How many monsters does a D&D edition need to start with?

[Ice9]

Oh, those dragon charts - the charts that were enough of a pain in the ass that they released a whole book of completed stat-blocks for dragons.

Let's see - assuming we're trying to bluff an Adult Black Dragon. I'll just look up the Sense Motive and ... oh. Ok, so I have to figure it out:
* Consult Black Dragon chart - HD is 19, Int is 12, Wis is 13
* Below the chart - Black Dragons have Hide, Move Silently, and Swim as class skills
* Up to the skills section - Dragons have 6+Int skill points and 9+N class skills (N = # of knowledge skills), three of which they almost always max, plus the three specific to Black Dragons. So, this particular Black Dragon may or may not actually have Sense Motive. We'll be fair and do it randomly: 4 free skill choices / 12 skills (only going to count the three best Knowledges) = 1/3 chance of having Sense Motive. This is assuming we stick to maxed out skills for simplicity.
* Ok, so the Dragon has either +1 or +22.
* Of course, said Dragon also has 7 feats, so it's possible that another +2 or +3 should be added to the total.
* That's without treasure, of course. I'm aware that most current MMs don't include which treasure is being used by the creature, but if we're talking about an ideal format, I would in fact include it as at least a sidebar.

That ... is too many steps.

[K]

Oh, those dragon charts - the charts that were enough of a pain in the ass that they released a whole book of completed stat-blocks for dragons.

Let's see - assuming we're trying to bluff an Adult Black Dragon. I'll just look up the Sense Motive and ... oh. Ok, so I have to figure it out:
* Consult Black Dragon chart - HD is 19, Int is 12, Wis is 13
* Below the chart - Black Dragons have Hide, Move Silently, and Swim as class skills
* Up to the skills section - Dragons have 6+Int skill points and 9+N class skills (N = # of knowledge skills), three of which they almost always max, plus the three specific to Black Dragons. So, this particular Black Dragon may or may not actually have Sense Motive. We'll be fair and do it randomly: 4 free skill choices / 12 skills (only going to count the three best Knowledges) = 1/3 chance of having Sense Motive. This is assuming we stick to maxed out skills for simplicity.
* Ok, so the Dragon has either +1 or +22.
* Of course, said Dragon also has 7 feats, so it's possible that another +2 or +3 should be added to the total.
* That's without treasure, of course. I'm aware that most current MMs don't include which treasure is being used by the creature, but if we're talking about an ideal format, I would in fact include it as at least a sidebar.

That ... is too many steps.

You have to use something like Pathfinder's skill system. No monster creation system would ever pass the "fast" or "easy" metric using 3.x skills.

Then you can just put it into the chart.

[Ice9]

I was assuming max ranks rather than allocating individual skill points anyway. The PF dragon would have 6 free picks out of ~11 skills (~55% chance for Sense Motive), but otherwise the process is the same.

You could put skill totals and feats into the chart, as well as other numbers one might care about, but at that point you're getting close to just having an oddly formatted series of stat-blocks.

[K]

I was assuming max ranks rather than allocating individual skill points anyway. The PF dragon would have 6 free picks out of ~11 skills (~55% chance for Sense Motive), but otherwise the process is the same.

You could put skill totals and feats into the chart, as well as other numbers one might care about, but at that point you're getting close to just having an oddly formatted series of stat-blocks.

The ideal system is:

1. Check chart and see skill ranks and skills for that level.
2. No second step.

Feats that modify pure numbers need to not be in any new edition of the game.

[Foxwarrior]

You mean, every dragon is equally good at lying?

[K]

You mean, every dragon is equally good at lying?

Yes, unless the DM chooses powers or feats that make it better at lying. For example, there are feats in 3.x that give social abilities not by adding numbers.

[tussock]

That's really simple in 4e, where every anyone is equally good at everything, unless the GM chooses to make it Hard or Easy. But that's not what you're arguing for, K, so never mind.


