Bounded Accuracy: Where are the bounds?

Bounded Accuracy is a great concept, but if we are going to use it as a fundamental concept for game design, shouldn't it be clearly defined? I know 'Bounded Accuracy' is really just code for 'flatter math', but I believe placing actual boundaries on the system will create the internal consistency that the game requires. 

What do you think?

 

"Make things as simple as possible, but no simpler." - Albert Einstein

I know 'Bounded Accuracy' is really just code for 'flatter math'

No, it isn't.

Go read  this.

D&D Next = D&D: Quantum Edition
Ok. But do you believe there should be actual limits in the game system? If so, what should they be?

"Make things as simple as possible, but no simpler." - Albert Einstein

No, they serve no function.  It's all arbitrary math at that point.

Things have to do something, or they're not worth specifying.  Specifying something that doesn't actually do anything helps nothing, and probably causes problems you wouldn't otherwise have had.
D&D Next = D&D: Quantum Edition
I think having firm DC limits would be helpful. Something like DC 1- 20, where DC 1 - 10 for mortal/natural actions, DC 11 - 20 for fantastic/supernatural actions. People can easily grasp a scale of 1 - 10.

"Make things as simple as possible, but no simpler." - Albert Einstein

You're missing my point.

You don't specify a limit, you just...set the DCs.  The limits set themselves.
D&D Next = D&D: Quantum Edition
My problem with bounded accuracy is not in combat, is about DCs outside of combat (wish the biggest grasp is that we are rolling d20s), i already have a big discussion with Mand12 about it that went nowhere
That's because you're fretting over insignificant details like the presented DCs, before the skill system is even close to being finished.
D&D Next = D&D: Quantum Edition
Mand12, ok, I understand what you're saying. I agree. They do. But as it is, the DC's have nothing to do with the actual number (e.g. Moderate - DC 13) and they feel arbitrary. I'd just like to see something a little more intuitive (like my example of 1 - 10 above).

"Make things as simple as possible, but no simpler." - Albert Einstein

Mexrage, I don't understand your concern...  ?
 

"Make things as simple as possible, but no simpler." - Albert Einstein

Mand12, ok, I understand what you're saying. I agree. They do. But as it is, the DC's have nothing to do with the actual number (e.g. Moderate - DC 13) and they feel arbitrary. I'd just like to see something a little more intuitive (like my example of 1 - 10 above).


They are arbitrary.  You can't make them not arbitrary.  Even your example of something you find intuitive is abitrary, and to other people it may not feel intuitive.  There's no solution, no possible way to make it not feel arbitrary to everyone.
D&D Next = D&D: Quantum Edition
Mand12, ok, I understand what you're saying. I agree. They do. But as it is, the DC's have nothing to do with the actual number (e.g. Moderate - DC 13) and they feel arbitrary. I'd just like to see something a little more intuitive (like my example of 1 - 10 above).


When they do finally massage the skill system, I DO hope they set some sort of guidelines for what the DCs mean. They have some now, so I assume that would continue.
As a DM, it's easier for me to make DCs for actions I never expected with a rough, "that would be HARD" measure.

Unless I misunderstand the thread.
That said, these things probably get written after the system is finished and working.
A few guidelines for using the internet: 1. Mentally add "In my opinion" to the end of basically anything someone else says. Of course it's their opinion, they don't need to let you know. You're pretty smart. 2. Assume everyone means everything in the best manner they could mean it. Save yourself some stress and give people the benefit of the doubt. We'll all be happier if we type less emoticons. 3. Don't try to read people's minds. Sometimes people mean exactly what they say. You probably don't know them any better than they know themselves. 4. Let grammar slide. If you understood what they meant, you're good. It's better for your health. 5. Breath. It's just a dumb game.
I'd call it Fixed Background Accuracy, but that probably wouldn't be any clearer.

