They were only addressing the attack tables/ THAC0, etc invovling attacks, so I didn't bring up "accuracy" in terms of skill checks.
Yeah, I've never been happy with "bounded accuracy" in terms of skills.
Yeah, such an insane expert (certainly possibly at 17+ levels) has about a 75% chance to do the "Nearly Impossible". But for many players, a tier 4 super expert with advantage should be able to hit DC 30 with the much loved 65+% chance of success which runs rampant through 5E from early on.
And those same players prefer that a PC in heavy armor with DEX 8 and no stealth proficiency SHOULD fail at stealth most of the time. Level doesn't matter at that point. Now, if Peter Paladin wants to be better at stealth, proficiency goes a decent way to increasing the odds, especially at higher levels.
Peter Paladin only has about a 6% chance to beat DC 15 without proficiency. Give him a +2 proficiency bonus and it doubles to over 12% (still crappy...). Give him guidance and advantage, like Edgar Expert has, and now Peter Paladin has a 47% chance of beating DC 15. Make Peter Paladin tier 4 and he even has a 42.5% chance to beat DC 20.
So, Peter Paladin in heavy armor, dumped DEX, and no proficiency in stealth will suck at DEX (Stealth) checks, regardless of level, because no resource has ever been put into it. But that's pretty much how it should be, shouldn't it???
Except that he sucks at stealth checks
against level 1 enemies. He sucked at them even when the characters were first level. Now that they're 20th, he's effectively getting guaranteed failures--and has learned absolute bupkis if he were to go back and deal with weaker threats.
That was not true in 4e, and the universality of the half-level bonus is precisely why. When the party is going out and doing things that are within or just beyond their comfort zone, his Stealth will be not great. (I would know, I have played such a paladin, though his name was Seth.) But if at, say, level 15 (analogous to 5e level 10), he were to need to go sneaking through areas populated with the kinds of threats he'd faced at level 1?
He would be better at stealth than before. He would, in fact, have actually learned a thing or two. It wouldn't be enough to really make him all that
good at it, a total bonus of +7 at level one is okay but not great, meaning he'd have solid chances to sneak past such things. (This, I must admit, I have not seen, but that only because my 4e games have been curtailed more than once by DMs having IRL issues that pulled them away from TTRPGing for the foreseeable future.)
Well, we're dealing with extreme cases here. I didn't think bounded accuracy was really meant a super-focused expert from being really great at the stuff they are, well, supposed to be really great at. And of course the Paladin isn't getting worse, it is that his opponents are paying better attention.
What I thought bounded accuracy was supposed to do with skills is make reasonably hard DCs still a bit of a challenge later on?
The explicit aim was to reduce the size and amount of bonuses characters could receive, so that the numbers would be lower (and thus easier to do math with) and progression more tightly controlled. A PC getting more than +25 by level 18 in 5e is equivalent to a 4e character having +50 by level 27, something even ultra-experts hyper-specialized in one and only one skill would struggle to achieve. A much more typical skill bonus for many characters would be...well, about equal to their character level (half from half-level bonus, the other half from training, ability score, items, etc.), so around 27. Meaning, for a 5e character, roughly Expertise with a +0 modifier or Proficiency with a +5 modifier, and nothing more.
Hence, despite explicitly trying to curtail extreme bonuses and keep numbers within a neat, tidy, narrow range...bounded accuracy has actually not done all that much to bound accuracy. Instead, what it bound was off-label stuff.
That stuff barely moves, and may even stay essentially flat across a character's career. Your weaknesses never get less weak, unless you radically refocus your character to address them, paying a steep price to do so. Meanwhile, your enemies get stronger; hence, instead of a treadmill, we have people straight-up losing a Red Queen's race.
Again, because in AD&D it sucks to begin with.
You probably wouldn't be surprised to know that I feel that
5e bonuses and effects suck to begin with (seriously, "competence" is now apparently succeeding about 15 percentage points more often!) So if AD&D is supposed to suck when treating 5e as one's baseline...
5e's design goal is to be popular, with as little objectionable content (to anyone) as possible. Unfortunately, being straightforward about that goal would have in all likelihood damaged its success. They have, I believe, succeeded admirably in that goal. How we feel about that is irrelevant.
Popularity isn't a design goal any more than sales are a design goal.
You cannot point to a part of the design and say, "This is what causes popularity." Popularity necessarily is much later down the chain of cause and effect than the goals phase of game design. Designers certainly
hope that their designs will be popular. They will, in all likelihood, ask players about how they liked various things. But popularity itself
is not and cannot be a design goal.