Monster generation in general though, 4e used that, and it was crap explicitly because the monsters used the generation system. Everything generated the same numbers and if you could beat one you could beat them all. If you cant beat one, you can't beat any. Then it told you to add arbitrary amounts of action denial without changing the reward, because that ... I don't know.

3e too, though it's numbers were from HD and monster-type and size and class equivalency and maybe not-quite-a-sorcerer and as many templates and class levels and ..., and then you handwave a CR. It might not look like a system at all, but it totally was and they really tried to make monsters stick to it, which made a lot of them have terribly inappropriate powers and numbers, even though they just remade functional AD&D monsters with it.



Hell, the only way to check your generator works at all is to make the classic monsters with it and test those out, whereon you may as well print them and take people's money for saving them the work of doing that all over again.

[Cyberzombie]

I used to think this would be the way to do it. Screw all this clutter! Make progressions, templates, cross-referenced ability packages. You can fit so much more in so much less space, and all it takes is a little bit more prep.

Now I've been running a game for a bit, and my current feeling on that is "Fuck no". Monsters that require me to add any component parts together, beyond what I can do in my head on the fly (such as commonly used spells, but not obscure ones), are only coming out for special occasions, and can fuck right off the rest of the time.

It's not only that "a little more prep" adds up really fast when you're talking about 2-3 monsters per situation for 10+ situations, half or more of which you may not end up using. Although that is definitely a factor.

It's that sometimes you need a monster on the fly, because the players did something unexpected and now there's a Troll in a party hat rampaging around at the duke's ball. When that happens, a MM with full stat-blocks means I can just flip to the page and keep the action moving. Procedurally generated stats means either "Hey guys, time out for 20 minutes while I prep some monsters, which may only be relevant for 1 round or so" or "Nope, you can't go off the rails, it's too much damn work".

I liked 3E, but one of it's biggest flaws is monster and especially NPC design. There were so many good options for going off the rails, but the speed of coming up with NPC stat blocks strongly encouraged railroading to the point where the PCs were kind of afraid to do anything unorthodox since it meant a big break in the action as the DM has to prepare a combat he never expected.

The amount of work it involved to make NPCs/complex monster lead DMs to underprepare, which added to the problem. Given it was going to take a half hour to create an NPC, the DM rarely even bothered unless it was something he was reasonably sure his PCs would attack.

It was one of the larger failures in game design in 3E. NPC/Monster design needs to drop all the clutter and give the DM some effective tools to create encounters on the fly.

[Username17]

Oh, those dragon charts - the charts that were enough of a pain in the ass that they released a whole book of completed stat-blocks for dragons.

Yeah, the Dragon Charts are a bad way to convey information. There are arguments to be made in favor of various ways to arrange information on a monster entry, but I can't imagine any good arguments for something that looks especially similar to that one. Fundamentally, you don't usually need to compare higher and lower level versions of the same monster, so putting the 15th level version right next to the 12th level version is just a way to confuse people. Dense charts with numbers for different levels of creature may be of some use to a monster deisgner, but they are pretty amazingly unfriendly to try to use in the middle of play.

When it comes down to it, most monster books use a format very similar to that of the AD&D Monster Manual. Because it's actually a very good format. You have the name of the monster, a series of important game stats arranged vertically, and then a textual description of what it is and how it interacts with its environment (combat and otherwise). It's very hard to improve on that. And when people flippantly tell me that they can improve on it by making dense tables of numbers, I am incredibly skeptical. No you fucking can't. Because dense tables of numbers are not very usable and don't convey a lot of important information like why you should care about this monster.

I'm sure you can cram a whole stat line into a line of text or two and then cram monster stat lines 50 to a page, but you don't have any monsters at that point, you just have a list of stats. Essentially, you've just made the 4e monster builder. You could do a better job than they did, but you're still generating gelatinous cubes with big reflex defenses because they are high level or something equally dumb.

