(This topic is focusing on party-based RPGs, and mainly on those that are non-tactical; that is, ones where you don't spacially move your characters during combat.)

There's something interesting that I noticed involving tanking in RPGs, or specifically how you can try to control which characters the enemies attack.

In older games, it was typically dependent on the position of the character in the party line-up. Examples of this include Dragon Quest 3, Final Fantasy 1, SaGa 1 and 2, and there's also partial example in Wizardry 1-7 (monsters can only attack slots 1-3, though 5 started to add some exceptions to the rule) and Might & Magic 1-2 (only the first n characters can participate in melee, where n changes from battle to battle).

Some modern RPGs, I've noticed, have borrowed the threat mechanic from MMORPGs, where each character can build up threat by dealing damage or healing (sometimes by using abilities that specifically have that effect), and enemies target the character with the highest threat. Examples of this include the Romancing SaGa remake (on PS2; ever wondered exactly what the Chameleon spell does in that game? It lowers threat), Crystal Project, and apparently Chained Echoes.

There's also games where tanking is not supported by the game mechanics, like SaGa 3 (though the original version at lest allows characters to cover others, but sadly the remake removed that option), Might & Magic 3-5, and the like. (Tanking doesn't seem feasible in False Skies, for that matter.)

Between these methods of controlling who the enemies attack, what are your thoughts, and is there one approach you prefer?

I like the MMO style of each enemy keeping an aggro rating for each of the players. But I think it only works if each class is very unique and fills a particular role. If all the characters are similar I don't think it works so well.

I love the idea of rows, with aggro increase abilities (both passive and active). Tales of Arise, though an action JRPG, has abilities on accessories that allow you to increase or decrease aggro. Because of monster AI they are mostly useless (fully stacking the +aggro up ability on your tank results in nearly no focus by enemies). The AI seems to focus almost solely on the character the player is controlling, and switching characters, even without attacking to draw aggro will usually cause the enemy to change focus to your new character.

But I love the idea of it, when it works. Abilities that steal all aggro for a few turns, passive abilities that make a tank more likely to be targeted are all good, in my eyes.

I never really thought about that. Maybe because I never played the games you mentioned. In others, I suppose I would notice it only if it works against me. If they choose to go after the character I prefer, I ascribe it to good planning on my part. If instead they go after one other character that is more vulnerable, I think it is good tactics on their part.
Only more recently do I see taunts and similar actions, and have explored those possibilities.

If you do happen to come across a short video where the new mechanics are well displayed, please to share it here so that I can see them; but please don't waste your time with it.

I don't have a strong opinion towards either method, it really depends on how it's implemented.

There are visual novel and RPG hybrids, Loren the Amazon Princess and Seasons of the Wolf, have both positioning and aggro. However, posititioning seems to be more dominant since there are not many ways to manipulate aggro. As long as a front row character remains, back row characters protected from most melee attacks, but can still be targeted by magic and ranged attacks. Additionally, tanky characters often have a ability to guard character of the player choice, taking damage in their place, the enemy parties will often use this ability on party members whose health are getting low.

In Star Crawlers, there are a few classes that have a tanking-oriented skill tree. The soldier class is arguably the best for that role, with abilities to draw aggro and absorb damage. That said, healing abilities are rather weak and consumables take up precious inventory space. Thus, all classes have a damage mitigation ability, so the party can endure multiple encounters.

I don't have a video, but I note that Crystal Project has a generous demo, and it's easy to see the mechanic in play there (especially since one of the starting classes starts with a taunt ability). I'm reminded of one of the first games I played with a taunt ability, the first Etrian Odyssey where the Protector class has such an ability. Unfortunately, that ability didn't actually work in that game, making that strategy useless. (Incidentally, I would say that Protectors aren't that good until you get to post-game bosses where they're practically required, or else an boss will do 4-digit damage to your entire party when you only have 3-digit HP.)

I always liked the idea of activities in battle creating aggro, because that seems more like how actual battles would go and let's you strategize more during the fight.

