The timing window on the short recasts is actually backwards. During Stormblood, we had some discussions on the tank forums around making cooldowns more timing dependent. The classic example is where you only need Rampart to mitigate a single hit, so you pop the action 18 seconds early. That makes your effective recast 72 seconds, because the cooldown starts trending down earlier. A really specific case of this that I used as an example back then is where you use Living Dead shortly after pull on God Kefka to mitigate both a 0:21 Hyperdrive so that it will also be up for a 5:17 Ultimate Embrace.

In the case of something like Holy Sheltron, the equivalent would be if Knight's Resolve activated on the last 4 seconds of the effect rather than the first 4 seconds. It's a bit like those Mario RPG style games, where timing a button press mitigates more damage than you would normally. It also means that you can do more fun things like tankbusters without castbars and interesting animation tells/fake-outs/roll catches. Unfortunately, it seems like they implemented the principle backwards without really understanding it.

Quote Originally Posted by ASkellington View Post
...
I personally think Clemency's implementation has problems. Hotbar space is at a premium such that to add new actions you often need to remove some. We're getting to a point in jobs' development where expansions tend towards animation updates on pre-existing actions (i.e. Broil 12). Having a button that you don't press outside of progression/emergencies isn't effective use of hotbar space. I think it's fine for that functionality to exist, but it should also have a reason that you want to press it under normal circumstances.

A compounding fact in this is that Endwalker indirectly devalued Clemency in 6.0 by adding lifesteal(!) to Holy Spirit/Circle, which already has half the MP cost. The end result is a button that you don't want to press under most circumstances. I think what is missing is a conditional proc that intermittently makes it damage neutral to give it baseline value. That in turn would open up the option of moving PLD's sustain off of passive lifesteal effects on their offensive spells to deliberate use of Clemency.

The Abyssal Drain/Carve and Spit merger wasn't really about introducing trade-offs in single target healing, which DRK has been historically terrible at anyways. It had more to do with the fact that DRK's gameplay has become progressively oGCD heavy with relatively few GCDs, so it was a way of cutting back on the amount of actions that you need to weave in during burst. Again, if you made Abyssal Drain into a non-damaging healing spell that is a potency loss, people are unlikely to use it. You also create this bizarre thematic mismatch where Paladins are siphoning life from their enemies with holy magic while Dark Knights are healing themselves up with non-damaging dark magic.

I think defensive actions should always have some condition under which they are damage neutral. You want to reward players for using their actions cleverly, not punish them for choosing to press a button on their hotbar. Otherwise you're better off removing the action altogether. I think what's missing is the 'clever use'. Right now, heals and mitigation are just sitting there waiting to be used. I want to see more ways to reward players for getting the timing of these actions right, instead of passive effects that occur automatically and reactive effects that you can use whenever you feel like.