We walked through the full reductionism model. It's not going to fly. There aren't enough spaces in 100 entries to even fit in all the shit that players need to summon, befriend, or ride. Even 10 levels of play requires a very large number of real ready-to-eat monsters. You could certainly trim a bunch of stupider monsters from the roster, but the roster has thousands of monsters in it to begin with.

---

But enough about why the reductionist model is really obviously a bad idea. Let's talk about what you'd actually put in your Monster Manual.

For starters, while the game should probably launch going to 10th level, the monsters should go to 15. Because you need boss monsters and stuff for you to aspire to defeat. There should be more monsters at the lower levels, and less monsters at the higher levels.

OK, let's talk about summons. You need 12 elementals (S, M, L for summoners of different levels). 16 would be acceptable. You also want Good, Evil, and Indeterminate summonables for 10 levels. Probably at least three choices of each for each level and of each alignment. So your 7th level Demonologist would have three choices for a summoned mook and your 7th level Angel Summoner would also have three choices and they'd be different choices. So that's 90 options there. We're very noticeably past the 100 mark just making the ready-to-eat monsters for player character summoners.

-Username17

[Desdan_Mervolam]

I agree with Frank. Trying to customize monsters is always time consuming to the point where it's only worth doing if it's a tentpole encounter to the campaign, otherwise you're spending lots of time on an encounter that will be over in a few rounds and won't be remembered by the party after a few sessions. Best to have a wide variety of interesting monster that you can pick from and are viable off-the-rack in case you need to improvise something.

[MGuy]

I'm going to need someone to explain to me why we seriously can have both a functional, relatively easy to use, monster generation system and then a bunch of generated monsters motherfuckers can simply reference when they are feeling too lazy/are too short on time to stat up a monster. You get the best of both fucking worlds. You have your made monsters then if you feel like changing them you can slap a bunch of templates and abilities on them with an easy guide on how you do just that. If you DON'T feel like doing that then you can seriously just lift the already pre-generated monsters right from the book. You get all your icons + the tools you need to do what the fuck ever.

[Ancient History]

Well, technically they do have that. It's just that the monsters are called NPCs. In GURPS, for example, you can set a point limit and then grab your books and make as detailed and fine-tuned an NPC as you want, or you can just slap a bunch of templates on a basic guy until the points add up about right. Hell, that was half the point of my PACKS system. In D&D3 you had a relatively easy time slapping a template or couple class levels on NPCs.

As for building monsters...y'know, I've never quite seen a good system. D&D3 sort of attempted it with some of their chimeric monster templates and stuff, but it was fairly labor-intensive. It's the sort of number-crunching where a computer program most comes in handy.

[OgreBattle]

D&D3e classed them by race, with the problems of succubi temptresses and tentacled horrors being the smae class

D&D4e classed them by role, which makes more sense in theory, but the execution...

Is there a way to make it work though? Brutes and Skirmishers seem distinct enough, one walks up to you to kill you and the other moves behind you to kill you. Controller I can see as another role, with leader type abilities being distributed for everyone, if they want it.

[MGuy]

Whether or not you might classify a monster generation system is GOOD isn't what's confusing me. The current argument seems to assume that if you have one you don't have the other. I mean K even frickin' said you have your spider whatever and then you can, from there, if you want, set a level and slap more shit on it to make it into the spider of your choice. I don't understand how that prevents you from then generating monsters from that system and providing those pregenned monsters. The way people are talking its as if just because you have a monster gen system you CAN'T then have monsters available for those who don't want to fucks with the system. What's more people are talking about not wanting to spend the time USING the monster gen system to generate monsters for their games but offer no alternative. I mean if you DON'T have a good monster gen system how are you supposed to adjust monsters that aren't viable for the target level that you none the less want to use, without one? Guess? What exactly are people wanting to use if not that? You can argue that it may be bad, sure, but you're otherwise talking about writing up 1000 monsters just from levels 1 to 10 without any guiding compass for no reason when you can get all those monsters just as easily by making a comprehensive gen system and going from there. The only problem I see people raising are concerns about whether or not their can be a 'steve' and I'm not really even 100% if this is even an actual concern or whether or not I give a damn.

