Here's something I hadn't even considered the last time I posted here, but that actually seems like it's really important if you're adding new types: how you express the identity of your type.
In Generations II and VI, the types Dark, Steel and Fairy had extremely simple and exaggerated identities. Most of the Pokémon of these types were entirely designed as "here is the most basic meaning of this type."
In fact, to demonstrate their "meaning" even further, all three of them were represented by a new Pokémon of the new type
that evolved from an old Pokémon that didn't have the type, and not only did these Pokémon exist, but
players had to see them in a context that made their type relevant and obvious. Jasmine, the first Steel-type specialist in the series, used a Steelix, which was designed as a direct contrast to Onix to show what Steel
meant as opposed to Rock. Karen, the first Dark-type specialist in the series, and Valerie, the first Fairy-type specialist, used Umbreon and Sylveon, both direct contrasts to Eevee to show how Dark and Fairy differed from Normal. Sylveon is especially notable because... well, it's actually the
only cross-Generational evolution we've had since Generation
IV, and there's a full Generation on either side of it that got none at all. It's pretty obvious that this was its main purpose and that Game Freak sees this kind of introduction-by-contrast as necessary!
But there was also a pattern in their designs. Nearly every Dark-type in Gen II is dark-colored (especially blue or grey) and portrayed as cunning and cruel in behavior. And then, to show the full scope of the design,
just one exception was introduced: Tyranitar, not a sleek, cunning, dark-colored predator but a bulky, pale green... well...
tyrant. Most Steel-types were slow, tanky and
literally covered in metal, but then we have, again,
one exception: Scizor, a bright red, slim Pokémon with punches like bullets and big, metal pincers. Nearly every Fairy-type in Gen VI is cutesy with a big head as if still a basic Pokémon; Fairy was a little more diverse than Dark or Steel when they started, but although there were two sides of this spectrum, there was still no middle ground between extraordinarily round and chubby (like Slurpuff and Dedenne) and extraordinarily slim and tall (like Gardevoir and Florges), and still a good half of them were similar shades of pink and white instead of diverse colors; even the rest were at least very light even if not pink and white. But then we have Xerneas, which has colorful and vibrant antlers but also a graceful, dark-colored body, actually an unusually small head rather than the standard Fairy's super chibi look, and a feeling of sheer majesty.
To put it one way, when you see most Fairy-types, you think awww! but when you see Xerneas, you feel awe. In every case, when the types were first introduced, there were quite a few designs that were on-the-nose, blatant pattern-followers and only really one single Pokémon of each type that broke the established trends.
I don't think this was Game Freak simply being uncreative. When the new types were introduced, both the patterns and the single exceptions are necessary. The patterns make it so you
know the type when you see it and
can't forget the type exists - if this is the first game with a new type, there shouldn't be any confusion when your opponent has it, or you'll be frustrated and taken unfairly by surprise. At the same time, the exceptions show you that that's not all there is to the type and keep it from being too bland so you don't get sick of it or find it unnecessary - but remember that
each type started with only one exception!
For the entire rest of the Generation, these new types were treated as new and exciting. Especially in Gen VI, tons of people were hyped over Mega Altaria, begged for a Water/Fairy Mega Milotic, and so on - it was
the awesome new thing, it was established that it was a pretty strong type, it would make all of the old favorites even better, and just about everyone wanted it... and then with the end of the Generation and the loss of its "newness," the hype cooled off, and it was only after it did that Game Freak started branching out and giving more varied Dark, Steel and Fairy Pokémon like Absol, Mawile and Tapu Koko. These were Pokémon that didn't scream their types at you when you saw them, but that were easily swallowed as Dark, Steel and Fairy and accepted right away.
To put it in perspective, imagine if we switched the order of Generations VI and VII, and we had the massive range of Fairies from the latter before any of Generation VI's super-pattern-oriented examples - if we the cute emotion-sensitive Cutiefly and Ribombee, the dangerous and scary Mimikyu that tries to hide that and only really wants love, the honored Island Guardians that are actually selfish and hardly offer any protection, the manipulative alien mushroom... The type would be a bit more interesting, but you'd be left wondering what the type meant and unable to tell when your opponents were Fairy, making it impossible to learn the matchups. A big part of why they worked so well in Generation VII is that they were a
direct response to Game Freak's testing the waters in Generation VI: these new Fairy-types showed traits that fans had been associating with them but that Generation VI's actually hardly displayed (an insidious, manipulative or two-sided nature, like the Island Guardians, Shiinotic, and perhaps even Mimikyu... heck, even Ribombee was based on the bee fly,
which is exactly what fans wanted Fairy to represent long before Cutiefly was shown, although the actual Pokémon is much more positive so this is perhaps not the best example). Game Freak waited until it was clear how people interpreted the type themselves before taking it beyond a one-sided "pitch," because that way, the newer designs fit with the ideas people already thought about as Fairy-types and didn't end up questioned. It was really important for Generation VI to test the waters with the type and then for Generation VII to follow through and take the most accepted and popular direction.
