Wow, haven't seen this post before, and the OP brings up some excellent points, pretty much articulating exactly what it is that I, personally, do not like about the current endgame setup.

The fact that so many players cannot get past the mechanical hurdles required to complete encounters like Titan EX, let alone the 5th Coil, is something that will stifle the game going forward, not to mention it is actually quite boring once you do memorize the dance as you simply repeat the same steps over and over again. Having the entire party suffer for the mistakes of only one or two other players is simply irritating, especially since, after you explain the mechanics to them, there is literally nothing you can do if they just keep failing at it.

And yeah - the gear is irrelevant. Once you have enough to clear the DPS checks it literally does not matter anymore, the only thing that matters is whether or not absolutely everyone in your party has memorized the dance - if they have, great, you're done! If not, get ready to practice for hours on end until they either figure it out or quit! What fun (hahahahahahaha)!

I REALLY like the suggestion of tuning the mechanics to cause gradual difficulty increases after repeated mistakes, with a smaller number of skilled players able to carry through a number of other players if they're good enough. This kind of tuning would also allow for smaller parties of skilled players who would be able to gear up faster since they have fewer other players to roll loot against, which would really benefit the elite players while not disenfranchising the newer ones. I understand the concern over players gearing up too quickly, but it's just like the OP said - if the content is engaging enough (which would definitely require some randomized elements in combat), people will play the content for months, if not years on end, even without a carrot on a stick at the end of it all.

I really hope these suggestions come through to the development team. I think the game could only benefit from implementing at least a few of his suggestions. Here's to hoping 3.0 proves itself a positive change in direction, and not merely more of the same!