[Username17]

Is there a way to make it work though? Brutes and Skirmishers seem distinct enough, one walks up to you to kill you and the other moves behind you to kill you. Controller I can see as another role, with leader type abilities being distributed for everyone, if they want it.

Well for starters, you can't have "hit dice", class progressions, or similar accumulation shenanigans for monsters. As Koumei pointed out, summing a bunch of cumulative effects across 8 levels and 3 digivolutions is bullshit. At no time should it be advantageous to use a calculus function to determine the final statline of a zombie ettin.

So basically we're down to level charts for numbers, selections off ability lists for monster powers, and some very simple templates who primary purpose is to open up new ability lists. Like 4e if it had ever bothered trying to make monster ability guidelines of any kind.

Which is obviously not going to give you monsters that are quite as satisfying as ones which have been custom crafted and had their stats pre-jiggled to better fit the theme. We keep using the Gelatinous Cube example precisely because it's so retarded. It's a fucking 10'x10'x10' cube. It doesn't dodge things. At all. And yet, the 4th edition system gave it a "reflex" defense of 16 - the same as the boneshard skeleton it is teamed with because they are both level 5 brutes. Most outputs of the charts and formulas are not going to be as pants on head retarded as that, but it's clear that you're going to have better overall results taking the outputs of the monster formulas and fiddling with them a bit to make them more thematic than just using the outputs raw.

Also, even if you can make a Gynosphinx after just making a few quick selections off of pulldown menus on an android ap (the best possible case for a make-a-monster system), that's still going to be slower than selecting "Gynosphinx" from a pulldown menu straight off. Any time you're making multiple choices from menus and adding things together, it's always going to be slower than taking the monster ready-made.

So bottom line, you still need a fuck tonne of monsters. You can and should have a monster-maker that is more user friendly, requires less math, and has fairer outputs than those of 3rd or 4th edition. But you still want and need several hundred monsters that are ready made. And those are the monsters that are going to get the most use.

-Username17

[Schleiermacher]

D&D3e classed them by race, with the problems of succubi temptresses and tentacled horrors being the smae class

D&D4e classed them by role, which makes more sense in theory, but the execution...

Is there a way to make it work though? Brutes and Skirmishers seem distinct enough, one walks up to you to kill you and the other moves behind you to kill you. Controller I can see as another role, with leader type abilities being distributed for everyone, if they want it.

Classing monsters by type did have its problems, but since most monster abilities are spesific to the kind of monster anyway, I think it could have worked with better execution. (Spesifically, execution that tried to make the value of 1 HD more or less equal across the board.) Displacer Beasts and Chimeras don't really have a lot in common, even though they're the same "class".

Classing monsters by role, if by role you mean combat tactics ala Brute/Skirmisher/Controller, I think is a fundamentally bad idea, because I think "roles" of that kind are a bad idea in the first place. As opposed to roles like boss monster/glass cannon/wall/imp, which really have more to do with levels and defenses than anything else. Now, defenses can be class-based, but generally not on the dimension of "good defenses" vs. "bad defenses" -every class should have good defenses against something, and some weaknesses.

[shadzar]

I'm going to need someone to explain to me why we seriously can have both a functional, relatively easy to use, monster generation system and then a bunch of generated monsters motherfuckers can simply reference when they are feeling too lazy/are too short on time to stat up a monster.

because if you have a good enough system, then why stat any monsters out? if you have to stat out a ton of monsters to give to people to play with, then why develop this grand system?

yes, in today's computer world it COULD be done to make a system easy to do it, but the best D&D software yet is still Core Rules from EVermore88, and it had problems. many of which was the restraints put on it like the CB of 4th that had too much error-checking to make sure you had only "core" rules in use or whatever.