I prefer to be able to move my characters, not blobbers. And even in case of blobbers, being able to change positions. Lords of Xulima does the positioning part quite nicely, between regular melee weapons, pole weapons, range weapons and spells and taking actual position on a row into account to determine reach, and characters and enemies can change position during combat. It completely fails on the aggro part though, being like an RPG Maker game there, entirely random.
And on that aggro part, while they can make for interesting game mechanics, skills that directly affect it feel wrong. Proper enemy AI would mean determining what characters actually pose the most threat at the moment and which can be taken out more easily. Taunt-like skills can play a part, but more in the sense of having a chance to enrage and break concentration, with various potential effects if the target fails to resist.

I much prefer row/rank and not "threat". "taunt"-like skills can still exist without the whole "threat" systems and the baggage they entail.

Then again, I'm not keen on overt roles ("tank/support/spike/sustain/healer") either. That stuff should have remained in the realm of shallow MMO (but I repeat myself).

Not to say I can't like a game with it/them (Crystal Project instantly joined my top-10-ever, possibly even highly in the rankings). I just prefer design otherwise.

I think I once saw what I think is an RPG Maker game that had aggro.

(Worth noting that I actually prefer there to be some randomness so that things don't get too predictable.) Even if that skill is something like Romancing SaGa's Chameleon or Crystal Project's Hide, which could be explained as making it harder for the enemies to find the user? So, are you saying that casting a protection spell on a character should reduce aggro/threat toward the target? Having some character roles can be useful, as it helps avoid the problem where each character is too similar.

On the other hand, it can be occasionally interesting to have skills that might not conform to the character's primary role. For example, I like it when healing abilities aren't just for healer classes. (For example, Crystal Project has many classes that get a bit of healing (including Scholar and Valkyrie, for example); on the other hand, False Skies is more strict, with the only non-healing classes with healing moves being Generalist, Chemist, and the two classes that can summon (though healing via summon is unreliable).)

Also worth noting that, even when there are characters with strict roles, I like to have a viable generalist. It's why my final Final Fantasy 3 3D party actually uses a Red Mage; that class can use endgame weapons and armor, can cast a healing spell that's strong enough, and has a revive spell in case it's needed (particularly if you're like me and refuse to use items that are in finite supply or if you've used too many Phoenix Downs already). It's just that they're not able to use the strongest weapons (unless you're able to get Ultima Weapon), and they don't get many uses of the level that has the good healing spell.

Having played classes that are outright agro magnets in MMORPGs, I'm going to say that this works best if you have some means of either controlling where the agro goes or at least mitigating the damage.

In single player games, I prefer random agro when battles are short. I also prefer to not have to get super tactical every battle - save that for boss level encounters.

yea, 'taunt' the first thing that came to mind....

There should be a tactical layer involved, preferably based on opponent type. Non-intelligent species will just uhm protect or hunt while intelligent species should be able to determine the biggest treat and act accordingly (yes, even with a row system)

The way I see it, the tactics with random battles should be geared towards minimizing resource use. There's a lot of random battles, so the player should have to be careful enough to not run out of resources before reaching the end of the dungeon (which may or may not have a boss).

Also, I think it's useful to have the occasional "problem enemy" (that is, the sort of enemy that can cause major issues if not handled properly), so that the player can't just mindlessly go through all random encounters without any thought. With that said, perhaps these shouldn't be the only encounters in the area, and there should be a significant reward for killing them; for example, perhaps they're worth more XP than the other enemies in the area, so there's an incentive to taking the fight (rather than, say, running away). Also, attacks that would break boss fights if they weren't immune, like many status ailments (including instant death) could be made useful in these encounters.

There's way too many RPGs where random encounters are so easy the game becomes boring (at least when you're not fighting a boss).

A related issue, and one that does affect modern indies like Crystal Project and False Skies (as well as some classics; Final Fantasy 6 has one particular large dungeon level that's particularly awful about this) is a lack of enemy variety in a given area. When every fight is the same, it gets boring.

(Worth noting that I'm speaking from the perspective of a high-attrition game design; in a low-attrition game, when running the player out of resources through a series of battles doesn't happen, you can't have large dungeons with lots of encounters in them.)

depending on the game, they both work fine and i have no preference.

i would also recomend Monsters Den: Godfall, which have a combination of both

Too bad gog rejected that game/developer. :(