ok...again. 7 comes up because if you are a DM when you use creature sheets, it says "7(2d6)". The HP is 7. There is no minimum and no maximum. Their HP is 7. Thats it. End of story. Goblins have 7hp. Young reds have 178. Thats it, end of story, no questions. You don't roll 17d10 and maybe end up with a 255hp dragon. You fight against a young dragon with 178 hp. There is no range, ever. They have what they have. So saying 12 could be doubled to 24 on a max rolled goblin - there is no maxed roll goblin. There are goblins - they have 7 hp. A DM might custom tailor an encounter or a particular tribe, or whatever but that is a DM being very comfortable with the game mechanics and THAT SPECIFIC PARTY to say "they can handle this"or "this would be a shit stomp, lets bump gobbo attack to +5 and increase AC by 1 and add 5hp".
No, the 7 is there for 1 reason and 1 reason only, because if you rolled the HP for every single monster at a table it would be time consuming and so it gives someone a quick reference point if they do not want to do so. The monsters HP is a variable range, with a min and a max, if it was not, there would be literally no reason of providing the additional data (2d6) because it is a redundancy. As much as I assume you would like to be the arbiter of truth and all that is right in the world (your ego is certainly big enough for it), that is not the way things works.
You can't just do it and be like "yeah im pretty sure any party of 4, regardless of configuration, should be good with this". The game just doesn't work like that. If you look at rollable encounters for a level 4 party:
Two winged kobolds, one drake (This is from Hoard of the Dragon Queen)
Field HP = 114. Drake resistant to weapons = 214 hp effective, 3 targets.
For a party of 5 level 4s, you are running into 114hp with a 55% hit chance from kobolds doing 6 damage - so 6 a round. Pack Tactics makes that 90+% so we call that 10 around from the kobolds. A drake doing a 70% chance to hit on bite at at 10 + con save 75% chance to save 5 poison and disadvantage as well as the 55% chance to hit tail doing 6. So about 12DPR from that. Total of 22 DPR. AC 13 on kobolds means that first round theyre probably both dead but knocked out 10 damage. You can play them ranged and send in the drake for a round and add another 12 before players take them out. So you rotate into round 2 with ranged another 22. 44 damage in 2 rounds now. Kobolds drop we're back to doing 12DPR on a drake with 200hp. 5 level 4s at 45hp on the high which puts them at roughly 200 hp across 5 targets. Well we just took 1 almost out of the fight in 2 rounds with only 3 creatures. Add 2 more winged creatures and on is definitely down and on lost saves because they keep targeting them.
Players go - Barbarian, makes 1 +6 (55% chance to hit) with a greataxe - 1d12+4 (with 18 str). 10 damage reduced to 5, so 2.5DPR. Rogue goes, dual wielding short swords. +5 to attack (55%) for 1d6+3 (6) and sneak attack 2d6 (6) = 12/2 is 6 plus another 55% chance with 3 on 1d6 because no dual wield feat, reduced to 1 = 7/2 3 DPR. We'll just say that average is going to work for a ranger as well and then we have a cleric and a wizard. 2d6 force on cleric for spiritual weapon is 6 at 55% = 3, not reduced force damage and a wizard is probably going to use either catapult, mm, or a chill touch/toll the dead. So, now you are added 6 average.
Add all of that up and you get 2.5 + 3 + 3 + 3 + 6 = 18 DPR v the 12 of the drake, each at 200hp. 5 rounds doing average but rogues exist and you might get 32 total on a crit from a sneak attacking that changes life significantly. Having poisoned PCs attacking at disadvantage levels the DPR. We all know dice exist and there are clever things to do so that moves the 5 rounds of the averages only taking everyone to 0 and moves it to about 3 with some healing done. Lo and behold, a 3/4 round fight that was 3 creatures and still a challenge.
But, like I said, 10 goblins with 45% doing 5 damage so 2DPR is 20DPR. Lethal to a level 5 party.
Firstly, I never said that any party regardless of configuration should be able to beat it, so do not put words in my mouth. Secondly, I would not say such a thing, because I do not believe it to be true, in fact, I solidly believe that if a party is not right for an encounter, they should die, horribly and as much sympathy as I may have for them, that is perfectly fine and is the way of things. Thirdly, as nice as this lecture is, this is not an encounter that exists within the game and nor is it entirely relevant as not all encounters (and parties) are the same. Your "lethal" encounter for 1 party, may be an easy encounter for another group. Power Gamers exist and so do players who want to experience nearly no combat at all and are mostly there for socializing, a good GM is 1 who is able to accurately judge their players and build the right experience for them.
Larian is building a "GM" which reacts the same to every party, in a different medium. They cannot do nuanced gameplay and make goblins play like tactical geniuses for a power gamer, or be more fair to someone who is just there to socialize, so there needs to be some compromise. They cannot predict what the players will do and to be able to account for that, the size of a "fair" encounter, will probably vary greatly from your source book.