adding your own material to Core Rules helped with that, but the programmers don't understand well enough the freeform nature of D&D so even Core Rule had terrible magic item creation. try creating a sword with a "BoH" built into the hilt. it doesn't allow things to be classified as weapon and container, so you either pencil it in in the description, or have to make a second item for the container and make sure you take both so you can actually have the programmer allow you to put things into the hilt to know where they are.

you would have to have someone that can write a platform independant program and fully understand D&D and the monster generation as well or better than the designers, then the program would need to work without being connected to the internet, it would ahve to output the monster, etc....


but if you have such a program, then why sell the system to people? why not just spit out those monster combinations as books and sell them to people rather than give them the simple system to do it themselves?

it can be done, but from a business standpoint, especially HASBRO, why give something for free when you can charge for all the little parts?

[deaddmwalking]

because if you have a good enough system, then why stat any monsters out? if you have to stat out a ton of monsters to give to people to play with, then why develop this grand system?

I'm personally in favor of both. You should totally have a good system for monster generation, and throw out as many monster stat books as the market will bear.

If you go to just about any grocery store in the country, you can get all the ingredients to make anything you want - even something totally gourmet with dozens of exotic spices. But they all also have a 'ready made' section, and most of them even have a 'deli counter' of some kind with hot food that doesn't even require a microwave.

Convenience requires that you provide the tools for people to make something amazing if they have the time and you have Hot Pockets and Rotisserie chickens for those who either want something quick or need something 'right this damn minute'.

[Ice9]

I'm definitely in favor of having a monster-creation system. Even if it's slow, it can still be useful for special occasions where you want to hand-roll a BBEG.

The one FantasyCraft uses has too many steps, but does attempt to resolve the "correct stats gameplay wise" vs "makes sense for the creature" dilemma. IIRC, you do something like:
1) Create stats based on the concept.
2) Calculate what level the monster is, based on those.
3) Now adjust the stats to fit the benchmark for that level.

Regarding stuff like Gelatinous Cubes, maybe components that adjusted what level you used for various parts? Like:
Slow and Indestructable - Use the durability for two levels higher, but worse than 1st-level defense.

[Cyberzombie]

Which is obviously not going to give you monsters that are quite as satisfying as ones which have been custom crafted and had their stats pre-jiggled to better fit the theme. We keep using the Gelatinous Cube example precisely because it's so retarded. It's a fucking 10'x10'x10' cube. It doesn't dodge things. At all. And yet, the 4th edition system gave it a "reflex" defense of 16 - the same as the boneshard skeleton it is teamed with because they are both level 5 brutes. Most outputs of the charts and formulas are not going to be as pants on head retarded as that, but it's clear that you're going to have better overall results taking the outputs of the monster formulas and fiddling with them a bit to make them more thematic than just using the outputs raw.

There's no problem with a chart for numeric guidelines, the problem only happens if you forget their guidelines and treat them as hard rules. Believe it or not, when 4E's exception based monsters were announced, I thought it lead to very interesting monsters. With monsters are no longer shackled by strict rules like in 3E, you could potentially have a gelatinous cube with a godawful reflex/AC and just way more hit points to compensate, or some similar alternate defense. You could do potentially anything you wanted with 4E monsters, because everything was just guidelines. They were no longer bound in that hit points, hit dice, skills and saves were all linked.

Unfortunately despite monster designers being handed all that power, we ended up with the most boring monsters of any edition, and the reason for this is most likely the over-egalitarian design philosophy.

It comes down to one issue: 4E design didn't believe in weaknesses. It was so devoted to being balanced that it was part of their design philosophy that no monster ever neutralized a PC's schtick, and no PC ever dominated a monster of similar level. That's what lead to things like wizard flight and golem magic immunity being taken out of the game. It also lead to any kind of numerical weaknesses a monster once had being reduced to only a slight numerical edge, because they didn't want the ray-based sorcerer crushing the gelatinous cube. Everything had to fit in the same boring cookie-cutter mold.

[Krusk]

It seems like it would be worth starting a thread or a conversation about What are the core monsters vs How many core monsters.

