Please be advised that if you use the connect app, it doesn't always correctly parse links to lemmy posts. If they're not working for you, you can follow the whole plotline on my site. (These comics are in reverse chronological order, so start at the end and work backwards.)
I guess if you throw an egregious amount of magical power at a problem it DOES solve it!
So... when you design a system to use "bounded accuracy" it can fall apart completely if you include a large number of stacking buffs.
WotC's original idea was that the way you'd "buff" someone for a check would be to give advantage on it, and the options for "flat bonuses" (or die-roll bonuses) would be extremely limited... but in hindsight, if you have the right classes and a lot of prepwork, you can force a ~30 roll on nearly any check you care about.
The biggest culprit is, as everyone knows, Pass Without Trace, but don's sleep on the Artificer ability Flash of Genius - it's incredibly impactful.
Honestly even without the dice buffs, it's pretty easy to break - Expertise is another big culprit. At T4, with regular proficiency, you can get up to a +11 modifier without magic items or spells, meaning your average is going to hover in the ~20s, ~25s if you can get advantage, but the range is still bounded from 12-31, only a 10% chance of achieving the 'nearly impossible' for a character representing the pinnacle of ability and adventuring prowess. Makes sense.
With expertise, the modifier goes up to +17, average roll 27, 32 with advantage, range of 18-37, which gives a 40% chance of achieving 'nearly impossible' DCs. Makes less sense. If you happen to be playing a rogue with expertise in Thieves tools and sleight of hand, and wearing gloves of thievery, your skill floor is 32(!!!) - You literally cannot fail to pick a lock or a pocket unless the DM creates a scenario that breaks bounded accuracy. And then, of course, you can throw BI, guidance, etc. etc. on top of that...
If I had to homebrew a quick fix, I'd suggest something along the lines of "When making an attack, check, or saving throw while under the effects of a spell or feature that allows you to add additional dice to the d20 roll, such as Bardic Inspiration, you may only gain the benefit from one feature on a single roll. If multiple features would add dice, you may choose which feature to add the dice from when you roll."
Totally ranting at this point - I think 5e's got good bones but there's a lot of problems inherent with simplifying down to a single scaling number. Saving throws swing the opposite way - depending on the creature you're facing, it can be literally impossible for a PC and/or NPC to make a saving throw against DC 20 and up if you don't have proficiency, and by design you won't have proficiency approximately 2/3 of the time (this is also why named monsters get legendary saves in 5e.) Not saying we need to go back to 3.5/PF with like a dozen different sources for bonuses, definitely not that, but proficiency needs to be less binary for a more balanced and playable game, especially at higher tiers.
If I had to homebrew a quick fix, I’d suggest something along the lines of “When making an attack, check, or saving throw while under the effects of a spell or feature that allows you to add additional dice to the d20 roll, such as Bardic Inspiration, you may only gain the benefit from one feature on a single roll. If multiple features would add dice, you may choose which feature to add the dice from when you roll.”
You don't need to because bounded accuracy doesn't need to exist, often it's actively detrimental. See the common houserule that some checks can't be attempted without proficiency.
I understand WOTC's idea to simplify by making Advantage the primary way to modify rolls, but it runs into a wall pretty quickly: only one effect can help at a time. Therefore you've still got things like Guidance and Bardic Inspiration providing numerical bonuses and the issue WOTC was trying to solve is only mitigated. And occasionally you get things like the above with an easier 30.
On the other hand, our level 12 Bard has +13 to most of his CHA-based skills (+5CHA, 4 Prof, expertise) so rolling 20-25 with just a flat D20 is more common than not. Which means unless the people he's persuading are heroically resistant to charm, he's probably getting his way within reason. Not having level-based scaling really amplifies players into super-human.
I feel what it's really "missing" is the bonus "typing" from previous editions - where you couldn't "stack" bonuses of the same type - so you might have access to a lot of buff spells that give you a flat bonus, but you can't stack them if they're the same general kind of thing...
The reason they moved away from typing all the bonuses (and keywording everything) was "simplicity" - and that does make some sense, but "bounded accuracy" has a built in assumption that you can't stack lots of bonuses...
There's some ways you can house-rule to try and adjust this... but they're pretty significant system changes, so... is it worth doing, or is it better to focus on scenario design to avoid this mattering so much...
I feel it's WotC's intention not to keep stacking bonusses.
All the plusses and minusses in earlier editions could really turn a fantasy game into a mathematical slug-fest, and without the stacking bonus that doesn't happen all that much.
But it does fall flat rather quickly for me, as well.
Well, by "fall apart" I mean that the purpose of a dice system is to introduce uncertainty. If you remove the uncertainty, then there's not much point rolling dice :)
The game is a mixt between Monopoly and Chess. Where one is almost entirely luck based and one is almost luck less. There is luck and ways to orient it.
Since its an hybrid system, it makes sense for players to invest so many resources to make a check a 100% guarantee.
Well, 95 % at my table since I like to crit fail or crit success checks as long as its reasonnable with the player's consent.
Oh yes... This kind of thing is pretty okay when the scenario is one that DnD 5e is designed for - a day of adventuring with 10 encounters, or a dungeon crawl, or the like. At that point, burning a huge amount of resources on a single check has a real, significant cost to force that success. If you provide these kinds of challenges mixed into the expected scenario-design, then it works well. It means you can say "this matters enough to me to hamstring my later power"
The problem mostly comes in when you're in situations like this - where you don't expect to need all your resources in the day, and burning through them let you force a success on the only check you care about.
Recently no joke, we played a day in a game. We knew an attack was coming, and do you think me as a bard kept my resources for it ? Fuck no. I gave each of them to our mage that tried to catch the mega fish in the fishing competition that morning.
When people say this, they usually mean the story falls apart. Without struggle there’s no story. No one wants to hear the story of the white boy who lived without hardship with rich parents who made B’s and went to Yale and became a lawyer.
Bounded accuracy is probably the single worst design I've ever heard about in ttrpgs, and I've looked at some of the math you gotta do in fatal. The things actually designed for it make you feel pathetic, the things that aren't make the game boring.