It's Friday morning and you have conceived of an intelligent homosexual mushroom, or an 18 foot tall gerbil that shoots rays out of its eyes, or a race of three-eyed gnomes with remarkable skill at throwing pick-axes, and it occurs to you, "I will draw up the stats for that monster!" So you do and you post them on your blog, giving the mushroom, gerbil and gnome their own special names (fuungay, grrbel and triomes), plus hit dice, armor class and lists of special abilities ... and presto, the Great Beastiary of Internet Folly is a little better padded for your efforts.
Posting monsters is a favorite hobby for D&D bloggers; some blogs do nothing but post monster after monster, some of which make someone happy, somewhere. Demographics guarantee that with a large enough pool of potential readers, one potential reader capable of hitting the buttons to make a complimentary comment on your blog shall justify all the monsters the gentle reader has ever posted.
Here, however, is my issue with it.
Monsters are not made useful by hit dice or armor class. They are not made useful by numbers appearing, or frequency, or rating their intelligence or size or any other statistical characteristic. Statistics are garbage. They do not tell me word one about how this monster is useful in my campaign.
Nor do the general biological or simplistic sociological characteristics usually assigned to such monsters. It's nice that your monster is a battlewagon, a clanlike humanoid or an imaginative demon, but here's a notice for you - no one NEEDS those. We have plenty of battlewagons and humanoids and demons ... and in all honesty, if the problem you're attempting to solve is that your genre-savvy players already know the monsters in the book, you're missing the point.
Stop obeying the book.
Your players do not have the right to know that goblins have 4 hit points or that Night Hags have a 50% magic resistance (don't quote me on that, I don't have the book in front of me). They are not entitled to know a bugbear's hit points, a rakshasa's AC or a sphinx's spell list. Those things are NOT rules that you, as DM, are required to follow. If you want the goblins in your world to have 80 hit dice, then the players have to suck that shit up - boo hoo, poor them, they didn't get to have superhuman knowledge in advance. Fuck that they've read the books through and through ... you are not required to play the statistics in the various monster compendiums by ANY law of DMing that I know of.
So if you want to throw your players off balance, and deactivate their genre-savviness, change the numbers and watch them squirm.
Yes, yes, yes, I know, you still want to create new monsters. Let me suggest a better guideline that splat-stats. Consider, if you will, the gelatinous cube.
Here is a genre-savvy monster type. Experienced players will guess a gelatinous cube is about to be in the offing just by the phrase, "the dungeon floor has no dust."
The cube serves a UNIQUE purpose. It's a jelly, but it can be hit by ordinary weapons, it has an extraordinarily low AC, it doesn't cause much damage but it delivers a non-killing and yet highly nasty special attack (paralyzation), it gives a bit of treasure and it is mostly invisible even under light. Thus, I can throw it at a low level party and it probably won't kill that party with damage, but it MIGHT manage a total party kill if everyone blows their saving throw. Cubes have 4 hit dice which are good for experience, and the easy hit means everyone, even the mages, will get to do something. In general, it is a damn good slashing monster perfect for parties without magic weapons and without many hit points.
I can make cubes as big as buildings for high level parties, give them hundreds of hit points, increase their ACs by arguing that 90% of the hits parties make aren't hitting any "vital nodes" and it is STILL a good monster. Logically, a good earthquake will break it into hundreds of cubes, while an ice storm may freeze one part of it and yet the rest may still be dangerous. It has elements of weirdness in it that adjust with flexibility to a multitude of attacks.
What is needed, then, is a monster that doesn't fit some previous niche. It isn't enough to make it bigger or faster or magical or in some way the exact same only now its a spider or a crab or a peacock. A peacock that's a lich is still a freaking lich ... the fact that its also a dead peacock just isn't enough to make it interesting!
The only thing is, most, if all, the niches have been filled. Those sorts of combat vaguaries are the easiest ones to work on, and therefore were the first slots to have monsters plowed into them. What's needed is not just a new combat niche ... but an entirely new ecological niche entirely.
When you sit to create your new monster, what is truly, awesomely needed is a CONTEXT. How is the existence of this monster going to change the sociological characteristics of a world? We build walls to keep out the orcs; we organize woodcutting parties to go out and rid the area of violet fungi or aggressive treants. We string nets over the town to keep out the stirges. How, exactly, does your monster affect the people of your world? How do people interact with it. I'm not talking just a border, with their land and our land ... we have that for hundreds of monsters already in existence. How does YOUR monster change the rules, the precepts or the habits of the other creatures and peoples in your world?
Work on that problem. Cows and chickens and horses may make boring monsters, but the world is radically different because these creatures exist. A good monster - a really good monster - will offer some truly profound change to the worlds of the people reading your blog.
Don't ask me how you'll do that. I haven't invented a new monster in years. I haven't needed one.