There's something special on both sides of this: the initial period of "everything that's <Dark/Steel/Fairy> looks like this and is like this" made it easy for fans to adjust to the new type and kept anyone from being surprised, confused or unable to figure out what Pokémon were those types, with the exceptions like Tyranitar, Scizor and Xerneas being enough to show that the type wasn't just one-sided but still rare enough to make it easy to understand and recognize the type. And then the later period of more liberal distribution was what
made it a Pokémon type and a regular part of the series, not just a Generational feature and a simple gimmick. The reason I bring this up is that it takes several games for the type to acclimate and "become" a proper Pokémon type, which is a luxury most fan developers can't afford and a huge difficulty. If you only have one game, there's a very tough balance to strike - if you're not careful, your type will either be
stuck in that period of newness (where everything fits a pattern except the rare exception, the type is largely one-sided and you rely on the excitement to carry you) or rush past it too quickly and be vaguely defined (so you get cool designs, but nobody is quite sure what the type means, you don't necessarily match the popular expectations, and the introduction of the type feels overwhelming). This is the reason Game Freak put these phases in different games: because they're both important to the type's development and acceptance into the gameplay. So how
can a single standalone fangame do this well?
Uranium actually did a great job of reflecting both sides of this and did it in the right order, certainly helped a ton by the frequent exposure to the Nuclear type through the plot, but even more so by the sheer length of the adventure. All three areas I pointed out above were addressed in Uranium:
• The first Nuclear-type Pokémon you were forced to meet were Nuclear variations of already known Pokémon, so you could see how it was different from those other types.
• The first Nuclear-type Pokémon had extreme design similarities and were instantly recognizable for their prominent dark palettes, green auras and generally high contrast between black and neon designs. Right away, the first Nuclear-types you met carried a strong, obvious identity so you could never be unsure if your opponent was Nuclear-type.
• It's the later Nuclear-types that diverge from these designs and look more unique and interesting, and they do so with strong purpose after players already have an understanding of the type and will recognize and accept these new additions. For example, Nucleon, Hazma, Geigeroach and Xenoqueen all represent their Nuclear types in different ways, but they make it abundantly obvious by having it in their names (Nucleon), being part of a directly Nuclear-type-related plot point (Hazma) or being met in an environment where only Nuclear-types can survive (Geigeroach and Xenoqueen). You would know these Pokémon were Nuclear-types no matter what they looked like, so they were free to represent the type in more unique ways and all focus on different aspects of it without following the examples set by the first Nuclear-types.
It's good to follow Uranium's example. Even if your new type isn't a major part of the story, you need to spend a
lot of time making absolutely sure that players understand the type before you branch out and make diverse designs.
I think games with an episodic or demo-based release schedule have a lot more room for new types than completed games that are only released when they're finished. Which is funny, because I've never directly advocated for demo releases before, but if you have a new type, it's actually a pretty helpful tool. Here's how I would introduce it:
• In the first demo that features the new type, you should have an evolution (or form, like Uranium) of a canon Pokémon that follows really simplistic, blatant design patterns for the new type instead of trying to be unique about how it portrays the type.
• In that same demo, every new-typed Pokémon should follow a similar, super obvious design pattern - I would say that at least the first five or so lines of the new type that you introduce should be extremely blatant about being products of a pattern and not varied interpretations of the type. Have maybe one exception to this showing your own idea of another side of your type and establishing a wider range for its "influence" than what you started with.
• In the time between that demo and any later ones that introduce new Pokémon, take note of any feedback you receive, what everyone seems to think of your new type, and what Pokémon they seem to think should get it.
This feedback is your type's new definition. Ignore the patterns you set with earlier Fakemon and base any subsequent designs or retypes on what this feedback tells you about your type.
I'm always in favor of a developer making the game they want and not trying to focus on their audience, so this isn't a set-in-stone "rule," but it's definitely the closest method I can see to doing a new type well and making it understandable to new players. Fairy made a complete 180 shift between Generation VI (one-sided, cutesy and friendly, and especially pink) and Generation VII (anything from creepy, selfish and mischievous to empathic and holy, often with pastel colors but a generally wider range), but it was natural and unnoticed because it was already what people thought about Fairy and wanted it to be.
If you can do
that with your new type - set a simplified version of patterned "expectations" and then be ready to widen your range based on what
you see people think your type means instead of what those patterns had you making - it will probably fit really well in your game and actually be one of the highlights of it instead of something that drives people off. If this thread (including my own previous post) can be any indication, new types drive people off if they're not done really well... but we did just see what a great job Game Freak did with Dark, Steel and later Fairy, and the Fairy-types we got in Generation VII as the fruits of the exact method I observe here were really some of the highlights of the Pokédex, so I want to believe it can be done - just be extremely careful of how you do it.
(This "formula" doesn't work if you don't have Fakemon, but does anybody ever add entire new types and not any new Pokémon? At that level... there's not really any salvaging your type anyway.)
-----
A much more minor note, but one thing that Game Freak has been careful about is replacing types. When a new type is added, any old Pokémon that get updated are either single-typed and get it as a second type or are Normal-type and lose Normal. To make something like Volcarona or Chandelure Light-type is a lot more difficult than, say, Watchog - both of them are extremely dependent on both of their existing types. Chandelure is based on will-o-wisps and haunted chandeliers and it's known for stealing life force, while its origin is focused on purgatory and burning, and its whole line is based on light sources. It's an obvious candidate for the Light type, but it's extremely hard to justify making it anything short of triple-typed (
this is not a suggestion) because it can't lose either of its current types. So you really need to be careful that your type doesn't unintentionally step on the toes of other types, or you'll end up backing yourself into a corner like this.
-----
I'm still not doing any new types myself, but I wonder if this different/observational perspective might be of better use than my last post because I have a better understanding now than then of how Game Freak manages to pull it off so well.
Edit much later: the horizontal rule cut out some parts of this post, so I've replaced them with dashes to restore it to its complete form!