Under this system, you can see a plate-armored foe and know that it's going to be around AC18-20.  That, along with similar examples, helps give a sense of consistency to the world. Flatter math is part of what makes the system work, not the system itself.
Mexrage, I don't understand your concern...  ?
 



My problem when we are rolling d20...where every mod or bonus is 5% less chance to fail, having +6 compared at something at level 18 to +3 at level 1...is only 15% less chance to fail...a barbarian or fighter at level 18 will struggle almost as much as a level 1 with the same task that involve his strength check, a contest, etc... And what kills me is that the designer that exposed the bounded accuracy said that this was made, because on 4e DCs scale with level (wish is a complete lie and proof that he don't know crap about 4th edition) and that he thought that a fighter at level 1 struggle as much as a fighter at level 20 at breaching a wooden door, and because of that, they decided to drop the half-level bonuses (wish was the main source of overall bonuses on 4e)
Under this system, you can see a plate-armored foe and know that it's going to be around AC18-20.


That is, indeed, the point.  I'm going to steal your example.
D&D Next = D&D: Quantum Edition
Flatter math is part of what makes the system work, not the system itself.

Exactly. Regardless of what the DC ends up being, or the bonuses that are set to as a counter of said DC, when the group sees the iron-banded and barred door, they know they are up against a real obstacle. I don't care what level they are.

Mexrage, I don't understand your concern...  ?
 



My problem when we are rolling d20...where every mod or bonus is 5% less chance to fail, having +6 compared at something at level 18 to +3 at level 1...is only 15% less chance to fail...a barbarian or fighter at level 18 will struggle almost as much as a level 1 with the same task that involve his strength check, a contest, etc... And what kills me is that the designer that exposed the bounded accuracy said that this was made, because on 4e DCs scale with level (wish is a complete lie and proof that he don't know crap about 4th edition) and that he thought that a fighter at level 1 struggle as much as a fighter at level 20 at breaching a wooden door, and because of that, they decided to drop the half-level bonuses (wish was the main source of overall bonuses on 4e)



well, my understanding of level is that it represents experience and with experience can you more easily perform the task you could when you were a novice and you now have the skill knowledge to have a chance for succeeding the skill situationas there would have been a sure failure when you were a novice

in this example
It should be easier for the level 20 guy with an appropriate skill to get the door up compared to the level 1 guy since
with experience do he know what muscles to use when making the attempt and he know what part of the door he should attempt to affect in order to maximize his success chance and he even know if it is best with a more gradual pressure or a more burst like pressure and so forth
Bounded Accuracy is a great concept, but if we are going to use it as a fundamental concept for game design, shouldn't it be clearly defined? I know 'Bounded Accuracy' is really just code for 'flatter math', but I believe placing actual boundaries on the system will create the internal consistency that the game requires.

It would be nice to know what WotC's target numbers are, but I really doubt they will publish them. There where a lot of people who wanted to know what the target numbers for AC, attack rolls and damage numbers where for 4e and WotC never made them public. I'm not sure exactly why, but I guess it some combination of not wanting to fix hard limits into the game and not wanting it to be so obvious when they push past the limits. Bounded Accuracy isn't an absolutely fixed limit after all, if they produce a supplement for divine characters at some point, it will probably push past the limits for mortal characters in several ways.

....

We'll know exactly what WotC's target numbers are, because they'll say things like "The DC to bust a barred wooden door is 17 (Hard)."



Oh, and they did make those numbers public for 4e, (last page of the DMG rules update document) so you should check your research before posting wrong things.
D&D Next = D&D: Quantum Edition
I'm ok with Bounded Accuracy so far as Combat goes. Personally, I like lower numbers and keeping them relative across the board. Should they increase? Sure, but not to un-godly lengths JUST to make the numbers higher. For one, it completely eliminates lower-level monsters as even a threat and thus, sorta breaks my immersion. Goblins should be a deadly threat at really low levels, should give adventure's pause at moderate levels, and an annoying nuisance at high levels. They should never be totally ignored because they're numbers are super low and can be completely wiped with 1 spell. For another, it seems that the numbers increase to compete with monsters defenses and attacks, and they only increase to compete with adventurers as they level. Just seems like a treadmill that goes nowhere except to show math bloat. I'd rather a Adult Black Dragon have an AC of 19, with 1st level adventurers able to hit it 45-50% of the time but give him resistances and HP to mitigate just how powerful he is.
Mand12, I disagree, but that the fact that we disagree sort of proves your point. 

