I agree to some extent, some games more than others. Stretch goals like platform support, modding tools, paying developers more, I think should have priority in some genres of games. I wasn't a fan of the dungeons levels and extra classes as stretch goals in Project Eternity, I think designers should have control about that. In terms of the dungeon, you can just make each level smaller, so you have some room to manoeuvre.
I think RPGs are different in the way they scale to other types of games, adding more writers and more programmers scripting to.flesh out Wasteland 2 and Project Eternity isn't necessarily going to make a RPG bloated, it's going to make it awesome. 6 axis shooters within a large universe like Elite and Star Citizen will also scale really well, you're not going to ever fill up that world too much by adding too many content creators.
The Defense Grid team did a great and realistic campaign in terms of stretch goals, their model wouldn't have resulted in the problems Andy Schatz worries about. I hope people take note of that kind of planning, and apply it to future Kickstarter projects.
Double Fine Adventure is a classic example of not being able to deal with going above expectations, at least as a campaign. Double Fine I think just said "you trust us", we;ll do something with this money, and for the most part I and other backers trust them.
I don't know whether it's possible, but I'd like developers to stop the stretch goals when they're not sensible any more and just say, we'll use spill over for the next totally DRM-free project.