Time is supposed to be more or less flowing in combat, happening in a sort of simultaneity that is abstracted for functionality. If characters can delay their turns, then they can order themselves exactly as they want to in the initiative, which will cause them to inevitably group all together in the exact order and optimum rhythm they desire, every single combat, every time, without variation. At this point, initiative itself has become redundant and has no purpose to existing at all. Being able to delay your turns would make combat even more formulaic and boring than it already is.
Being able to wholesale delay your turn also has substantial ramifications for spell and ability abuse, many of which revolve around the timing of turns, which being able to slide your turn down the order past other enemies destroys the balance of.
In 5e, there exists the Ready action, which is limited and does not actually shift your turn in the initiative, and this is for very good reason.
I'm in support of 5e's ready action being properly implemented.... but that would also recall a proper implementation of reactions, so...
I agree with every word
I agree and disagree.
The part where I agree more is the risk that, without being forced to make do with a given turn order in each fights, some players could end up developing combos that they'll use all the time, making the combat a bit stale. There was also the risk of analysis-paralysis.
These are part of why Larian scrapped their initial plan of having all your team play, then all the enemies play, as I think Nick Pechenin explained in a video I watched at some point (btw, I wish Larian had Nick demo the game instead of Swen from now on ... showing off how you can win while having no clue about how the game works isn't the best way to sell the game to me ... though perhaps it's selling it well for different types of players).
Another part was probably that you could be exposed to taking a lot of damage on your fragile characters without a chance to react, and that whichever team would go first would have way too big an advantage.
I'm less convinced about the power/balance issue. Say that the turn order is Shadowheart, Enemy 1, Enemy 2, Enemy 3, Gale. If Shadowheart Delays her turn to next-to-Gale, this opens up combos. But maybe Shadowheart could have killed E1 (and perhaps E2, if dual-wielding of using AOE damage) on her turn. Not doing this means that by the time it's Gale+SH's turn, the team has perhaps received one more attack than it would have otherwise received. And this is a cost. Is that cost big enough to compensate to gain of having SH+Gale playing together instead of separately ? I don't know.
Also, if Delay was to exist, I think it should then probably cost a Bonus Action (or Action ?). This would increase the cost of Delay to "(B)A + allowing more enemy Actions". Perhaps it then becomes not-so-clear that Delay is obviously beneficial, all the time. Maybe the use of Delay becomes really situation-dependent. What I'm saying is : there are levers that could be pulled to try to balance a Delay action. Which may or may not be powerful enough, I don't know.
Anyway, all this is pure speculation.
But I would certainly welcome a Ready action.
For example, Shadowheart could Ready a Create Water spell and drop it when Gale's turn start. Basically, Ready is a form of Delay, but with an added cost : you forsake flexibility. If your plan was to cast Create Water + Ray Of Frost, with Delay, if Gale or Shadowheart died when the enemies played or they split or else, you can switch to making a new plan. With Ready, you cannot. Ready may also be more restrictive, I don't know if "When Gale starts moving" is allowed in 5E (PHB doesn't seem to forbid it). Anyway, I presume that 5E was play-tested more extensively than BG3, especially play-tested for combat, so maybe that's why the 5E designers had Delay/Ready be as it is in 5E.
