What makes a good Job mechanically?

[Shurrikhan]

Low floor, high ceiling.

Since the first often inadvertently helps the latter, that's not as hard as it might first seem.

Or, another way to put it, it's a kit where (A) the cost of any single failure is low enough, (B) the net reward for more complexity (itself a risk) is high enough, and (C) optimal actions are contextual enough... that most players aren't phased towards a singular or rote understanding of the job's intended playflow despite its components all being quite intuitive.



What I think a lot of people miss is that difficulty in terms of stringency or rigidity often harms (lowers) the ceiling as much as it harms (raises) the floor, despite there typically being a greater "skill gap" (or, variation in performance from successfully avoiding the most punishing fail conditions).

Personally, I don't mind a person flailing around getting some 60% of the throughput of complete mastery on that given job. What I mind is simply when the rewards of engaging with the minutia are so low, despite the much greater complexity in optimizing them, than the core elements that a player trying to actually optimize some part of those is likely to perform slightly worse (at least at first) than those following a painfully oversimplified/rote idea of what the job should do. We should feel encouraged to go that step more, but it's fine for it to have rapidly diminishing results.

:: Yes, there are some obvious analogs here to the "2 minute meta," prior to which there were a greater number of precise timings to match (your separate 1:00 and 1:30 raidbuff recasts, etc.) and therefore slightly instances of / opportunities for optimization, but the costs of failing each or especially the largest overlaps of raidbuffs were far less. While we want some performance gap, certainly, it shouldn't come all from something so simple or, especially, with little ability to be salvaged.

Of course, any discussion of kit is generally incomplete without also having a discussion of encounter design. Without some degree of stringency (without also being too stringent), and a total number of mechanics across the instance, enough for different job features/profiles/etc to see different results... kit differences and their affordances end up paper-thin.