I don't know how you got killing 4+ of them a round?
a Fighter wielding a greatsword, which, after accounting for all modifiers, deals 2d6+5 damage excluding superiority die as well as the dipping mechanic. 12 HP goblins, on hit die 58.33% of the time. Against the 10 HP Goblins, 83.33% of the time. You can reasonably expect 4-5 enemies to die every single round in that example
Fighter gets 1 hit, so even if its for literally infinite damage, just one dies. Also ON HIT, hes hitting 60% of the time. 40% of that time on a 12hp goblin, it doesn't die. You have 3 other people who do less damage - barring a rogue - and they are *certainly* not killing one a round. So the concept of killing 4+ is right out, and youll be lucky to get 2 based on the math for hitting their AC and the average DPR. So, the math doesn't jive beyond what I said above regarding 10 with 2 dropping a round, I haven't seen anything that changes that. If you want to take a video of you starting a fight against 10 goblins and without cheesing the absolute shit out of it, kill 4 a round. The numbers you posted for Shatter and Channel divinity are, again, for that particular configuration and they still have a surprisingly good save rate against it so assume max 75% of the damage on the board and thats if they havent already taken your wizard out and the cleric is spending the entire fight healing people and casting aid.
While, yes, its possible, and I can do it, it also requires a fair amount of optimization and cheesing. Want to run two wizards because you are on and you like gale? Well, good luck!
The fighter has an 80% chance of hitting, without advantage. 96% chance if you do have advantage.
The game provides you with potions of speed and other consumables, which, if you use intelligently during challenging encounters, it makes that "4-5 dying per turn" not much of an achievement. Please, keep telling me how super optimized this party is. Its a Warlock, a Fighter, a Cleric and a Rogue and we all know what Shadowheart's statistics look like. If I was trying to be optimized, I would have ditched either her or the fighter for Gale. Why? Because Thunderwave and Magic Missile with the necklace do far more damage than a fighter or Shadowheart could hope to do, especially with all the vertical terrain elements available within the game. As much as you "disbelieve" 4-5 enemies die per turn, its happening.
The fact that your mindset is, "set the cleric to casting healing word or aid" already tells me why you struggle to see this happening, that is not how I am playing the game. You know what the cleric is doing? Nuking things with either a longbow or guiding bolt. The best damage mitigation is dead goblins.
So while having to be challenging for the hardest core min-maxed 4 PC MP game and also give something beatable to "WTF is a D&D lol?" player is difficult, the way you simplify it is reduce headcount. Bodies means DPR, flat out. If you are using utility spells, lockdowns, CC, it falls apart when you realize that holding 1/10 of the force in Hold Person was silly, its meaningless based on what their output as a collective is still. There are no real "impact" spells because the round length and crowded fields. So, what do you do? You make encounters like the ones that come as rollable encounters from the D&D modules that have a hell of a lot more playtesting behind them than this. 4 goblins with 2 pet drakes? Wow, crazy ass fight. Locking down a drake is a HUGE DEAL now. Casters and Druids, when they get introduced, are being neutered by the spam. So, what do you do? Give the dual wielding martial classes a bunch of buffs and push around the scatter a bit so they aren't getting attacked so much and roll until a side is dead.
Beginner DMs make the mistake of thinking an epic battle is an army coming after you and when its 45 minutes a round and everyone is bored to tears you learn to not do that anymore. Get some tough things in there, drop that headcount, and make spells and abilities matter because youre taking out a larger percentage of their round DPR by holding, sickening, weakening, or knocking down something while granting your martials some advantages and areas for them to take advantage of the chaos and chip damage youre creating so that 1 swing WILL kill something.
Fortunately for you, Larian has realized this and not every enemy is a wave of Goblins. For example, there are 2 Minotaurs in the underdark which are intended as a challenging fight, all on their lonesome. I can also give examples of fights like the Bulette, but I think you see my point. But, just because not every enemy needs to be a wave of Goblins, does not mean that a wave of Goblins does not need to exist.
The EA has 2 set piece battles, which are both potentially entirely optional. There is nothing forcing a player to do them and even if you feel compelled to complete 1 of the quest lines (either to defend the druid camp, or kill the druid camp), the 2nd set piece battle (killing the goblin camp) is still pretty much entirely optional. Set piece battles have their place, provided they are scarce. There are a sum total of 2 in the entire first chapter and neither of them are forced on you, I would say this fully meets the definition of "scarce."
The tactics being used right now are to reduce the effect of the HP and headcount on the field. They are not things you do in D&D. You think what ability would do the most interesting thing and tilt this. No one runs and finds a doorway or a ladder in D&D. That just isn't in the DNA. They saw a 3D world and decided to use that battle map as an excuse to litter the world with bodies. I think the most true-to-D&D encounters in the game currently are the madcaps, bulette (although its jump aoe damage is bullshit), and minotaurs.
Maybe its not what you do. Maybe to you, D&D is a theater of the mind and you are, "just here to see pretty explosions" but D&D has its roots in war gaming and if you look to those origins, things like using line of sight, taking advantage of height and forcing enemies through choke points would certainly be in the DNA of D&D. That is what D&D is to me, it is a glorified fantasy squad based tactics simulator and like it or not, I am a part of this game's audience just as much as you are.
Long fights always favor martial classes, pick a game with expendable resources, that is a true statement.
Provided casters are the ones with the limited resource, yes, this is the case and that is fine. In D&D this is the case. A class doesn't need to feel useful in every circumstances, I am perfectly fine with casters feeling useless every now and again and my preferred class in D&D, for your information, is a Sorcerer. Yes, a good GM will occasionally make me feel useless within the context of the game, because it adds a touch of realism to the setting.