Bounded Accuracy and such?

[2D8HP]

I've been playing and enjoying 5e WD&D and reading posts on this Forum recently, before that I played other RPG's until 1992, starting with TSR D&D in 1978.

Most posts in this Forum detail how 5e is different from 3e/3.5/4e WD&D, and Pathfinder, not the TD&D that I have dim memories of, and this quote fascinated me:


Here are some 'rules' for 5e

1. Bounded accuracy
It means that any roll has a chance to fail or succeed.....
:confused:

I've seen much mention of "Bounded Accuracy", but never really bothered with what it meant, but if the quote was accurate, it simply means that one only rolls if there's a chance of both success and failure.

Why would there ever be rolls when that wasn't the case?

Were there mechanics for degrees of automatic successes and failures in previous WD&D Editions?

[tyckspoon]

I've been playing and enjoying 5e WD&D and reading posts on this Forum recently, before that I played other RPG's until 1992, starting with TSR D&D in 1978.

Most posts in this Forum detail how 5e is different from 3e/3.5/4e WD&D, and Pathfinder, not the TD&D that I have dim memories of, and this quote fascinated me:

.
:confused:

I've seen much mention of "Bounded Accuracy", but never really bothered with what it meant, but if the quote was accurate, it simply means that one only rolls if there's a chance of both success and failure.

Why would there ever be rolls when that wasn't the case?

Were there mechanics for degrees of automatic successes and failures in previous WD&D Editions?

Bounded Accuracy is a design principle that means, roughly, that the bonuses and stats achievable by characters should never be so high that the randomizer (the d20, in D&D's case) stops mattering. In 3rd Edition and its relatives it was possible and in some cases pretty easy to get skill/attack/save bonuses to the point where the actual d20 was pretty irrelevant. In 5th Ed the range of possible stat numbers is much tighter and more controlled, so that a character should almost always be able to have an effect on another creature or character if the die roll goes there way.

[Ninja_Prawn]

I think what people mean when they say that is that in 3.PF, you can have +X bonuses to rolls that are so large that the number on the die basically doesn't matter, and ACs/DCs so high you can only hit them if you have massive bonuses. In 5e, a goblin has +4 to hit and an adult black dragon has an AC of 19. That means a goblin can hit a dragon on a 15 - that's 30% of the time! In PF, it's still +4 to hit, but the dragon's AC is 28. The goblin can't hit it short of a nat 20.

The flip side is that high-CR monsters have shedloads of HP in 5e.

[Unoriginal]

Another exemple: in 5e, a level 20 Fighter with a maxed out attack stat, attacking in melee with a legendary +3 sword, would have +14 to hit.

Which mean they have 1 chance out of 4 to miss a guy in plate armor with a shield.

[lunaticfringe]

There is a Cap on difficulty. (30)

The game is not built on the idea of Level increasing Accuracy by great degrees. You should always be able to reasonably hit things.

An 1/2 CR Orc has 15 AC, a CR 13 Storm Giant has 16.

[Knaight]

I've seen much mention of "Bounded Accuracy", but never really bothered with what it meant, but if the quote was accurate, it simply means that one only rolls if there's a chance of both success and failure.

Why would there ever be rolls when that wasn't the case?

It's more than that. While rolling in times of uncertainty only is a general principle across RPGs, it's often character specific. Gilgamesh gets to roll to see if he can shoot down a thousand incoming arrows in flight, automatically succeeds at shooting down one incoming arrow in flight, and automatically succeeds at pulling a bow. Robin Hood automatically fails at shooting down a thousand incoming arrows in flight, rolls to see if he can shoot down an incoming arrow, and automatically succeeds at pulling a bow. Some paper pushing civilian automatically fails at shooting down a thousand arrows in flight, automatically fails in shooting down an arrow in flight, and rolls to see if they can draw back the bow in the first place.

This represents at least three distinct tiers* of rolling, where you automatically fail at a tier above, automatically succeed at a tier below, and roll when in tier. There may be tiers or fractions of tiers both inside and outside of the three example characters.

3.5 absolutely had that. Having a +20 bonus relative to another character puts you a whole tier ahead of them at something. Having a +40 bonus relative to another character puts you two tiers ahead of them. Getting that +20 comes near automatically with a mid level character, an attribute bonus of +6 and a skill of +14 can be reached at level 11 without putting any real effort into it. Getting that +40 generally requires at least a little effort, but isn't actually hard. Even getting +60 or +80 at high level isn't that difficult.

4e clamped down on that a little, but a max level 4e character gets .75 tiers above automatically in every skill without trying compared to a 1st level character. Getting a whole tier above is entirely doable, getting four tiers above is as far as I'm aware completely blocked.

Bounded accuracy as a philosophy is essentially that at least in combat everyone stays in the same tier as the baseline character. They don't necessarily have to stay in the same tier as each other, and the philosophy can have exemptions, but things like gaining +20 to attack or +20 to AC or saves above your starting point isn't just assumed any more. A character terrible at doing something might have a -2, even the best get only +17, and even then that's mostly cut off from combat, where +13 is more common, which is more like +9 given that +4 is a pretty routine starting value. Heavy optimization can push these numbers, but they're what the game was designed around. Getting a few magic items is also expected, so call the upper range +17 (skills), and +11 (combat).

