The only way that 4e is less gritty is when DMs hand out wands of cure light wounds like candy (which hasn't happened in any 3e/Pathfinder game I've ever run in).
Hand them out? You mean "when someone spends a feat on it"?
I mean, I agree with the basic point - 4e is less gritty, and this is by design. But wands of cure light wounds aren't some crazy thing that a DM throws at the players - they're a crazy thing the players are perfectly capable of doing on their own, using just the basic rules in the core books.
Lanefan said:
how many 3+ year 4e campaigns have you seen?
3+? Dunno, but that's a little bit dirty pool considering the edition was only released 4 years ago. (And paragon tier was rather broken until 2-1/2 years ago).

I have a Dark Sun campaign that's been going strong for 2 years right now, and certainly has enough potential to go 2+ more.
Right now, that's longer than my 3.0, 3.5, Star Wars Saga, or WFRP2 campaigns, and it's about tied with my CoC d20 and Arcana Evolved campaigns. (The former only ended because some of my younger players graduated, and the latter sputtered out due to AE's intensely broken spellcasters after 10th level or so.)
I'm sure all editions have their players at all levels. From what I saw higher levels were just number adjusted versions of lower levels.
Number-wise, that's true, and again by design. However, the capabilities of the PCs broaden continually, their abilities become vastly more effective, the threats they face are trickier, and the nature of the campaign changes if the DM is any good. (Heroic is "save the town/village," Paragon is "save the kingdom," Epic is "save the world" is how it's explained.) It's actually a lot like Next proposes to be; bounded accuracy is just 4e math with the treadmill turned off.
Further editing, re: challenging the PCs. MM3 math helped a
lot. That's why, up above, I called Paragon tier basically broken until MM3. Also, because as I mentioned PCs' capabilities broaden, L+0 generally ceases to be as meaningful by upper Heroic. You tend to need at least L+1 at low paragon, L+2 at mid- to high-paragon, and L+3 or more at Epic for a baseline. I am not saying this is good; it's a flaw in the design, but it's at least closer than CR/EL.
-O