Fighters and Wizards from 1e to 3e
Simply (or rather unsimply now that I'm at the bottom

) put:
1 - 1e UA Fighters were combat monsters. Absolute monsters. They had unmatched (i) Saving Throws, (ii) HPs, (iii) AC, and (iv) ridiculous offense: opponent HP ratio. They devastated foes before them. Their lack of out of combat resources didn't really matter, especially if you were primarily playing Pawn Stance Dungeon Crawls where everyone had a role and JackFighter001 was there to stomp faces, full stop.
2 - 2e Weapon Spec Fighters (especially after C&T Fighters) were even more powerful than 1e UA Fighters. Again, they had unmatched (i) Saving Throws, (ii) HPs, (iii) AC, and (iv) ridiculous offense: opponent HP ratio.. They devastated foes before them. However, the offense: opponent HP ratio was even better for 2e Fighters. Further, if you used the "Heroic Fray" rules, which by design were invoked player-side, they also stomped hordes of mooks/minions in a way that matched Wizards AoE potency. Further, Kits and an expanded proficiency system that allowed them to trade in (what really became superfluous weapon proficiencies after a few levels thus they became fungible assets to round out an archetype with out of combat functionality) weapon profs for non-weapon profs created a situation where monster combat Fighters actually were functional out of combat as well.
3 - Then we have 3.x. On all of the major metrics, they lost ground. They went from having the best (i) Saving Throws to having terrible, terrible Saving Throws and being inherently vulnerable to AoEs and mind-affecting/fear attacks. They went from having unmatched (ii) HPs (and having those HPs serve a considerable, functional purpose) to having several classes achieve parity or exceed them (especially considering "functional HPs"). The same goes for (iii) AC. Finally, their (iv) ridiculous offense: opponent HP ratio was hit nearly as bad as their Saving Throw paradigm. Monster HPs inflated dramatically and mean Fighter damage didn't close the gap to achieve their prior status. Whats more, other classes did close that gap (and then some, especially with specific builds). (i) and (iv) were the biggest net losses for the Fighter with respect to prior edition iteration (both PC

C caste and Fighter:adversary challenge math).
It can certainly be argued that low level UA and weapon spec fighters in prior editions were overpowered at lower levels. I don't know if the 3.x designers had this in mind when creating the 3.x Fighter, but from 2e weapon spec/proficiency system/C&T Fighter to the 3.x Fighter, the power gap and the out of combat functionality gap was enormous. Just gigantic. Further, during that exact same time, the Wizard, already quite powerful in 2e (and world-altering at higher levels), went from quite powerful to:
- Having the Save DC math perturbed in their favor. What's more, the spell resistance paradigm changed as well in their favor. What's more, they now had means (Feats and Utility Spells) to perturb these paradigms in their favor even further.
- Having their spells:level ratio scale like never before (just as class mechanics).
- Having their already perturbed spells:level ratio perturbed considerably further with the magic creation system (especially bog standard Scribe Scroll at 1st level). Any Wizard whose wealth by level or Transition Scene asset strategy didn't involve Pearls of Power, Headbands of Intellect, DC neutral Scrolls and Wands wasn't trying. It would be akin to a Fighter wading into combat unarmored with a club. Suggestions that Wizards should eschew these assets ring as logically teneble to me as the Fighter sans armor + club strategy above. If we're making the leap of logic that what has unfolded in the world is a "contingencies against Wizards" paradigm, then it seems certainly the causal mechanism for this (that PCwizard001 would seek and attain, mundane items that are their + 1 sword and armor anaogue) should be expectant as setting default.
- Many spells in 3.x flat out became more powerful.
- The introduction of the "Ability Score Damage Paradigm" created a scenario whereby Wizards could leverage a new vector of SoS/D. Average Dexterity? Ok, you're dead.
- Spell disruption mechanics that, while still allowing for process, are trivially (utterly trivially) bypassed.
All of these things gave Wizards a situation where spell load-out wasn't a constraint anymore (so the offense:defense:utility spell load-out consideration became pretty close to nonexistent), they had more vectors to best disable an opponent (Ability Score attacks or attacks against specific, horrendous STs), and a better ST paradigm than ever before, with a wider and mroe potent swath of utility spells than ever before.
Meanwhile, the Fighter was looking at his former self (the former master of combat with UA and 2e weapon spec) and wondering where that guy went.