So. Maybe the solution isn't in a bounded scale. Maybe the solution is in creating an internal consistency of numbers in D&D. if a '5' actually meant something, it might feel more 'real'.

As it is, Levels scale from 1-30, stats go from 8-20, DC's go from 7-25, Stats go from 8-20 for PC's, 3-30 for NPC's and Monsters, Modifiers go from (-5 to +5), and damage goes from 1 - ?...

I think we can do better. 

"Make things as simple as possible, but no simpler." - Albert Einstein

....

We'll know exactly what WotC's target numbers are, because they'll say things like "The DC to bust a barred wooden door is 17 (Hard)."



Oh, and they did make those numbers public for 4e, (last page of the DMG rules update document) so you should check your research before posting wrong things.



yeah...they changed that during essentials, we actually had static DC tables for alot of things, infact my book of rules compedium still have the static tables (thought that was done by the Rob "Awesome" Heinsoo)...thanks Mearls...we all hate you for breaking 4e

I use that table, but i don't use the dificulty based on the party level...i use it for the level of the challenge or the obstacle...wooden door is a hard dc of level 1 even to a level 30 party
....

We'll know exactly what WotC's target numbers are, because they'll say things like "The DC to bust a barred wooden door is 17 (Hard)."



Oh, and they did make those numbers public for 4e, (last page of the DMG rules update document) so you should check your research before posting wrong things.

One of the players guide book also had the target numbers for PC defense and attack
yeah...they changed that during essentials, we actually had static DC tables for alot of things, infact my book of rules compedium still have the static tables (thought that was done by the Rob "Awesome" Heinsoo)...thanks Mearls...we all hate you for breaking 4e

I use that table, but i don't use the dificulty based on the party level...i use it for the level of the challenge or the obstacle...wooden door is a hard dc of level 1 even to a level 30 party

Actually the tables were there from the very first DMG. They only later updated them because they found out that their own target numbers were wrong
So. Maybe the solution isn't in a bounded scale. Maybe the solution is in creating an internal consistency of numbers in D&D. if a '5' actually meant something, it might feel more 'real'.


That's the whole point of Bounded Accuracy.  A DC 15 Strength check to break something is a DC 15 Strength check to break something, no matter what level the person doing the breaking is.  The DC is tied directly to the objective nature of the thing in question.  That's what Bounded Accuracy does.

In fact, a stated goal of the system is for the DM to just know what DC to make a thing, as a result of this internal consistency of numbers.  "easy" "medium" and "hard" are no longer moving targets.

Yes, PCs will improve over time.  What they might consider easy relative to their own capability may change.  But the objective label won't.  The fact that they can, actually, hit those "Hard" DCs easily is the real character advancement, not just keeping pace with the treadmill.
D&D Next = D&D: Quantum Edition
I understand and agree. Do you think they've succeeded?

I would contend that they haven't. And I'll quote the 'DM guidelines' as support:


"Here’s another secret: You don’t actually have to set the DC before the player rolls the check.


Decide whether the character succeeds based on the check result. You’ll probably find that your gut feeling (and the player’s) squares pretty well with the set DCs presented here. A number below 10 is never going to make it. A number in the low teens is good enough for an easy task. A number in the middle teens will succeed at a moderate task. And when a player rolls a 16 or better, there’s usually little question that the character succeeds.

Your players will never know."

 


(also, if it is against the rules for me to post this, I will remove it.)