As for the early D&D comparison, there weren't skills. Still, the combat figures had a range of +16 in 1e, and +19 in 2e, without magic items. Throw those in, and you get +20 and +24, as fairly routine numbers that reflect changes in item availability. AC patterns also reflect this in both games, as well as 5e. 5e has dialed it down, and more than just dialing it back down. Of course, HP has gone way up compared to earlier editions.

*All the good terms for this are already in use, so I'll just note that this is a different definition than tier usually has around here. It seemed like a better option than class or level.

[PhoenixPhyre]

To be specific and pedantic,

Bounded Accuracy (BA) is the name given to the lack of assumption that AC, save bonuses, attack bonuses, and save DCs scale (strongly) with level/CR. A related lack of assumption (usually subsumed under bounded accuracy) is that ability check DCs strongly scale with level either.

Note the words "lack of assumption." This is not a positive assumption--there is no positive statement that defenses don't scale with level (they do), but that the math should work even if they didn't. As a result, BA is not broken when saves/etc scale with level. If the designers did their math right, such scaling things (like expertise) are bonuses but don't break the math. Because the math is only bounded on one side. The lower side.

What comes of BA?
+ No assumption of needing +X items. Even without them (both +ATK and +AC), you still have a reasonable chance of hitting/avoiding a hit (as in, you hit on lower than a 20 and are missed on rolls greater than 1). Having those items is a bonus, but not a requirement (like it was in 3e/4e, where lacking that progression was fatal).
+ Monsters stay relevant longer. A no-magic plate wearer hits a static AC of 21 (barring some special things like defensive duelist). Even a +3 ATK (most CR 1/8 creatures) still hit on a 17+.
+ Scaling happens in damage and HP. A level 1 Barbarian has ~14HP, a level 20 one has ~285. Similarly, an average monster faced (in packs) by a level 1 party does ~7 DPR. An average monster faced (in packs) by a level 20 party does ~75 DPR.
+ Large(r) groups of low CR monsters can replace single high CR monsters--this avoids the weirdness of killing 6-8 ancient dragons/balor in a day. Instead you might face a small army of orcs led by a beefy orc (attacking in waves), and finish off with a major demon and some henchmen. To me, that keeps the world saner without postulating the existence of large amounts of epic-level threats.

[2D8HP]

To be specific and.. ..
I didn't understand all of what you wrote @PhoenixPhyre, but what I did understand mostly seem like good goals to me.

[PhoenixPhyre]

.
I didn't understand all of what you wrote @PhoenixPhyre, but what I did understand mostly seem like good goals to me.

Sorry. I tend to be a bit verbose. But I agree about it being a good thing. Bounded accuracy is one of the things I like most about this edition.

[2D8HP]

Sorry. I tend to be a bit verbose. But I agree about it being a good thing. Bounded accuracy is one of the things I like most about this edition..
I'm not sure it my lack of literacy that makes my understanding a little hard so much as my memories of Oe D&D, and 1e AD&D being decades old, and my being relatively ignorant of (I only glanced at the rules), and having never actually played 2e AD&D, or any WD&D before 5e.

Something that initially confused my was that people would post along the lines of "compared to previous editions 5e is...", in ways that seemed the opposite of my perceptions, but I've come to realize that folks mostly meant the in-between D&D that I skipped.

I've read some stuff on how 5e is different from 3e/Pathfinder and 4e, but my ignorance of those games hinders my understanding.

FWLIW, I like 1st level 5e a lot, it feels like 3rd level in the D&D I used to play, but going up in levels in 5e has diminishing returns in fun for me (the highest from 1st has been 5th).

In games that have started at higher levels the highest I've played at has been 11th which I fond a bit confusing and dull.

If I'm understanding the term "Bounded Accuracy" correctly, than I think I would like even more of it, or just have level advancement by slower than in the 5e games I've played.

[DivisibleByZero]

http://bluishcertainty.blogspot.com/2016/06/bounded-accuracy.html?m=1

Read that.
It is word for word what Rodney Thompson wrote in the Legends and Lore article explaining what BA is before it was lost from Wizard's site, as everything always eventually is.

[2D8HP]

http://bluishcertainty.blogspot.com/2016/06/bounded-accuracy.html?m=1

Read that....
Thanks!

:smile:

[Ninja_Prawn]

before it was lost from Wizard's site, as everything always eventually is.

https://sa1.narvii.com/6562/82ee7f219b60feefefe7bebdc17625ed8cbc6045_00.gif

[Eric Diaz]

One thing I haven't seem mentioned here is that some numbers tend to get a "higher floor" to avoid a "higher ceiling"... Reliable Talent, Indomitable Might and even the Great Weapon Fighting style come to mind.

There is probably a better explanation than mine out there.

In any case, I like Bounded Accuracy, but I think they went TOO FAR in 5e. A kobold beating an Adult Red Dragon AC more than, say, 10% of the time is too much for my tastes... and a 20th level fighter with STR 20 and proficiency in athletics should have NO chance (well, maybe less than 1%) of being beaten by a 1st level wizard with STR 10...

I know, I know, you wouldn't ask for a roll so the system is okay as it is.