We've got numbers frank throws out which sound like a good starting point to me. Is it worth starting to write that down and come up with what the ideal MM includes. No stats or anything, just "low level demons, imp, medium demon bearded devil, high demon balrog". As (if) we agree on what those abstract tiers look like, we should start to work it into a "Ok we want 10 levels of DND, that means at minimum 1 demon every 2 levels, so imp gets to be level 2 and beard gets level 6. We probably ought to scrounge up a level 4" structure.

At minimum those guys doing the next edition of tome should want that info, but I could see it being really helpful with everyones (my) personal Fantasy Heartbreaker too.

[FruitThatAteItself]

What about there being a huge book of plug-in monsters (like the 400+ iconic monsters) and a monster/NPC creation system in a different book?

[PhoneLobster]

Is it worth starting to write that down and come up with what the ideal MM includes. No stats or anything,

Oh gee yeah, just what the gaming den needs ANOTHER big list with no mechanics attached in a wanky conceptual thread.

Then we can talk about how awesome it would be to have 1000 monsters (name only) just like people wank over having fifty bajilionty classes (name only) for the Magic The Gathering as DnD thread and so on.

Woohoo. Hold me back.

[shadzar]

because if you have a good enough system, then why stat any monsters out? if you have to stat out a ton of monsters to give to people to play with, then why develop this grand system?

I'm personally in favor of both. You should totally have a good system for monster generation, and throw out as many monster stat books as the market will bear.

the thing is, WE, do not make the game or have the rights to it.

yeah you could build a system to allow creation of anything with templates added onto a baseline (the "human" race). the problem is, how many people would use that to create everything or a few things and how many people would like to just grab something already made and run with it.

like mentioned you dont want to have to build everything, maybe a BBEG and a couple sub-bosses as special for your game, otherwise you just grab the standard goblin and throw 20 of them on the fly and go with it.

also in that system is it forced? does it require the Ag...whatever from the Wandering Monster article as a spider herder that collects pixie parts to sell to sea hags, or can you just use a monster how you see fit?

some people will want a system for making mosnters, others will want a system to group monsters, some will want both, some will want neither.

now you are dealing with 3 system to satisfy 4 groups of players.

the money will do it the easy way with the most profit.

i surly wouldnt add the depth of ecology that FORCES 6 interlinked creatures outside of a tight-designed campaign setting. James Wyatt is a moron for trying to make a default game setting as detailed as FR since not everyone is going to want to play the D&D game setting, thus why generic Earth (Oerth, Mystara) was what was always used for D&D rather than FR, eBerron, etc.

and don't think for a minute that people will not want that encounter design system built into the mosnter creation system so you can slap down some brutes with minions, etc and they all figure up stats at the same time with some sort of encounter budget; because they don't want to have to make their own groups or dont want a living game, but a board game where things work in these groups cause the game says so and they wait for the players.

4th didnt work. best to do would be to jsut make monsters and sell them for higher profits, and for more sales of various books. that is what the money will do so it doesnt have to force al this other stuff on people, or it will go the more complex route and fail by forcing people to buy crap they have no need for, and don't want the wasted pages of in the books. a 20 page system for encounter design and 30 pages for monster design in a 300 page book is 1/6 wasted space that everyone would have to pay for, and 50 less monsters that could be ready-to-go.

IF they made the monster design book its own thing to include the system and have it work to begin with, then they could sell it separate, same for the encounter design system. but you know ANY system seen as "helpful" to core for some will be forced on all.

[Username17]

It seems like it would be worth starting a thread or a conversation about What are the core monsters vs How many core monsters.

Fair enough. I would put out monster books that were geared towards an implied setting. An implied setting with more meat on it than Nentir fucking Vale, that's for damn sure. You'd still have little boxes for "Wemics in Eberron" or whatever, but it would be focused on your edition's implied setting.