"Make things as simple as possible, but no simpler." - Albert Einstein

I think we're getting somewhere with the current DC's, but there's actually only 7 Difficulty Classes: Trivial, Easy, Moderate, Hard, Very Hard, Formidable, and Nearly Impossible.

Listing the DC's as 7 - 25 is misleading. 

"Make things as simple as possible, but no simpler." - Albert Einstein

No it isn't misleading, as there's nothing stopping you from labeling another DC as "medium-hard" and putting it between Moderate and Hard. 

They've given us the full color spectrum, and while we can pick out Red, Orange, Yellow, Green, Blue, and Violet, that doesn't mean that we only have six colors.

As far as whether they've succeeded or not, since they haven't actually finished skills, monster, or class design, no they haven't succeeded.  But they have succeeded at outlining a compelling theory and statement of design, that they're in the process of implementing.  Don't get hung up on the details of any given packet to prove that they've hit the target or not.  We have a long way to go.
D&D Next = D&D: Quantum Edition
I agree. And I don't mean to get 'hung up' on the details, I'm here to contribute to making D&D the best game it can be, like you. I'm just asking questions because I want D&D to 'make sense'. 

I think using 7 difficulty classes is actually better than 25, because then they mean more.

"Make things as simple as possible, but no simpler." - Albert Einstein

Under the hood, you find math that can help codify the reasons behind some things. When done correctly, it makes sense.

The DC categories rise in increments of 3. Training in a skill adds +3 to your ability check. Thus, having a skill increases your effectiveness by one step up the difficulty chart.
Under the hood, you find math that can help codify the reasons behind some things. When done correctly, it makes sense.

The DC categories rise in increments of 3. Training in a skill adds +3 to your ability check. Thus, having a skill increases your effectiveness by one step up the difficulty chart.



Good find. I didn't put that together.

"Make things as simple as possible, but no simpler." - Albert Einstein

Oh, and they did make those numbers public for 4e, (last page of the DMG rules update document) so you should check your research before posting wrong things.

Those are the numbers for monsters. Since PCs and monsters don't follow the same rules in 4e, the target numbers are not the same.

We will have the target number for skills, but that isn't the same as knowing what the expected bonuses for PCs are. Without out those numbers, we don't know how the target numbers actually compare to what PCs are expected to be able to do. Plus, bounded accuracy covers a lot more. We also need BAB and expected damage rolls at least, and probably expected AC and HP along with saves by level so we can determine if a PC or a class in general violates those standards.

And since the monsters have both offensive and defensive stats, defining just the monster half defines the whole system.

Funny how that works, hm?
D&D Next = D&D: Quantum Edition
Bounded Accuracy is a great concept, but if we are going to use it as a fundamental concept for game design, shouldn't it be clearly defined? I know 'Bounded Accuracy' is really just code for 'flatter math', but I believe placing actual boundaries on the system will create the internal consistency that the game requires. 

What do you think?

 

I think "bounded accuracy" is at least as "not D&D" as "fighters casting spells in 4e."   One of the things that makes D&D so interesting is the way your character constantly improves as you gain levels.  For classes that don't cast spells, the clearest and most obvious ways you improved were hit points and to hit and saving throw matrixes.  D&D Next keeps hit points but losing to hit and saving throw improvement seems like a major deviation from "D&D," to me.  

Of course, that's not a real difference, just like fighters didn't really cast spells in 4e.  Characters in D&D: Next are still going to advance, they'll gain more & better spells, more dice, and more & less desireable maneuvers.  At least some classes do get attack bonuses at higher levels.  They'll pick up magic items that give them bonuses.  The advancement is still there.  But the concept, itself, creates an appearance that it's absent, and a mere appearance was enough to create an unjustified, but dreadfully effective, outcry against 4e.  

D&D Next needs to be careful, lest it be D&D Last.

- Warlords! Join the 'Officer Country' Group! Join Grognards for 4e, the D&D that changed D&D.


D&D Home Page - What Class Are You? - Build A Character - D&D Compendium

One of the things that makes D&D so interesting is the way your character constantly improves as you gain levels.


