I agree that WAR on-demand mitigation with IB is more awkward and convoluted than PLD or DRK which are both pretty straightforward, but I wouldn't paint as bleak of a picture as it seems you are.
Yes, IB is stance-locked but WAR also has the most built into their kit to easily transition between stances with them being ogcd and having no resource cost. There is the 10s recast so you can't pop into Defiance, IB and pop back to Deliverance immediately but with Unchained you can negate most of the negative effect of being in Defiance, losing out on only the 5% dps boost from Deliverance and not being able to FC within that time.
Then there is the question of how important or needed on-demand mitigation is in this game. I would argue not much due to how scripted and planned for all relevant incoming damage is and how low impact the fluff damage is. Honestly where I see the on-demand mitigation of Sheltron and TBN being used most is to supplement fluff damage soaking and for stacking with other defensive CDs to take a TB. So really the "on-demand" part means very little and they really act as more "supplemental" mitigation to fill in gaps and bolster other abilities. The fact that they are available so frequently is in a way necessary for them to layered in here and there as they should be.
So then that leads into how these easily available "supplemental" mitigation abilities and the rest of the defensive kits of PLD and DRK relate to WAR's kit.
Holmgang has a much shorter recast compared to other defensive "ultimates" and so can be used to handle more TBs which then frees up more of the defensive CDs to be used for fluff damage.
That is actually not to dissimilar to the end result of the defensive CD kits of PLD and DRK that have to be more spread out and split to handle TBs and fluff damage, relying on proper TBN/Sheltron usage to stack up with lesser defensives to create greater defense overall for TBs instead of relying on a more available "ultimate" for them, and for gap filling defense during fluff damage since more of their CDs had to be used for TBs.
So if we focus on boss fights, completely ignore IB as part of WAR's kit and relegate the tanks to not being in tank stance, then they actually have a semblance of equability.
The main point of contention against WAR and particularly Holmgang is that because it is an "ultimate" and straight up prevents death as opposed to % mitigation and/or damage shields, it is able to sometimes cheese very specific scenarios because the stacked regular defensives have a damage limit that if surpassed results in death while Holmgang doesn't.
However I would say these situations tend to be more the exception rather than the norm, so it then brings into question whether the problem is necessarily the fault of the job/ability design or the content design. I personally lean more towards it being the latter but I see where people are coming from when arguing the former.
Then there is the situation of if the defensive CD kits balance out under the scenario lined out above, then what happens to that balance when you throw in the availability of IB into the mix. Yes it is not something that players will want to use when "playing optimally" and things are going as desired, but it is an additional safety net that can help when things go sideways or you are really pushing into the unknown like in prog. So that is something that WAR has, even if reluctantly, that differs from DRK and PLD. I will add in the caveat that the flexibility of TBN and Sheltron does overlap somewhat with this and does provide a bit of a similar safety net, but since those abilities are also part of the basic defensive strategy of DRK and PLD there are questions around how much they can be relied upon as an additional safety net without negatively effecting base defenses as opposed to the more "in addition to" sort of safety net that IB can provide.
So I don't feel that it is really accurate to say that WAR is lacking in mitigation compared to PLD and DRK because of IB being stance-locked and at the same time I feel that some of the rhetoric on how overpowered WAR's defensives are because of Holmgang can often be a bit overblown.
If I had to point to any one thing that I feel comfortable saying is ridiculous and op in regards to WAR's defense, it would be IR+Inner Beast/Steel Cyclone spam. I mean you will rarely ever use IR+IB spam and only in dungeon mass pulls for IR+Cyclone, but when you do use them it is crazy. Seems like a pretty nasty design oversight imo.
Yeah, Holmgang was definitely originally designed as more of a thematic ability, I mean just look at the name, that then got invulnerability shoe-horned into it. I'm not totally down on what it has become but I would shed no tears if the whole pull, root and needing a target bits went away, it got renamed and it kept just the invulnerability part.
As far as making IB not stance-locked and reworking HG, that could work but there would be a few considerations that would have to be dealt with.
First being the recast time of HG and how the greater availability of IB to supplement defense would affect the overall state of WAR's defensive situation. Likely HG would need to have a longer recast time to compensate and potentially the recast times of other defensive CDs would need to be looked at.
Second is that FC and IB would then both be available in Deliverance and would be competing for gauge resources, leaving using IB over FC as still being a dps loss which leaves it as still being something to be avoided instead of something to be layered in like TBN and Sheltron. So it removes the whole stance hurdle but still leaves it a dps loss to really take advantage of its availability. PLD avoids this issue since both of it's gauge abilities are defensive in nature so there is no defense vs. offense trade-off and DRK avoids this because it utilizes two separate resources and proper TBN usage results in a resource switch-off which helps keep it's frequent use closer to dps neutral. If IB were to be stance,-independent, something would have to be figured out so that it can work around this inherent problem of IB and FC sharing a resource.
Agreed. They just seem like something that was added in at some point to have a semblance of variety, never done anything with and then just left to fester.
That is not a "spaghetti code" issue, it is purely a design mistake. If the root effect locks an actor in place, restricting movement including directional movement to turn around and then a designer decides to apply that affect as part of an ability but overlooks that it may lock them facing away from each other and therefore unable to attack, that is on the designer and honestly should have been caught in either self-testing that all designers and developers should do or in qa. Even if the designer scripted the effect themselves, that is still a design mistake and not a "code" mistake and especially has nothing to do with the issues implied by the term "spaghetti code", being that the code is so unorganized that it results in breakages and bugs.