I don't follow your logic here. Almost all those feats were terrible because they didn't balance well against other feats. +2/+2 wasn't worth the price of admission. Gaining training in a single skill wasn't in 4E either, but aquiring a new skill was often a worthwhile perk as part of a different feat that operated more like a bundle. Dodge was awful in 3.X, but the only way they could manage the over-the-top stuff like Uber-cleave and Spring Attack was to essentially levy penalties to be paid via bad feats on the way up to the shiny stuff at the top of a dependency chain.
Designing and balancing when abilities come in bundles is a lot more flexible and feasible. Smaller things like gaining a language or a circumstantial bonus can be included in this system much more easily and less painfully (for the players) than depending of prerequisite chains spanning multiple levels to justify larger and smaller abilities in a system with more granular player control.
- Mary Lund
It was a snark remark on your idea that combining the smaller parts of which a mega-feat is made, would open design space, i.e. that it is an opportunity for designers to fill splatbooks with feats that grant X+Y+Z, A+B+C and then X+Y+A, X+Y+B, X+Y+C, X+Z+A, X+Z+B, X+Z+C, and so on and so on...
If that's a good opportunity based on having mega-feats that are easily "disassembable", I want to point out that we already had them disassembled in the first place when each part was a single, smaller feats.
This is why it reminded me of the +2/+2 feats: once they realized they could easily make more like the original Alertness, we started to see lots of them, while it would have been so much simpler to have one feat granting +2 to two skills of choice.
---
I understand your idea that designing a bundle can be better for balance, for instance those Dodge+Mobility+SpringAttack together would have less problems if e.g. Mobility is used more rarely or worth less, as long as the whole package is worth right. A weak and a strong "sub-feat" together would compensate each other.
This is the way classes are designed, so that they don't have to worry if wizard's spells are much better than a wizard's knowledge bonus.
What I want to point out however, that this may not really be the best approach to feats for 2 reasons:
1) Feats are (or at least were, until now) meant to customize your PC on the smaller level, kind of like
fine-tuning. Maybe one PC wanted to learn a new weapon, another wanted to shoot 2 arrows in the same round, the other one wanted to be more protected against poisons.
5e had a great idea for those who
don't like fine-tuning: Specialties. Want to become a weapon-master, a sniper, or a poison master? Take the whole specialty, and let it suggest feats for you.
With mega-feats, we simply lose this level of fine-tuning (unless we want to wait until supplement #38 comes up with the "right" combination of sub-feats). We do not gain anything new, because we already had those "bundles" in the form of Specialties.
2) Classes, which use the "bundle" design approach, grant their abilities
gradually over many levels. Feats grant their abilities immediately, all at once, so if we now have mega-feats, we are going to have significant complexity bumps whenever our PC gains a feat. Of course, spellcasters get similar (or bigger) bumps whenever they gain a new spell level, but still this is not so nice IMO, if we could just have the same spread over more levels (this assumes the rate of mega-feats will be 2-3 times slower than the current rate of feats by level, but it might not be so...).
They said before that mega-feats will be used to represent prestige classes. But the whole point of prestige classes was the
progression of getting those abilities. Not that the 3ed prestige classes were usually a great design (they were not), they suffered from the opposite problem of being too "empty". Still, Specialties were much better in that regard, because they traced a progression rather than getting a sudden bump.