I agree, and fortunately, your character will continue to constantly improve as you gain levels.

But, it may not improve in accuracy, automatically, for all characters ever, no matter what they are doing with themselves. 

Why do people insist on accuracy as the only acceptable form of character improvement?  Well, largely because non-spellcasters got screwed and couldn't have nice things, so to-hit was all they had left.  They can do better, and get real, meaningful improvements, as the spellcasters have been doing for decades.  But they're blind to the possibilities for character advancement beyond just to-hit, after growing accustomed to being abused for so long.
D&D Next = D&D: Quantum Edition
I think "bounded accuracy" is at least as "not D&D" as "fighters casting spells in 4e."   One of the things that makes D&D so interesting is the way your character constantly improves as you gain levels.  For classes that don't cast spells, the clearest and most obvious ways you improved were hit points and to hit and saving throw matrixes.  D&D Next keeps hit points but losing to hit and saving throw improvement seems like a major deviation from "D&D," to me.




Removing the 1/2 level shenanigans from 4th Ed like a gangrenous limb was the single best (mechanical) thing to ever hit my 4th Ed campaign.

 D&D Next keeps hit points but losing to hit and saving throw improvement seems like a major deviation from "D&D," to me.  

Of course, that's not a real difference, just like fighters didn't really cast spells in 4e.  Characters in D&D: Next are still going to advance, they'll gain more & better spells, more dice, and more & less desireable maneuvers.  At least some classes do get attack bonuses at higher levels.  They'll pick up magic items that give them bonuses.  The advancement is still there.  But the concept, itself, creates an appearance that it's absent, and a mere appearance was enough to create an unjustified, but dreadfully effective, outcry against 4e.  

D&D Next needs to be careful, lest it be D&D Last.




Guess what? Your to-hit and saving throws still improve, just at a slower rate. That's for all classes, not some of them. The good thing is that your improvements really are improvements, not just bonuses that are promptly matched on the other side, Red Queen-style.
And since the monsters have both offensive and defensive stats, defining just the monster half defines the whole system.

Not really. Using that information and the player character information from the character builder you might be able to reverse engineer a chart of the player character ACs, attack bonuses and damage by level and role. But there would be no way to be sure it matches WotC's expectations.

Bounded Accuracy is possibly my favourite thing out of everything in this edition.

So you have a door in your path.. It is either a Trivial, Easy, Moderate, Hard, Very Hard or Impossible challenge. Bounded Accuracy means that the Easy wooden door remains Easy for all 20(30) levels of character advancement.

The characters themselves do not rapidly or inherently gain bonuses for checks against breaking down or unlokcing  a door as they gain level (+1/2 level bonus for 4e or skill point increases in 3.5). This means that that iron hinged door which was a Moderate challenge is always a Moderate challenge. Sure, the character might pick up an ability which makes him better at picking the locks of doors throughout their advancement - but this is real advancement of the character - and so the Moderate door stays a Moderate door - but the Rogue finds all doors easier to unlock.

This is in wild contrast to previous editions, where the DC of a Moderate door changed with level. So the players never knew even roughly how hard the door before them was going to be to break down. 

Another effect of this was it caused a terrible case of DM Fiat. Because the Easy - Moderate - Hard DCs shifted with level, a DM never really knew what to set the DC at. Example follows:

I have a DM who threw level 30 Hard DCs at a level 8 Bard I played, because he knew the Bard wouldn't have a problem with the level 8 hard - despite I had put EVERYTHING into making that Bard good at Diplomacy - to the point that he was useless at combat. So because he used my Bard's check results as a guideline, my Bard was still only mediocre at Diplomacy - despite the fact that he could crash through checks that would have made a level 30 character struggle. Having a set, "This would be a hard thing to convince the city council of" would have solved that problem, completely and utterly.




Then, coming back to combat - yes, it means Plate armour is AC18 no matter what level you are at - or Leather is somewhere between 11 and 16 depending on the Dexterity of the wearer.

