While very little definite information has yet been released about 5th edition D&D, they've released some new information on mechanics in the past day or two that I want to comment on, based on some educated guesses made in the context of previous editions.
"Instead of the fighter getting a better and better
attack bonus, he instead gets more options to do stuff as he goes up in
level, and his attack bonus goes up at a very modest rate. I think it
offers a better play experience that the orc/ogre can remain in the
campaign, and people can know how the monster would work from a
previous experience, but they remain a challenge for longer." - Monte Cook
So let's talk about scaling bonuses - Warning, this is about to get mathy. Let's say you're trying to pick monsters for a fun and challenging fight in a 4e campaign. A monster becomes too powerful to use in a fight when its AC and other defenses are high enough that PCs need to roll very high (natural 17+, I'd say) to hit. There's also a band, which I'll arbitrarily identify as requiring natural rolls of 13-16, where the game might be fair, but it's mostly frustrating. For comparison, I'll note that we're talking about ACs of 19 to 23 to challenge 1st-level characters. Based on the Monster Statistics By Role chart on p. 184 of the DMG, this represents 3rd to 7th level for soldiers, or 7th to 11th level for brutes and artillery.
Is this a valid way to judge what you should use to challenge the party? Not remotely, but the point I want to make is that 4th edition paid a lot of attention to likelihood of hitting an opponent, and scaling AC and defense out of reach is a primary way of signifying that an opponent is too tough for you right now. The same is true, of course, of low-level monsters threatening the PCs; once they need high rolls to hit the PCs' defenses, they should really just pack it in unless they are there to slow down the PCs while a boss of some kind does the actual damage.
You reach a point, though, where it just doesn't matter how many low-level monsters you throw out there, because hitting only on 20s is not a valid threat. PCs reach the point of sneering at orcs, even dozens of orcs; a little later, they can sneer at dozens of ogres, and so on. Obviously, you always have the option of creating a new stat block that is a higher-level orc - that's entirely beside the point. You've outleveled orcs as a threat. The same thing happened in a different way, at a different time, in 3.x D&D. We regarded it, in both editions, as the natural order of things; in heroic fantasy, should the high-level champions of the realm be threatened by orcs?
5e seems to be saying, "Yes, you should still have to worry about orcs, at least in very large numbers." Orcs and ogres "remain in the campaign;" they're saying (I presume) that you could have an epic war against the orcs, because orcs don't (or don't quickly - we'll see how the math plays out) reach the point of averaging near-zero damage per round against the party's fighters.
So it's a really interesting goal, keeping a given statblock relevant over most of the range of play. This has a couple of drawbacks, though. People like watching the numbers go up, frankly. I think we can still expect hit points and damage output per attack to scale with level - certainly they've made it clear that wizard spells improve in damage output at higher levels. But if the attack bonus you have at first level is the attack bonus you're going to have for several levels, and you're never really going to improve relative to the scaling opposition (higher level monsters, not the orcs of my earlier example), then having anything but the best possible number in your attack stat (for fighters, Strength) feels like starting out the game in the hole.
This was a problem in 4e (discussed in more detail here) that the system never credibly addressed, to the best of my knowledge. Hit rating was the be-all and end-all of combat effectiveness, because 95% of everything you could do to affect an opponent involved an attack roll. 3.x somewhat avoided this problem by making non-primary stats more useful for more characters, and by making attack bonuses scale faster than the defenses players were likely to face. It's not clear to me what 5e is doing with monster defenses, but slow-scaling attack bonuses for players makes it likely to me that monster defenses will also scale slowly, one way or the other.
I think there's a reasonable solution to this, but first I want to address a tangential point in statting that I expect to hear as a counter-argument. Characters can be defined by their weaknesses, but I think the majority of players don't want to play characters whose primary stats as determined by class are those weak points, since that interferes directly with the character's most basic functions during gameplay. Weak fighters, clumsy rogues, slow-witted wizards - these are not archetypes that are going to catch on in any edition of D&D. A comedy-of-errors character in D&D either keeps his errors outside of his class function, or meets a swift end.
The thing about this is that it doesn't leave a lot of room for character variation, in comparison with fantasy and other source fiction, but I think there's an opportunity to tweak the 4e and 5e rules to cooperate a little better. 4ewas, and presumably 5e will be, balanced on the assumption that characters have an 18 in their primary attack stats.
What if the benefits of gaining levels included attack bonuses (at odd levels, let's say) until you have an attack bonus on par with what you would have if you had an 18 in that attack stat?
There are problems with this, I am aware - and I invite comments to that effect. The thing I like about the idea is that "this fighter is remarkably mighty" might be divorced, even slightly, from "this fighter is just better than any fighter with a lesser Strength score." Of course, fighters with a lower Strength score don't have the same damage bonus, and that's sort of a problem, but those fighters might enjoy sufficient compensatory advantages in other stats - particularly if 5e is better than 4e about making stats other than your class-and-build's primary and secondary stats useful.
Since I brought up the point of the source fiction earlier, I want to mention what I specifically had in mind: The Three Musketeers, or really the d'Artagnan Romances as a whole. (Also, their fantasycounterparts.) Porthos (and his counterpart) are noted for their great-verging-on-superhuman strength. Yet they are capable but not objectively superior fighters; Porthos does not outshine any of the other three in actual fighting ability.
Obviously, Dumas had no rules set he was attempting to imitate, and I don't intend to imply otherwise. My premise, though, is that you could have two or more fighters in a party (much like Pendragon) with different Strength scores, without one perpetually living in the shadow of the other, at least as far as their class was concerned. The later introduction of magic items, such as gauntlets of ogre power, might cause some really odd things to happen here by re-equalizing the two, or by putting the stronger even further ahead of his comrade, but a quote from the DDXP seminar on 5e leads me to wonder how they'll be designing such items in the first place: "We want to decouple magic items from character progression so they're not needed, and return that exploration and excitement of finding magic items." So who knows?
I think this kind of rule works best in random stat generation; when a player can plan too definitely around it (as in point-buy), I expect that it will be trivial to find a solution that returns stat generation to a solved game, which is what I'm trying to get away from in the first place. If classes and builds have secondary stats, as they do in 4e, new problems come along, as you may now be able to spare better stats (rolled or purchased) for your secondary stat... and for Constitution, the universally-needed stat. It also really only works for stats that have two functions for a given class, much as Strength governs both attack and damage bonus. In 3.x and 4e, this is pretty straightforward - for example, Int governs save DC and bonus spells for a 3.x wizard, and attack and damage for a 4e wizard.
The simulation goal behind this idea works a lot better in SIFRP. Since so many stats contribute in one way or another to combat, and many weapons use a stat other than Athletics for their damage values, it's much easier to have mathematically comparable fighters that simply shine in different areas. This suggested rule matters a lot less in games with one fighter, one cleric, one wizard, and one rogue, since there's not as much room to directly compare oneself to another character. For that reason, it might be one of those modular rules that a DM only brings out in campaigns where it would be useful, such as an all-fighters campaign.