely fixed by level and ability scores/size mods/etc. have basically no effect on anything" setup, because I played through AD&D where monsters were built like that and putting a belt of cloud giant strength on an ogre or troll gave you a Divide By Cucumber error because their ability scores and the way they interacted with other stats were largely undefined. 4e got away with that only because all of those spells and items were removed and monsters were basically cardboard cutouts that popped into existence for combat and didn't want you to think too hard about how they work, but that won't work for 3e.

Indeed, it's important that these be "Recommended", and not "Fixed No Matter What". That's why I called out the stats as "Recommended Total HP" / "Recommended Save Bonus", and indicated that novices (not experts) would be the ones doing the copy/pasting.

I do not like the idea that monsters need a specified stat array with +1 bonuses every 4 levels like a PC, because attributes don't serve the same purpose for PCs & monsters. On PCs, attributes are a method of sharpening identities and encouraging certain overlaps in gameplay. On monsters, attributes are just a tuning lever. For the small percentage of the time that an earth elemental gets played as a PC, you can have a premade set of stat boosts written out as class features in your earth elemental class. Monster classes have done this for over a decade now without confusion, so it's a pretty safe bet.

[Morphic tide]

I think there's too much focus on the idea of having the creatures be rendered into monster classes, rather than starting with the core of reworking Racial Hit Dice. Start with working out the mechanical niches that need their own chassis. For example, a Mastermind could have low health, high skills, poor BAB and only a good Will save, while the similar Lurk/Assassin could have a touch more health, good BAB and a good Reflex save and a lot less skill points, the difference between a quick-but-touchy fight and an indirect encounter. Meanwhile, a Controller could have poor BAB, high health, decent skills and saves for days, being a "keystone" monster that's hard to take out so that its minions aren't trivially removed from the fight and ideally need to be taken out first before focusing on the summoner becomes worhtwhile. Then figure out properties for the subtypes along similar lines as at least partially chassis adjustments. Demote Undead and Construct to subtypes as part of this, as they have enough mechanical impact to come with some backed-in downsides to the chassis.

If we're going to construct highly class-like organization schemes, then they should be balanced to be comparable to actual PC classes, with exception of severe outliers of non-PC-able abilities like Spawn creation. Given that we'd be redoing the whole Monster Manual alongside this, the biggest offenders can be altered as necessary, such as having a Leadership-like model for acceptable amounts of minions, and many health recovery mechanics could be made to be based on Lay on Hands, where there's a limit to how much health is recovered in a day, so that Regeneration isn't a pass to ignore health damage altogether, but rather one part damage mitigation (due to the nonlethal damage conversion), one part emergency not-dying power, one part trickle from an extra-but-limited HP pool.

Also, Gaze-that-slows-and-eventually-petrifies is actually perfectly viable, since you can always balance it to have it be unlikely-to-the-point-of-impossible inside a normal combat period. Like, 10-20% slow per failed save, recover 5-10% of that each round, some other penalties and maybe some conditional benefits relating to the fact that the character is becoming made of a more durable material. Could use this sort of thing as the general petrification mechanic for this rework, so the stuff that's automatic petrification is 100% on a failed save and the stuff that already gets to direct petrification, when used by PCs, ramps up in effectiveness over time. A partial petrification mechanic also allows for partial petrification on successful saves, so that a Gorgon isn't getting stomped completely because they had the misfortune of dealing with an all-Meldshaper party touting an A-Game Paladin for a healer, as they still have a good debuff to hold out if all the saves succeed.

Similarly, other major problem cases can be balanced with similar added granularity, such as Energy Drain having small amounts of ability damage when the character passes the save, allowing for the presence of lower-level monsters with such an effect. The existing secondary damage poison mechanic can also do a lot for this sort of thing, as it generally will tick after a fight's over. Overall, we could make a lot more things granular enough to be present at low levels, giving a lot more in the way of making low-level monsters into interesting-but-fair encounters.