And so, in assuming that characters gain options, rather than hitting more often - characters that were moderatly hard to 'hit' at level 1 can still be a threat (in great numbers) at level 20. Which means that you can always throw goblins at the PCs, and they will know exactly what to expect. 
So. Maybe the solution isn't in a bounded scale. Maybe the solution is in creating an internal consistency of numbers in D&D. if a '5' actually meant something, it might feel more 'real'.


That's the whole point of Bounded Accuracy.  A DC 15 Strength check to break something is a DC 15 Strength check to break something, no matter what level the person doing the breaking is.  The DC is tied directly to the objective nature of the thing in question.  That's what Bounded Accuracy does.

In fact, a stated goal of the system is for the DM to just know what DC to make a thing, as a result of this internal consistency of numbers.  "easy" "medium" and "hard" are no longer moving targets.

Yes, PCs will improve over time.  What they might consider easy relative to their own capability may change.  But the objective label won't.  The fact that they can, actually, hit those "Hard" DCs easily is the real character advancement, not just keeping pace with the treadmill.



I agree 100% with this. One of the things that always confuse dme about 4E was that in order to put together a 'by the books' (or RAW if you will) adventure, I had to use DCs that were reflected by the level of the PCs. So if they went into an old dwarven ruin at say level 4, finding that really hard to spot secret door would be DC 21. Now suppose they don't find it. If I decide that they need to go back to that set of ruins at level 26 for some reason (a dragon took them over, giants are using old dwarven runes found there to open a gate into the Elemental Chaos, whatever, then if I were to go 'by the books, or RAW, then the DC to find that door would now be a DC 39. How does that make any sense?

Now granted, I can easily decide that the DC is still 21, but then the challenge would not be up to par for the PCs anymore. I much prefer the way DDN is doing skills and DCs. Makes a hell of a lot more sense. Opening that lock to that gate? Always a DC 13. Not 8 then 13 then 19. Seems more realistic to me. Also the way they are currently doing skills? Nothing new. Go back and look at Proficiencies in AD&D 2nd. It's not exactly the same, but it's more in that neighborhood than 3/4e skills.   

  

My current feeling is that accuracy is too high. I want my hits to matter and if the game scales up by giving really big damage and really big hit points then I'm going to feel like the attack roll is meaningless and I may as well just roll damage and be done.


A lvl 10 dude using their main mojo with a 20 stat has an 80% chance to hit something with an AC of 15, and 18 is the highest AC in the playtest. I'd much prefer the accuracy point swing around a pole of about 65% or so.



With skill rolls it's even worse, 'cause there are things a master can do that a novice can't even attempt so in relative terms, an easy roll is not and should not be the same at lvl 1 as it is at lvl 10. It doesn't make any real-life sense even if it makes mathematical sense.


Course, Mearls did say he was gonna give skills another pass on twitter so I'm happy to see what he comes up with.

With skill rolls it's even worse, 'cause there are things a master can do that a novice can't even attempt so in relative terms, an easy roll is not and should not be the same at lvl 1 as it is at lvl 10. It doesn't make any real-life sense even if it makes mathematical sense.

Course, Mearls did say he was gonna give skills another pass on twitter so I'm happy to see what he comes up with.



You're missunderstanding. The difficulty is not in terms of character difficulty. A Rogue at level 10 will be better at falling than one at level 1 - if he wants to be.



Firstly there is the flat Ability boost, which can raise it 10 percent.

Then, if he wasn't trained, he can become trained.

Then, if the rogue really wants to be good at dropping from high places, he takes Skill Speciality and Skill Mastery.

Then, there is the +1 bonus that characters get to add to their skills.

So that pushes something that would have been hard at level 1... to be moderate at level 10 - but the DM still goes, "Well, its a wooden barred door, so I make the DC 13" If anything there is more improvement than in previous editions because the DCs don't shift with character level.