And that is very nice and all and sure, for a table it makes sense because it would take forever to roll the HP of every single goblin and there is no real reason to, but the rules don't say, "Goblins have 7 hp" it says, "Goblins have 2d6" hp. A Goblin can have anywhere within that HP range of 2 to 12 and it will be per the official rules. You know the great thing about computers though, they could roll the HP for every single monster in the game in under a second and it would not slow down gameplay at all. I am personally fine with Goblins that have 15 AC and 12 HP, as per the rules if Larian decides they only want to use maximum HP rolls for monsters. Sounds good?
I think youre missing the point. As a DM/game designer you are sitting there saying I have 4 level 3 PCs with 24 hp and 15 AC avg.
So there is 100hp on the field in 25 hp chunks. At 1d20+4 to attack from goblins that puts goblins at a 50/50 to hit the PC.
Shortbow 1d6+2 = 5 damage on average. So, with focus focus on a single PC, which the AI does, 10 creatures with a 50/50 means 1 pc dead a round.
Now, why would you do 10 goblins? thats a lot for 4 people, they have no way of killing 10 goblins with even 7hp because they have ac15 and youre 50/50 as well and fine you have 8 damage per round or 10 even per PC. At 50/50 you still are only talking 2 goblins a round. Well, next round, you heal up someone to 5. Well you get that same person down again with 2 goblins 50/50 on them. You got 6 left, and someone else is 3/4 down with the 50/50. Another 2 goblins go down. Down to 6 and they take the 2 that went down again down. PCs are out of spells nows, 2 are down or low single digits and you got a few left. Thats if no natural 20s happened, the goblins werent rolling well & or the PCs werent rolling poorly. Realistically, you have a very strong possibility of a TPK with a goblin group that size. Add a hobgoblin with martial advantage and you need half the amount of goblins to get a TPK.
Ok, so thats 10 goblins with 7hp. You run that scenario and you look at the averages and that what happens if they roll well. So you drop the number down because in ONE ROUND regardless of HP, the damage output is enough to easily drop a PC. So you say they can also roll poorly. Ok so not 10, not 5, give a buffer with 7 or throw 5 in with a hobgoblin. Lets keep 3 back in range and 2 goblins to walk in with the hobgoblin.
Ok, now I am a computer running numbers. What algorithm are you going to run to balance HP rolled versus the damage output? The fact is that first round is the important part and HP isn't calculated into it because the PCs have limited attacks and even with 1hp, 1ac, 10 goblins is enough to down a PC, easily.
So when do you roll hp? After you decide the damage output per goblin? Random the HP based on the fact that 7 or 5 and a hobgoblin is a good fight and then the variable is how much it is to take the damage output down? Well that changes it. You add another round to a 7hp goblin with 14hp and you just doubled the number of goblins on the field. So do you delete 3 goblins from your 7 goblin encounter because you tripled the hp on the field and they'll be getting as many turns as 7 goblins would?
The argument doesn't really hold up for that method. The things you are considering stay the same and youre controlling the power swing per round and determining the number of rounds you can survive. A rule of thumb in D&D is creatures die within 3 rounds. If you want it to be tough on players, we start talking unique enemies - orc bosses or something - you might push to 4 or 5, but thats still rare. Usually, the 3 round rule just becomes a lot scarier because the power swing can be so great from the enemy youre fighting. A natural 20 on the bossman can 1 shot a PC, the whole fight just changed.