Those are comprehensive gameplay ideas:
Consider this: it's an open world, or interconnected semi-open world game. In normal battles you fight one on one, and though the bones of the combat is similar it's more of an ATB style. You're Pokémon has up to four moves, with different moves taking longer to either to execute or a longer cooldown period and each move can be used a limited amount of times either in a single battle or between rest points. You dodge/or block manually to reduce the damage and switch Pokémon on the fly.
I think there's an interesting philosophy attached to developing a pure turn-based games that allows players to require
zero hand-eye coordination to make it available to children, the eldery and disabled people. But I cannot prove that this is why Pokémon focuses on pure turn-based systems. (it could just be to be compatible between games)
You can actually see the Pokémon within the world, but sometimes you might have to interact with the environment to find a Pokémon. Maybe it's leaving berries or food on the ground in an area with signs of a Pokémon, maybe it's hitting a tree to see if a Pokémon is in there, you get the point. And like Kingdom Hearts III, you have a camera function in your Pokédex to take pictures of Pokémon ala Pokémon Snap.
Funilly, all those three are mechanics in Pokémon.
- Leaving Pokéblocks (Gen 3) or Honey in some containers to attract certain Pokémon. But you check the trap later to see if something came.
- Hitting trees with Headbutt to make Pokémon fall from a tree (Gen 2). Sometimes the Pokémon would start the fight asleep. Hitting rocks also sometimes triggers a fight.
- Sun and Moon had a camera mode where you could take picture of Pokémon.
The issue I guess is the presentation: all those mechanics are only possible in dedicated spots (well, headbutt-able trees were all over the map), and don't have dedicated animation ("fade to black" or nothing) and so is not seamlessly integrated with the rest of the game and end up not being "memorable", and thus forgotten.
Comparing to Zelda puzzles before BotW: each puzzle element was manually created and did not have any signifiance outside of said puzzle. That lead to many types of torches that using different fire logic. In BotW, they integrated everything into a "chemistry engine". One single fire system controls all the fire puzzles. However in "classic" Pokémon, there is no action component to integrate with those mechanics. As a result it generally doesn't change much to the gameplay that those mechanics happen in a sub-menu rather than directly on the overworld, due to how rigidly it plays. And a sub-menu doesn't have to be animated to convey the action.
That's why Roaming "visible on the map" Pokémon are interesting, since they could integrate with all those points. They make the encounter system more dynamic and more memorable, by forcing Pokémon to appear and thus be animated, even if it doesn't change anything to the actual game mechanic.
"Sweet scent" could even be used to still have quick back-to-back random encounters.
The problem is generating the animation for Pokémon falling from tree, being attracted to a trap, "idling" around, might be what's blocking GameFreak.
By animation I don't mean the 3D rig per-se, but the logic behind a basic AI that would procedurally decide how to position and act based on the size of the Pokémon, its shape (how do you even decide?). Pretty sure they would decide it's not worth the money and just "fade to black", like they did time and times again when any "special" interaction happens in the game so far. Yes, indeed even a big budget doesn't help for problems that scale badly. But I'm not sure they always do the right cost-analysis (Pokémon turning around in Pokémon-amie → fading to black...).
One issue that roaming Pokémon bring is changes to the dynamics of resource-based exploration in the early levels. Tampering with RPG difficulty is a hard subject to study. If you ignore the starters, you could look at how LGPE handled the difficulty curve, while acknowledging how much of an outlier that game is (151 Pokémon, Go influences, no pressure to be compatible with the previous and next entries).
Conditions and elements matter even more, because like Divinity: Original Sin if you are in a water puddle and an electric move is used the damage or chance to paralyze increase. If a fire move is used it creates steam.
I'm fairly certain there was a gen where some attacks would interact with background elements during the combat. (I think it was mostly to obtain some items.) It was limited to certain moves only and I don't think it came back. That's one problem to think of, GameFreak considers that each game should have some "exclusive" mechanics because they still want to sell them long after the next gen is released (I don't think it's a positive).
Some interactions you mention are similar to weather which does have gameplay effect (fire is weaker during rain, thunder never misses under rain, etc.).
It's just... limited to a handful or so of
weather and
terrain with types, and some attack-attack interactions (
defense curl + rollout,
dig + earthquake...).
So again, the number of interactions being limited compared to an integrated, dynamic system. Mostly being stat boosts, not explicit, not animated. Making them less remarked despite the mechanics being there.
Similarly, some Pokémon have hidden moves that can be unlocked that serve you outside of battle, such as being able to cut certain obstacles. When you fly, you actually fly the Pokémon. When you surf, you actually surf the Pokémon. The environment and your Pokemon's abilities synergize inside and outside of battle.
I think it's going in this direction in the next five years. At least in LGPE you could ride certain big Pokémon to fly over any obstacle, in the real overworld. (after beating the elite 4)
It's mostly a question of animation, since other non-gameplay modifying "rideable" Pokémon each had their own custom animation in LGPE (not all 150, just the big ones). And a matter of GameFreak allowing "game exclusive" mechanics to actually become staples of the serie...
They have an excuse there, there's no way to automatise that. (But maybe with the advent of content-creation AI... Which still has a long way to go.)
There's puzzles using your Pokemon's abilities, platforming (maybe even with a little bit of The Last Guardian platforming with the aid of a Pokémon). All of the fun exploration and side content you would find in a Breath of the Wild style game, also taking inspiration from other RPG's.
Static vs dynamic world. In BotW when something goes wrong physics-wise you can take damage and die, wiggle around, teleport... I don't think Pokémon overworld would go into that direction.
If Pokémon abilities control progress, better make sure you defined them such as something they can't lose by evolving/getting rid of common attacks (similar to how HMs couldn't be removed unless you went to a specific city).
----
So the main blocks are:
- new "interaction" mechanics are relegated to their own sub-menu/"fade to black", because they aren't integrated to the rest of the overworld (because they often don't come back), making them less memorable/less presentable.
- existing interactions during combat (weather, terrain, attacks combos) are limited to a handful of ideas, because they require manual input from development, and are generally ignored by players.
I think it's mainly a presentation (animation, game structure) problem, before becoming a game mechanics one.
