If you think five-minute-workday DMs are accommodating (and I agree), imagine what one would call the DM who lets the whole situation in a "Masters of the Realm"-level game pause long enough for the party to retreat all the way to a home base where there's fortunately an NPC caster to bring back the dead?
(Of course, if anybody could cast teleportation circle, there'd at least be a way to the NPC cleric's town other than a long journey from a hostile place while a man down, but, well, that's a 5th-level spell, just like raise dead or reincarnate . . .)
The same DM who would be giving full casters the same pause in the same situation because either our party of all casters has someone who can raise dead (ie a cleric typically but an all bard or all sorc party will have someone who can, true) or...
Teleportation circle requires gp (minor issue) and a permanent destination circle of which the caster is both aware of and possesses the correct sigil sequence. It's also a bard/sorcerer/wizard spell that bards and sorcerers don't normally waste a spell known on and wizards don't typically memorize but might have in their book (maybe) and need that long rest to prepare. That NPC cleric isn't any more a device provided by the DM for wizard than he would have been for a fighter, lol; it's a contrivance equally applied to full casters and non-casters alike in your example.
But if the DM is providing that NPC cleric for the wizard he/she/they is going to provide it for the fighter and place him within a day's travel to match the wizard requirement anyway. And if the DM isn't going to provide that NPC cleric to the fighter he/she/they is not going to provide the cleric to the wizard.
If the choice of the DM is to provide that NPC healer only if the party carries that one particular spell that is an example the DM of forcing it, and even then it's not "full spellcasters" it's a wizard specifically.
I find that a PC dying at high level isn't more likely than dying at lower level just because raise dead becomes available. The campaign in the OP runs the second and third tier so arbitrarily arguing that a spell not readily available to all full spell casters will somehow make a difference in the third tier when it wasn't available to any spell casters in the second tier doesn't make much sense. The same conditions apply in both tiers. Arguing that all full casters have an advantage because some full casters have the spell doesn't make much sense.
That's preconstructing an issue where a spell might be needed but contradicted by the fact it didn't matter before the spell became available, and misapplying a benefit for some full spellcasters to all full spellcasters.
All right. What's the feat or Battle Master maneuver or whatnot that counters "The enemy wizard cast a hemispheric wall of force to split the party, so his allies could defeat the party in detail" again? Or would only a "gotcha DM" include one 9th-level wizard (using his spells as if he really had a high Intelligence score) among the foes of a 13th-level party?
Wizards are a player class and not an NPC monster in the first place. That's an example of a DM building an encounter outside of the monster manual and deliberately choosing a spell he/she/they believes with be highly problematic. Yes, that would be a gotcha DM because it's a deliberate choice attempting to target a perceived party weakness.
The monster manual does have the archmage, however, and that NPC spellcaster does carry wall of force. The archmage also carries counterspell and globe of invulnerability to cause issues to your all spellcaster parties.
The archmage is CR12 monster and starts battle with stoneskin, mindblank, and mage armor for 15 AC with 99 hp.
The first issue with your example is you have assumed the archmage won initiative. That's possible with a +2 bonus but I would argue the party of 5 fighters would have at least one DEX based with +4 or +5 to initiative at that point. The archmage also has a passive perception of 12 so surprise by a party of fighters is a definite possibility. I would argue that the fighters go first more often based on the low passive perception and low bonus to initiative the archmage has.
After not ignoring suprise and initiative, the archmage spends his first action on casting wall of force. No more actions until next turn. This might stop 3 out of 5 of our fighters from attacking but the other 2 are going to action surge a difficult encounter like that and deal a ton of damage. Dead is still the best condition a party can inflict. More often that archmage is looking at getting attacked 7 times (once by each fighter and a second time by 2 fighters) before inflicting any damage, and sometimes that archmage is going to get attacked by 2 fighters before inflicting any damage.
Do you really think 2 fighters cannot action surge burst damage 99 hp on AC 15 at that level even with stoneskin up before the archmage can stop them? It's one of their strong suits. Even if the archmage manages it that leaves 3 fighters to continue dealing with after the wall of force goes down.
That's before looking at tactics like approaching from 3 or more different directions to spread out and surround.
That's before closing into counterspell range because that party will have at least one eldritch knight for that option.
That's before the archer style battlemaster uses a ranged disarm attack to disarm the casting implement and a couple ranged push attacks to separate the archmage from that implement to prevent casting spells with material components and limiting options (including wall of force).
That's before resorting to something like the intimidation proficiency on the fighter class list to pull something off.
That's before less orthodox methods like tackling the archmage for a grapple and dragging him into a pond where he cannot speak to cast spells or starting a plains fire moving with the wind in the archmage's direction to use as cover moving in for the attack.
Don't get so stuck in your box that you forget there's an outside to it. ;-)
Warlocks are iffy for the single-class campaign because their narrow class spell list at high levels doesn't have the sort of broad utility that you find in the lists of other full casters. Resting a lot doesn't do anything to change the spell list.
There's a reason part of 5th edition's fix for full casters was reducing high-level spell slots; many high-level spells re-define the game. One of the OP criteria for the single-class party was capability, and there's just huge amounts of capability in the spells full casters get at Tier 3 that just can't be matched by anything else . . . except an accommodating DM handing out powerful magic items that duplicate the spells.
Warlocks pick up spells from their patron lists from which to select as well, have invocations in addition to spell casting, generally don't suffer the same low single-target DPR issues most casters have, and don't need to take that 24hr restricted long rest to keep on keeping on.
Warlocks typically know more spells than sorcerers and are similar to bards in that number known because of the extras often picked up in invocations. A spell from a list that is neither known nor prepared is irrelevant as it cannot enter gameplay at that time other than rituals. A spell on a preparation list might come into play later, and that is an advantage, at the expense of giving up another spell normally considered more useful by the fact it was selected as a priority. Again, a person cannot take an advantage some full spellcasters have (prep) and apply it to all full spellcasters.
I don't disagree that there's capability in the higher level spells for some classes that you seem to have applied to all high level casters. I'd call schrodinger's on that because the likelihood that any class would have all of those just because the class falls into the full caster category doesn't exist. Most of them need to replicate the wizard's capability along those lines by grab wish or an equivalent at 17th or 18th level, and warlocks can do that too with the genie patron.
I disagree that the need to hand out magic items to compensate exists because the DM never needs to hand out items at all. 5e magic item crafting rules lets anyone craft them (the process requires a recipe, not a spell caster) and the game was designed with magic items as optional in mind.
I don't deny that high level spells are useful. My vote was for bard and part of that reason was magical secrets.
I'm going to share an anecdote of the inherent flaw in spell casting. I was DM'ing for a group in which my daughter was playing (a fighter) and there was a wizard who blew through a lot of spells in an encounter and wanted to camp for the night. My daughter looked an him and said, "We have things to do and a village to save. It's not their fault you suck 23 hours a day just so you can show off. Suck it up buttercup we have a job to do." And off she went with the rest of the party while the wizard adjusted his behavior a bit moving on.
High level spells are great. They are useful. But they are not everything and most arguments along these lines resort to cherry picking those benefits while downplaying benefits like extra attack, and like I said above taking benefits from different full casters and applying them to all full casters when it's mostly the wizard spell list.