Logical extensions of setting legitimizing GM-force as "in the rules"
With respect to all of the various GM-force scenarios, none of those are "in the rules." There are various developed techniques and guidance on how to manage problems in core books, setting bokos and modules but these aren't transparent, core rules elements. Further, they are all subject to table social contract and many of them outright advocate the Calvinball strategies that we have all developed to deal with our spellcasters. When I was deploying them in 2e at high level when Wizards became Gods, I knew they were Calvinball and I knew I was using GM-force.
Rules are set forth to be, hopefully, as clear and as transparent as possible so interactions cannot be misinterpretted or over-leveraged for good or ill. The more opaque the ruleset, the more fuzzy the math, the more opportunity for GM-force there will be. With respect to opacities and silence within the ruleset, a GM's job is to adjudicate rulings where the ruleset falls short (due to this silence, opacity, or just poorly written prose). Whats more, their constraint in these situation isn't the authority vested in them by the ruleset. The authority (breadth and scope) of Rule 0 is therefore constrained by the authority vested in him by the players and the social contract of the table. At any table that is not a Call of Cthulu like table (where the GM is leading the PCs by the hand, through the story, their build choices just being set-dressing rather than shaping the narrative via interfacing the action resolution), it is fundamental that the GM is to adjudicate *** in the spirit of "fair-play" and use the most simple and coherent interpretaton with respect to designer intent. Use specific over general and apply the rules fairly and openly. This creates trust and vested players. All I am submitting here is that leaps of logic such as
- High fantasy/magic couldn't equal "very old world with battles of gods/primordials/devils/demons/angels and many fallen empires over the ages being causal for the influx of magic...and the PCs are unique, big, damn heroes in the here and now" (a fully functional narrative) but rather must equal "wizards, wizards everywhere and not a baron without a court mage" (the alternative which specifically allows for/justifies adversarial GMing)
plus a consistent, campaign-long output of "setting-driven complications", "adverse rulings where the rules are silent/opaque/incoherent (such as the awareness of the deployment of Charm by the Charmed)", "win-condition subordination with punitive obstacles or outright shutdowns" is very likely to demoralize Wizard players. They're placed in the uncomfortable position of feeling that *** above has been turned into "in the spirit of required adversarialism to the Wizard in order to maintain campaign and screen-time balance and the interpretation/logical leap to support that end" and the well of vested authority may begin to run dry. If it happens very, very sparingly, ok. If it happens with any frequency, expect Wizard players feeling like the GM:Wizard *** relationship is the latter rather than the formal; eg Calvinball. And it won't be pretty. And I don't think its fair to indict that Wizard player as "entitled."
Why GM-Force applies to GM-Paced, covert, GM-contrived win:loss paramaters and does not apply to Conflict Resolution Frameworks with framework:GM: player-synthesis-paced, overt, rules-prescribed win:loss condition
Finally, and again, in a conflict resolution scenario (such as Dugeon World and 4e) whereby a framework is in place that specifically promotes layered, dramatic, conflict resolution, this problem won't exist. You're deploying thematic resources, complications being generated from them is the mandated form of play and explicitly in the rules. Louie the Lizard, the obstinate chamberlain, the Ritual component pilferers (or any analogue thereof) regularly showing up to complicate things is fundamental to play. The GM is 100 % deploying complications overtly within a transparent framework that not only expects, but demands subordination of strategic power-plays inducing a "press-button win condition". There is no trust issue or vested authority compromise. There is no "is he being fair and reasonable or is he playing Calvinball to protect plot trajectory or to equillibrate screne time/protagonism?" There is only stakes-framed conflicts, thematic resource paradigms, and conflict resolution centered around PC decision-points > PC deployment of those thematic resources (so that preferred archetypes emerge through play) to impose their will on the evolving fiction/situation/conflict > GM considering intent/stakes-driven complications that account for what just transpired and the aggregate to date > and ultimate resolution once this process has effectively met the "conflict won/conflict lost" condition and the players plans are either furthered or set-back.
In other words, there is no tension of:
- GM interpretation/rulings in moments of rule silence/opacity/incoherency...with fidelity to fair play, honoring the player-vested authority granted to them.
- GM responsibility for dramatic pacing and establishment of a
covert win:loss condition...that may be utterly at-odds with "press-button win condition"...and thus may be willfully subordinated.
- A class whose entire shtick is predicated upon a "press-button win condition" paradigm.