Actually, it's 8-15, since no stat before racial adjustments can be above 15. But, it's still a bell curve.
Remember, in 5e, it is not possible, using point buy, to have a stat over 15. 16+ is superhuman for a 1st level character. You're claiming that it's realistic for nearly every single PC to be superhuman. It's pretty unlikely you'll get no rolls above a 15 with your rolling system.
And that's somehow more realistic?
That is 100% incorrect as I pointed out. The fact that you choose to use one of the systems that doesn't allow you to have a 16+ does not alter the meaning of a 16+. In the game, 18 is the pinnacle of normal human capability, not superhuman. That starts at 19. That hasn't changed since the game was first published.
I have no idea what type of curve is created by the relationship of the stats as a whole (rolled or otherwise). However, when referring to a bell curve, we're talking about the probability of a given number when you roll the dice. 3d6 produces a bell curve. In the standard array, I don't think there is any curve. The number is what it is. In point buy, it would be linear. You have several options, and each number has an equal chance. Once you've chosen the first ability, the next ability has a narrower range of numbers available, but the possibility of any of those numbers being selected is still equal, it's still a linear progression. The
cost is different for higher numbers, but that doesn't alter the fact that any of the numbers within the range are equally possible.
Minimum 8th level character, so, we're talking a tiny, tiny fraction of the game world. Also irrelevant for a conversation about chargen since we generally don't start at 8th level. But, as I said earlier, if you let me use standard array and then hand me two free ASI's, I'd be fine.
But no, it's not a tiny fraction of the game world. RAW, it's achievable by every single person in the world within a matter of days. If two 18s is unacceptable for a 1st level person at age 20, then it should be unacceptable for an 8th level person at age 20 and 15 days. The speed at which people can gain ASIs is very, very fast in 5e in relation to the game world. I certainly think that it's much more believable to have two 18s by 1st level that has been developed over 20 years of life (or more in the case of non-humans), than it is to go from 16 to 20 in less than a month. As you said, we're talking D&D, and them's the rules.
When discussing the "realism" of multiple people having multiple high scores, it's a question of whether each character created can achieve it. The 8th level limit applies equally across the board, regardless of the method of generation. By standard array, every single character starts with a 15 and a 14, and they gain at least a +1 for two stats due to racial bonuses. Thus, it's possible (even likely) for every single standard array character to have two 18s by 8th level, as early as 6th for a non-human fighter.
With dice rolling, it's possible (and even probable) that some will not.
To put it a different way, there is a percentage of rolled characters (20%? Somebody better at math will have to figure that out) that will not have a 14 and a 15 (or higher) as their two highest scores. Some will have no scores above 14, some will have no scores above 13. This is the result of the bell curve. It is impossible for a character with no score over 13 to have the same result as the standard array, they can't gain two scores of 18 by their third ASI.
With point buy, that's theoretically the same, but when have you not seen a point buy that people selected the maximum allowable for their two most important stats.
Hang on. I never said ANYTHING about it being a "horrible preference". If that's what you want to do, knock yourself out. Just don't pretend that it's something it's not. If you want to completely rewrite the game and then try to use your game as a talking point, it's pretty difficult since I don't play your game and you don't play mine. Frankly, I don't want to talk about your game, I don't care. I want to talk about D&D.
But, in any case it's not a horrible preference. It's simply that trying to rationalize it as somehow "realistic" is what I'm arguing against. It's not realistic. There's a reason that virtually no sim-based games use random stat generation. If it really was realistic, don't you think games like GURPS would be using it?
No, you didn't, and I wasn't implying you did. Just joking around, as I (and others) have the tendency to get a little too serious sometimes...It's not directed to you, or really anybody, just an acknowledgement that we have different priorities.
However, being "realistic" is a number of different things. To somebody like me, it's not having full control of the stats you were born with. That something like GURPS took a different direction means literally nothing in that regard. The process itself, which I have pointed out, is irrelevant. It's the probabilities/results that the process produces that would matter if, in fact, you were trying to be more "realistic."
Having said that, I do think that to produce "realistic" results, it would require random determination of some sort, and an ability to be outside the statistical norm. Everybody in the entire world having the same base number is most definitely not realistic. But "realism" isn't my core goal. I laid out our goals pretty thoroughly, and while one of those goals is to set in-world baselines and help with consistency (which is different from everybody having the same base number, and cannot be reached via a standard array), we developed a system to meet those goals. While this does bear a relationship to "realism" it's not the same.