[Durzan]

I think there's too much focus on the idea of having the creatures be rendered into monster classes, rather than starting with the core of reworking Racial Hit Dice. Start with working out the mechanical niches that need their own chassis. For example, a Mastermind could have low health, high skills, poor BAB and only a good Will save, while the similar Lurk/Assassin could have a touch more health, good BAB and a good Reflex save and a lot less skill points, the difference between a quick-but-touchy fight and an indirect encounter. Meanwhile, a Controller could have poor BAB, high health, decent skills and saves for days, being a "keystone" monster that's hard to take out so that its minions aren't trivially removed from the fight and ideally need to be taken out first before focusing on the summoner becomes worhtwhile. Then figure out properties for the subtypes along similar lines as at least partially chassis adjustments. Demote Undead and Construct to subtypes as part of this, as they have enough mechanical impact to come with some backed-in downsides to the chassis.

If we're going to construct highly class-like organization schemes, then they should be balanced to be comparable to actual PC classes, with exception of severe outliers of non-PC-able abilities like Spawn creation. Given that we'd be redoing the whole Monster Manual alongside this, the biggest offenders can be altered as necessary, such as having a Leadership-like model for acceptable amounts of minions, and many health recovery mechanics could be made to be based on Lay on Hands, where there's a limit to how much health is recovered in a day, so that Regeneration isn't a pass to ignore health damage altogether, but rather one part damage mitigation (due to the nonlethal damage conversion), one part emergency not-dying power, one part trickle from an extra-but-limited HP pool.

Also, Gaze-that-slows-and-eventually-petrifies is actually perfectly viable, since you can always balance it to have it be unlikely-to-the-point-of-impossible inside a normal combat period. Like, 10-20% slow per failed save, recover 5-10% of that each round, some other penalties and maybe some conditional benefits relating to the fact that the character is becoming made of a more durable material. Could use this sort of thing as the general petrification mechanic for this rework, so the stuff that's automatic petrification is 100% on a failed save and the stuff that already gets to direct petrification, when used by PCs, ramps up in effectiveness over time. A partial petrification mechanic also allows for partial petrification on successful saves, so that a Gorgon isn't getting stomped completely because they had the misfortune of dealing with an all-Meldshaper party touting an A-Game Paladin for a healer, as they still have a good debuff to hold out if all the saves succeed.

Similarly, other major problem cases can be balanced with similar added granularity, such as Energy Drain having small amounts of ability damage when the character passes the save, allowing for the presence of lower-level monsters with such an effect. The existing secondary damage poison mechanic can also do a lot for this sort of thing, as it generally will tick after a fight's over. Overall, we could make a lot more things granular enough to be present at low levels, giving a lot more in the way of making low-level monsters into interesting-but-fair encounters.

Of course, whatever we do, granular progressions and stuff should be fairly simple and consistent in concept and design. Uniformity is one thing that can make a game easier to learn, play, and have fun with.

The main thing here is the separation of creature types from HD and the stats that are derived from them is the primary goal. Having the more common monster chassises be refluffed into monster classes gives more uniform and easier to deal with customization for the DM, but trying to alter monster abilities into "monster class features" is a bit much for me. With exception of some of the more common abilities that could probably reworked into something akin class features, monsters in 3.5 are a bit too varied in their powers and abilities to be able to adequately account for most monsters.

Personally, I'd just create the monster chassis essentially be monster classes with little to no class features, and leave it at that... at least at first. In essence I would treat Monster abilities as racial features, and if I need to determine at what "level" they come online, I'd determine how much HD an adult monster of that species has and then reverse-engineer progression from there.

Only once we had the common chassis for many monsters, and figured out any patterns in special abilities among them, would I consider trying to work those abilities into the monster classes themselves. Thats the next step in this process

[Just to Browse]

Personally, I'd just create the monster chassis essentially be monster classes with little to no class features, and leave it at that... at least at first. In essence I would treat Monster abilities as racial features, and if I need to determine at what "level" they come online, I'd determine how much HD an adult monster of that species has and then reverse-engineer progression from there.

I'm strongly in favor of an approach like this. Monsters should be foes first and PC options second. Reverse-engineering an advancement progression after you build your critter is the way to go IMO.

[Bohandas]

As was touched on in this thread, having monster type and HD inextricably linked creates problems. So lets fix this. While we're at it, lets collab on reworking monsters from the ground up.

I think an even bigger problem is the inextricable linking between monster HD, monster skillpoints, and monster BAB