Well, my run was hardly the most completionist, so I may have missed some things that further emphasize the issue (my late-game fighter weapon was the silver sword, and I couldn't be bothered to fully comb Act 3 for everything there is after a while), and I opted out of using Haste after some time because it made the game much too trivial, so I have to concede on that front. What I did find truly broken was the potion of bloodlust which refunds an action after killing something, especially with how late-game tends to throw trash enemies at you in combination with the properly powerful ones.
However, the calculations didn't account for GWM's pseudo-Cleave attack (extra bonus action attack after a kill/critical) which still can provide a 4th/7th attack even without Haste, and that BM can burn the superiority dice to outweigh the accuracy penalty from power attack (using 3e terms here because that's what I am used to), so the increased miss chance from power attack being on can be practically ignored, especially if you consider how easy it is to get Advantage in 5e (if there is one mechanic I would definitely consider much too simplistic and reducing combat to milking that one way to heavily increase your odds, it's definitely that).
As for the spells, the surface DC is certainly bugged to all hell, which made Evard's Black Tentacles practically useless instead of a reliable control tool. As for damage-dealing spells, their scaling in 5e somewhat hampers their usefulness by itself, while the nature of their effect makes using them without metamagic (let's delegate a very important and interesting caster mechanic to sorcerers alone, why not, 5e?) so that you can guarantee save disadvantage or at least on average better damage rolls a crapshoot compared to just whacking someone with a +2 or +3 weapon a few times. Even then, high AC counters the ranged/melee attack spells, high Dex saves which many enemies have make AoE unreliable (and if they have Evasion, practically useless), and high Con saves (every beefy boss enemy, basically) or being a construct/undead counter all the necromantic/poison-based sources of damage.
I am afraid I remain of an opinion that 5e inverted the 3e's caster idea of "start weak, finish strong", because regardless of itemization and balancing the spell DC barely increases and outside of being a sorcerer or doing a sorcerer dip you have little to no ways to circumvent higher saves and damage variance, except for the level 10 evoker feature (which Larian did implement wrong, actually, and the damage bonus applies to every instance of a multi-target spell). You don't have the old Empowered and Maximized Spell metamagic, you can't easily bump your casting ability up to increase DC, you can't really weaken the enemy's abilities except by imposing Disadvantage on saves.
A fireball with its 8d6, doing ~28 to all targets in range, far more likely to be ~14 because saves and very few ways the player has to bump the DC high enough, potentially as little as ~7 because resistances, when the game tends to use high HP enemies in smaller groups, is not nearly as reliable as even 3 attacks with a +2 weapon (assuming our fighter is min-maxed, it's 3 attacks of, say, 1d12 + 7, so 8~19, 18~29 with GWM, resulting in ~13.5 / ~23.5 per hit, times three, plus potential criticals, of damage output per turn)...
...alright, as much as I enjoyed the exchange (thanks!), it's getting off-topic.
