I agree. I'm not sure how they do that with such a core part of the game, but I suspect its possible.
Nod. I supsect it isn't. But, then, I didn't think it'd bee possible to balance fighters & wizards in D&D, and 4e went and did it. So I have a record of being pleasantly surprised.
I think there's general agreement that starting from simple and adding complexity is the way to go. I think its an open question whether there is a similar "direction" between CaW and CaS.
I'd hesitate to claim a 'general agreement' around here.

What I'm sure won't work is starting with limited options/power, and bolting-on new options that add power. What might work while retaining some balance is starting with a complex framework 'behind the curtain,' and using it more as a design standard to build simple-seeming classes with most choices already made. Then, in modules, not adding complexity, but revealing it. In that way, the 'simple' character can be played along side the 'complex' one without it being at a great disadvantage.
The perfect example in 4e is Backgrounds & Themes. The game didn't have them initially. When they were added as options, the characters that opted for them gained additional benefits over and above those that didn't. While those benefits were mostly pretty minor, it's still a pattern that needs to be avoided, because, if multiplied by the many modules we should probably expect from 5e, it could result in vast gulfs in effectiveness between 'simple' and 'complex' options.
More concisely: 'simple' must not be inferior to 'complex.'
I'm only speaking of superiority in effect, not RP experience. You can be a great fighter in the ring, but the guy who sneaks up on you with a shotgun still wins.
Nod. He also gets the gas chamber.
CaW is presented as about 'thinking outside the box' and the like, finding ways to sieze. In other words, finding overwhelming advantage or exploit weakness. Cheating, in the context of a game. When it's happening in the context of the imagined world, that's fine in an RPG, even though it's a game, you're not cheating, you're following the rules to model your characters 'cheating.' When it's happening at the meta-game level, it's just powergaming, often quite disruptive and undesireable powergaming.
The former sort of in-character CaW style is well-supported by a balanced - and detailed - game system. The latter sort is supported by a poorly balanced system, and not, IMHO, at all desireable at the table, unless there's a unanimous enthusiasm for it.
My original point was that in earlier editions, magic was the shotgun.
Magic was often 'broken' yes.
I was expressing my opinion that most of the things that you can do to make D&D's abstract combat system more interesting for the fighter tend to be CaS. If magical combat effects in 5e go back to more CaW effects (Stopping Time, Holding, flat out Save or Die, etc.), then I suspect the LFQW problem will resurrect itself. Hopefully, I'm wrong.
There is nothing inately in-character 'CaW' about broken or overpowered or badly-written magic systems. Yes, they make the metagame take on a 'CaW' feel, as everyone scrambles for the most broken thing, and the game devolves into a sort of arms race.
To make the fighter and wizard work with either a CaW or CaS aproach in-game simply requires a balanced system. And that requires looking at casters as protagonists in a story, viable choices within a game, or established technologies in a theatre of war.
4e Characters are CaS machines because that was how they were intentionally designed, its one of 4e's selling points.
They were intentionally designed to be balanced. Balance supports both styles. Imbalance only suports metagaming, and makes it boring into the bargain if the imbalance is extreme.
I'm afraid you're conflating something. D&D /is/ a game. Of course it will need to have game balance. That's analogous to the 'fairness' of aproaching combat as a 'sport' - a structured contest with rules, but it does not force that style of play within the imagined world the game represents. Rogues may be balanced characters in the metagame, for instance, but to the imaginary monsters they stab in the back, they must not seem to be 'playing fair,' at all.
And, just to be clear. THAT ISN'T INHERENTLY A BAD OR GOOD WAY TO PLAY D&D. However, the fact that 5e is even attempting to "reunite" the player base or whatever-ya-wanna-call-it, indicates that that didn't go down so hot with the D&D public as a whole.
Every edition has had it's hold-outs. 3.5 just had the SRD and Paizo waiting to cater to those hold-outs. That's the only difference between the rejection of 4e by 3.5 hold-outs, and the rejection of 3e by AD&D hold-outs.
Maybe, but I'm not sure who's using it as a litmus test for a game. Part of the mechanical design issue is that a lot of CaW takes place outside of combat (sabotage, setting up ambushes, subverting allies, etc.) So, to some extent, CaW is served by making the Fighter more viable in the other pillars.
That would be awesome.
Oh, I mostly agree with you here, and its part of my point. The fundamental problem with the way D&D has traditionally viewed Wizards....as PCs/protagonists. Wizards as anything but a Deus Ex Machina or villain are a relatively recent phenomenon in genre, IMO. So, in a fairy tale, a grumpy enchantress turns you into a frog until you get a kiss from a princess. That's not Combat as anything, but take away the narrative context, and Polymorph is suddenly a CaW effect.
Magic in narratives where the Wizard is the protagonist tends to be much more limited, unreliable, and often has tremendous "backlash" of one kind or another.
I'm not sure I'd agree. Magic when it's being used on the protagonist's behalf by a supporting character is often that way, because the story has to be about the protagonist, not the magic that let him easily win through. When magic finally makes it into the hands of the protagonist, it often becomes much more reliable and understandable, and much less powerful. If it remains powerful, it's generally also in the hands of all the protagonists 'real' foes, and also any allies or co-protagonists in an ensemble.
In other words, magic has to become balanced in some sense to become the tool of the hero, rather than the tool of the plot, the helper, or the villain.
In 4e, magic was made balanced by making it no more powerful (though still more versatile) than non-magical heroic abilities. In Harry Potter, magic was balanced by making it the focus of the story - everyone who mattered (anyone who might be PC) had magic.
D&D, coming at the early stages of mage-as-protagonist made the understanible and disasterous mistake of giving PC magic-users antagonist/plot-device level magic, and trying to 'limit it down' to protagonist-apropriate levels (actually support-character-apropriate). And the game has been an imbalanced shambles for the longest time as a result of that initial error.
3e made magic much more reliable and removed almost all the "backlash" from previous versions...exacerbating the LFQW problem.
Very true. The AD&D caster was barely-playable. The demand of healing on the Cleric kept it from being anything but support, the physical weakness and profound limitations on casting made it hard for the magic-user to contribute consistently (or at all at all levels). 3e took note of the problem and made magic more useable and consistent, as required to model a 'protagonist,' and thus also more playable. It just failed to dial down the power to match, and the result was the optimization tiers and endless 'Fighter SUX' threads on the WotC boards.
4e finally achieved balance by further removing limitations from casting, but bringing casting down, and non-casting up, to the same level of 'protagonist-apropriate' power. An elegant solution, rejected for being different. As any improvement, sadly, must be.