So let's say you got your basic "Light is Spreading" scenario, where the world is emerging from a Dark Age where Bad Stuff™ happened. You have your expanding pocket of civilization, which has the basic playable races in it, and that should probably be like 15 races. If you just have 5 demihumans and 5 "evilish" races, people are going to rightly accuse you of ripping off World of Warcraft, and that is grief that you don't want. So you're going to want a three-polar system and put Orcs and Humans into the same empire. Putting together a list of fifteen or so playable races on release is not actually terribly difficult. The Human, Orc, Elf, Dwarf, Halfling, Gnome, and Drow have to make the cut, but the Half-Orc and Half-Elf really definitely don't. A lot of stuff like the Shadar-Kai can be delayed for a more "planar themed" monster manual down the line (possibly even call it the "Fiend Folio" on the grounds that it has a lot of fiends and shit in it). There's also the Changeling, Dragoncock, Gith, Goblin, Hobgoblin, Kobold, Mongrelfolk, Raptoran, Shifter, Tiefling, Warforged, and so on. You could fit in an actual Undead race here, and I think you should. You should probably have a playable Undead, Construct, and Demonic race (the latter two being the Warforged and the Tiefling - the Undead one could be something like the Vryloka or just plain old Ghouls). But clearly your job here is to cut it down to 15.

Now you want the Savage Races. And yeah, I totally capitalized "Savage". If you're going to be doing conquistador stuff, you need cannibals with stone tools and golden idols for you to stuff in your santa sack. So each of these guys is from an area that is worth crusading/exploring/conquering because it's still full of demon worshipers and monsters from the Dark Age. Each of these zones should probably have like two groups of savage humanoids. The Aztec Underground zone has the Grimlocks and the Kuo Toa, the Banemires has the Lizardfolk and the Bullywugs, the Ozymandias inspired Empire of Dust is full of Gnolls and Draconians, and the Underwater zone is the province of Sahuagin and Locatha. Other shadowy regions (the prairie, the high peaks, the tundra, the bleak woods, and so on) can be opened up later by adding additional content in future books with additional villain races (like the Ophidians and Armand).

So now you have 15 playable races and 8 more pure-villain races. They need domesticated animals. It's actually a very cool thing about the Kobolds and Goblins that they domesticate dire weasels and Worgs respectively. You'll want a couple examples for what each race might use as a familiar, a war dog, and beast of burden. The savage groups need a bit more love because you're going to be spending a lot of time fighting in their villages and dark temples. So each one of them should have at least three available war dog variants ranging from guard dog power to bear power. So for the Gnolls, you want to make sure you have an entry for "Hyena", and another one for "Hyenadon", but also a third entry for their Sarlac equivalent. Should be a nasty scavenger beasty. The Lizardfolk are going the full Dinotopia on this and get to use Ankylosaurus as their beast of burden. Anyway, you've just earmarked 85 monsters to have shout-outs to one of the races in their descriptions.

You also want a big chunk of extraplanars. I already mentioned 102 as a good basic number of straight summonable mooks. But you frankly honestly also want non-mook demons. You want at least a high, a medium, and a low level extra-planar monstrosity that is called on by the dark priests of each of the savage tribes. These can be in the normal rotation of summonable mooks or they can be masterminds in their own right. But it would be stuff like putting in the Marrash to go with the Gnolls.

You really want to get a bunch of monsters that are lesser than a first level hero. So monsters like the Corallax and Ash Rat totally should be core. Some of these can be creatures that fill happy forest friend slots or guard dogs of the Githyanki slots, but they need to be there.

Giants should show up at every level once they start showing up at all. I would say that a Giant should always be at least 2nd level (Ogre), with an additional Giant every level after that. Since you can have Storm Giants and shit, you can easily have Giants that play in the double digit level paradigms.

You want the basic five Dragons. And you want enough age categories so you have a Dragon to throw down at every single supported level from 1st to however high it goes.

Mounts are totally important, and different from "domesticated" animals. It's not all Gryphons and Nightmares, you probably want a lot of them. Go dig up old Heraldric beasts and shit.

-Username17