enemies often tries to run away from players with Opportunist..."the AI doesn't know that you have Opportunist, nor should they!"
This is intentional (we don't want the talent to go to waste), but clearly it's giving some issues at the moment.
Having a character choose not to move away from you as not to take an Opportunity Attack isn't a waste of a talent.
To fix this issue I would be to just bake OA's back into the system. All characters with a melee weapon get OA's (or with specific ones, like backstab and daggers). Golem Studios is in the early stages of demoing
now that gives attack preferences to monsters. Some stronger preferences will have them take an OA while others the monster will respect that threat.
Could do something similar to that so that melee character's aren't overly sticky and they get to use OA sometimes. Can even reveal these priorities to players with Loremaster so they can do things to try and trigger the OA's.
The laughably bad defensive abilities.
I think you're missing the primary reason these defensive abilities aren't very good, and some of your criticism I don't consider valid.
The primary reason that the defensive abilities aren't good defensive abilities is that offensive abilities are the best defensive abilities. The best defense is a Hard CC. To enact hard CC you need to break armors, which you need damage to do. So maximizing your damage output to break armors to allow for Hard CC is the most effective defensive ability in this game.
As for this:
However, even if Retribution would be buffed to 200%: you'd be rewarding the player.. for doing nothing
That's not true. To use Retribution well you need to find ways to encourage the enemy to be attack your players with Retribution. That requires some planning and positioning to use well. It is hardly nothing.
I'm not going to defend Retribution's mechanics as valuable to gameplay, I'd remove it as well, but to suggest the player has to do nothing to use it well is false. There is gameplay in using Retribution well.
I agree on the problems with circumstances you put forth on Leadership and Perseverance (as an aside Petrify was resisted by Armor in EA, seems they forgot to change this).
Binarity of outcomes, predictability, and the armor system.
You throw a lot of criticism here and I disagree with most of it.
Unintuitiveness of effect types and targeting armor
I'm actually not entirely sure I understand this complaint. Is it breaking immersion for you? Magic armor is visualized as a force field in Dos2 that blocks objects that are elemental from harming the player. For me this was neither immersion breaking or unintuitive. So I'm confused as to the nature of the complaint. Can you clarify?
What it does is that it leads to a completely predictable combat system on any given turn...
Simply put, the armor system completely negates any element of the Delta of Randomness. This is terrible design in itself, because that delta of randomness is one of the staples, I would say fundamentals of good turn-based gameplay, and the armor system breaks this completely on a systemic, tactical level.
I greatly prefer being able to plan, execute my actions and my adapt my plan are to choices the AI makes in targeting, abilities I didn't know they had etc. This makes each fight feel more like a tactical puzzle I have to solve vs hoping my CC's proc.
I don't think this armor and CC system is perfect, and there are changes I'd make but it is a great improvement over DOS1.
For an EC video, which are generally pretty bad, that one was surprisingly decent. Yet...it has no relevance at all to this circumstance. It's largely talking about randomness in competitive games and barely gets into the balance implications of randomness in PvE.
Saying that randomness is a stable of game design is also patently absurd. I can't think of a single great turn based game that has high reliance on randomness besides poker and accounting for randomness is just a low aspect skill of that game. While it is a common thing to use in turn based games, to simulate variance in ability to execute, it diminishes the value of the core skill these games test, strategy. Are you prepared to argue that Chess and Go are not just bad but terrible games? Maybe I'm misunderstanding what you are trying to say here. If my response to this doesn't seem on point can you clarify?
So how can it be resolved? Short of reworking the entirety of the game and reinstate a save- or resistance-based system, there's not a whole lot that can be done. If absolutely determined to maintain the armor system at it's core in how it works right now, several things would have to change.
I'm not entirely sure what you think needs to be solved, besides that you wrongly seem to think reliability in effect is bad.
Even if I agreed, there is a large variety of options in changing the armor system as it is now while keeping the heart of it. My primary concern is that dual damage type parties don't work well. An easy solve to this is having all damage target vitality and deal damage to an armor type. Status doesn't apply until the related type is broken but all attacks will deal HP damage. Suggesting you need randomness to have not terrible games is just moronic.
The Round-Robin Turn Orders.
We kind of agree here. It a confusing choice, not because round robin was chosen but because the rest of the system wasn't then modified to account for this choice.
Having initiative as it was in early access seemed fine. Players had to choose between putting points into the better offensive stats or going earlier in the round. The problem only seemed to present itself by having all enemies that had silly scaling initiative based on level.
There seemed to have been many systemic changes late in development (like moving Petrify to magical instead of physical) and they didn't account for the systemic effect of those changes. It's a hard thing to account for especially as you are near an end of a project and so close to the project. It's also solved by hiring an experienced system designer, something they can hopefully do easily now as this game seems to be selling quite well.
Initiative systems are hard to make not bland. Many games over the years have tried interesting things, none have really stuck and become persistent to the genre. The closest is FF's Active Time Battle but really only square uses that. FF Tactics, Radiant Historia and various others have tried to make changes and mechanics around it, however it really does require an whole system to be based around the initiative mechanic to do anything complex well.
Now to address your specific points.
The only time set-ups like this can be done reliable, to see a plan and see it take shape, to get that rush of "feelings" is when you're at the absolute end of the turn order in a little clump.
First I question the value of your rush of feelings. That's a longer discussion and I'll ignore it for now.
Why aren't you checking the initiative order if the enemies, to see where they are going and who isn't moving between your team so you can combo them? Just because you have to be a little careful with the order of your combo and which enemy you pick doesn't mean the opportunities aren't there. Be more selective in who you are choosing to combo, and account for which enemies go between allies you need to combo in future rounds, the game gives lots of tools for dealing with those units (teleport, slow from oil, clouds if they are ranged, etc).
Yes, yes I was, and they are - to a point...
And if you kill an enemy, you might expect them to be removed from the turn order, right? Well they are. But someone else will immediately take their place.
Right...this is again requires you to have better planning to know when to knock people out vs down. That you don't like that the game tests that element of planning multiple turns isn't bad game design, its just a part of this game that you don't like.
Both the round robin and straight initiative value systems have drawbacks. Neither are much greater than the other. The criticism should stop at how their initiative system isn't transparent, has elements that suggest it behaves different than it does and the rest of the system has issues with how the initiative system behaves (the value of Wit as you pointed out or gear changing your preferred character turn order because it has +init, etc).
I appreciate that you took time to post something long and that you care about. There is some valid criticism you provide and I think I was fair in pointing out where, but there are also places where you miss the mark by quite a bit.