"Illusion of Choice" depends on several other factors, namely:
- The breadth of "real" content (flat)
- The relative narrowness of content considered "real" content (percentile)
- Individual players' perceived importance of "real" content (e.g. does it matter if something isn't optimal in Savage if I only like running dungeons?)
- The breadth of applicability of talent compositions across content ("real" content's most optimal choices can be more limited in breadth than the less optimal, but choices should still be applicable across the board)
- The performance margin required to be considered a significant difference
- The degree to which players focus on options available or options not available (flat vs. percentile)
- The degree to which players focus on the percentage of options that are available or the number of options available (flat vs. percentile)
- The number of choices within the accepted margin (flat)
- The ratio of viable choices to non-viable choices (or, the percent of total choices which are viable) (percentile)
Unless customization is done very, very well, illusion of choice is a certain outcome. Whether that will prevent the addition of customization options from improving the player experience, however, a much narrower and more difficult question. And only the latter is actually worth discussing when considering positive design or development efficiency.
Personally, I am not a fan of customization, at least as it is typically implemented. I loved the heck out of outperforming far better geared and more experienced players using "trash" talent build variants in WotLK. But, I think that was a matter of me finding the gameplay I really wanted and then making it work for myself despite the barrier that was customization (and, of course, a hefty share of smug vindication) than it was a love of theorycrafting for its own sake. After all, every choice a player makes for Option A out of A, B, and C, is essentially stripping options B and C from typical use.
I see this trimming of the toolkit as practical only "when necessary" -- when a rotation would otherwise be too cluttered in its optimal form or have just an incommensurate or frustrating amount of nuances and conditionals and sub-strategies that could muddle the aesthetics the job could otherwise focus on more clearly across its playflow. Until we get to that point, customization should not be given unto menus, but rather in how we choose to use shared resources between our skills, be it basic resources like MP or more unique resources like Gauge or especially unique resource like Sen, or special effect generated from spending certain Sen or even time itself, be it over a fight or a particular burst window or in conflict with CD-immediacy with another skill, etc., etc..
And while I can see how that might be
useful eventually, perhaps even within Shadowbringers, I do not believe such a system is
necessary. I mentioned it only because I wholly believe that Role Actions are yet another bait-and-switch, replacing one flawed system with another in order to change the conversation rather than having to fix the flaws in execution of the first system. It's not the first time.
- When we lost the open toolkit system (the original "Armory" system, whereby you could synergize skills from any proficiency [class] into and through any weapon), rather than daring to look at why the execution failed, they blamed the concept and said the "build your own job was doomed from the start!" and gave us jobs. "But look! Jobs!"
- When they failed to finish and then outright cut the former Battle Regimen (Skill Chains' + Magic Bursts' red-haired cousin) system and its combo concepts that would have given Stamina a decent pacing they said the Stamina Bar was a flawed concept from the start because it slowed combat, and replaced it with XI's generative TP, which was even slower, more AA-dependent, and brought us down from some 5-16 viable combat skills to choose from per GCD to at most 3 (combo A, combo B, and sometimes a reactive skill). But alas, "Look! Combos!"
- When they implemented a dual-job class without allowing for job-based EXP or job traits which could affect their class skills they said dual-jobs were fundamentally flawed. This was "A lesson learned!" (Never mind that it didn't actually have to go that poorly...)
- And now? "Additional skills were fundamentally flawed because they obliged players to level multiple jobs." (Never mind that you could have allowed the Additional Skills as adaptations, whereby one swaps Keen Flurry for an untraited Bloodbath or vice-versa, which would have greatly mitigated the issue.) "The Additional Skills spent buttons on skills of minor importance, wasting bar-space!" (Never mind that there is a difference between infrequently used and the truly situational; situational skills can be left off bars, while infrequently used skills remain obligatory, but waste their slots for all but that second of their use.) And so, we could have simply removed Additional Skills and trimmed "redundant" skills. But alas, we can't just remove a system without replacing it! "And what's that? We had skills that were basically the same as another job's skill? Well, no longer will we foolishly waste space on these redundancies!" (Never mind that the space saved is insignificant and the issue could have been better solved by simply... diversifying the previously uncreative skills.) And so we have another system, Role Actions, even though it fixes none of the deeper issues it responds to.
It's often western 'money-grabbing' companies we think of as throwing previous visions for games or whatnot under the bus and spinning the narrative in order to seem responsive to the playerbase, reactive to past failures, and progressive in their roadmap when in fact nothing's really been done, but there's scarcely a ploy that you can point at from some Western company currently under the spotlight that someone who's been on XIV from the start wouldn't have seen already in this game alone. Thus...
When it comes down to it, I don't think XIV would ever abandon some shiny system without replacing it with another. Even failure has to be sold, and ever bit of "owing up" to said failure glorified and added to the march. Thus, if Role Actions were to go, then there's a very, very high likelihood something would have to take its place. Customization of that sort, then, I just think would be the least intrusive.