Bounded Accuracy. Seems To Be A Deal Breaker For Me.


 Never really thought to throw my toys in the forums here over what would cause me to reject D&DN but after playing through the Isle of Dread and the test adventures I have found one.

 A deal breaker for me is something that would cause me to reject D&DN outright. Before I started testing I had the following deal breakers.

1. 4th ed reprint.
2. Spellcaster power levels similar to third ed.
3. DM headaches a'la 3rd ed.

 The D&D skills or ability checks or whatever you want to call them do not work that well. A DC 16 check which is supposed to be hard is almost a guranteed success when 4-5 peopel are rolling it. Skill mastery and mighty exertion also totally ruin it.

 I admit I was spoinled by Star Wars Saga which had a great skill system which was basically 4th eds one but with a revised skill list and the way skills interacted with class abilities. Pathfinder uses and improved 3.5 skill system, and whatever problems 4th ed had I don't think the skills were a major part of it (except maybe +5 skill focus and +5 being trained being to high maybe +3).

 Maybe they are trying to implement a pre 3rd ed approach to the game but even their own system doesn't work very well. Not sure how many pre d20 players are left and I don't know if they actually care about skills at all.

 Why can't we use 4th eds skill system (tweak numbers if needed) with a revised 3.5 skill list? Skill focus gives you advanatage on a skill of your choice. If it isn't broken don't fix it. Insert some comment about reinventing the wheel. It would be a nice evolution of d20 and I don't think it would upset to many people from the 3rd ed and 4thed camps. It might upset the pre d20 players IDK.

 Put simply I have a large pile of 3.5, Pathfinder and 4th ed material I can use. All of which can be tweaked to something I like better than expertise dice wrecking the weak at best skill system. D&DN is kinda fun but it just seems bad in this regard.

 Fear is the Mind Killer

 Fear is the Mind Killer  

I think their line about "it's all attribute checks" is a load of wash 'cause the skills certainly are skills, even if you can make untrained rolls.


I like skill points based on class, but what makes it a mess is when you start to multi heavily and it becomes difficult to know how many skills you're supposed to have. I can't walk up to 3e and say "I want a fighter/mage" and then know what my skills are and while I never really messed with pathfinder it's my understanding that they managed to clean this up a bit.


What I would really like is for skills to be more than a zero-sum game. In Next right now, I either have the skill or I don't. I can't be someone who took the crash course and never pursued it, I can't do that because the skill either gives me all of the bonus or none of it.


And to me, bounded accuracy (as much as I really hate how narrow the bounds are right now) make the dabbler work really well 'cause the DCs stay relatively the same for most of the game. To me, this is just a failure to understand the advantages of their own idea.

The D&D skills or ability checks or whatever you want to call them do not work that well. A DC 16 check which is supposed to be hard is almost a guranteed success when 4-5 peopel are rolling it. Skill mastery and mighty exertion also totally ruin it.


This same problem crops up in practically every rpg in existence.

As DM, I don't let all 4-5 characters roll for the skill check.  Rather, I have whomever is the best character roll for it with everyone else using aid another (help in the playtest). 
Bounded Accuracy is fine.
It's just that the DCs are too low and the bound range is too small and small as a result.

Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds. Constitution Based Class for Next!

I think by now even the devs have realized that things like skill mastery/mighty exhertion are playing havoc with bounded accuracy.  But I see that as more of a flaw with skill mastery/mighty exhertion than with bounded accuracy.  I think bounded accuracy has its flaws, but I don't think the OP really pointed to any of them.  Even if you tossed out bounded accuracy, all that would mean is that at higher levels it becomes more difficult for the "group check" to power through even the hardest challenges, it wouldn't fix the problem at low levels and once you do something else to fix that (like ole one eye's solution) you don't need to toss bounded accuracy to fix it.  

The thing I don't like about bounded accuracy is this ludicrous concept of "real improvement."  The idea that if a +1 is assumed in certain places and just keeps you in line with assumed system math, it's not a real bonus and is somehow diminished in importance.  This is a common but irrational misperception, and one that is breaking down the system.  When you get a +5 and the DC goes up by 5, you still need the same die roll to succeed, yes.  But it's not the same task.  You are now doing a heroically difficult thing and making it look as easy as a normal person would find a substantially easier task.  It's not a treadmill, it's not boring.  What is boring is getting a +5 when all the DCs are still in the same place, such that it becomes impossibe to fail at anything, ever.  That is what "real improvement" means.  It's not a good thing.  Now I'll grant that in skills world, higher level characters with bigger bonuses are supposed to face harder DCs more often (a tenth level party should face "formiddable" obstacles a lot more often than a first level party).  But insofar as that's true you're still on the treadmill.  Insofar as it isn't true, like when you get a +3 bonus from a magic item that "isn't assumed by the system math," I think it's a bad thing.  That's what's wrong with "bounded accuracy" for me.
There's nothing wrong with the idea behind bounded accuracy.

The skill system just isn't implemented properly yet.  They haven't done the sort of tuning that you're requiring for it to not be a dealbreaker.  Which isn't surprising, since a lot of things are dealbreakers if they aren't finished being developed.
D&D Next = D&D: Quantum Edition
I adore Star Wars Saga Edition, as stated, it was a snapshot into 4th Ed design at the time (sorry for the bad rhyme); it seems they went more with the Tome of Battle action (those guys won).

I love Bounded Accuracy, the old best thing to happen to this game since sliced-bread. 



Wait!  Your game slices bread?  ::swears::  I knew I got a defective version!
The thing I don't like about bounded accuracy is this ludicrous concept of "real improvement."  The idea that if a +1 is assumed in certain places and just keeps you in line with assumed system math, it's not a real bonus and is somehow diminished in importance.  This is a common but irrational misperception, and one that is breaking down the system.  When you get a +5 and the DC goes up by 5, you still need the same die roll to succeed, yes.  But it's not the same task.  You are now doing a heroically difficult thing and making it look as easy as a normal person would find a substantially easier task.  It's not a treadmill, it's not boring.  What is boring is getting a +5 when all the DCs are still in the same place, such that it becomes impossibe to fail at anything, ever.  That is what "real improvement" means.  It's not a good thing.  Now I'll grant that in skills world, higher level characters with bigger bonuses are supposed to face harder DCs more often (a tenth level party should face "formiddable" obstacles a lot more often than a first level party).  But insofar as that's true you're still on the treadmill.  Insofar as it isn't true, like when you get a +3 bonus from a magic item that "isn't assumed by the system math," I think it's a bad thing.  That's what's wrong with "bounded accuracy" for me.



Ding! Exactly. Binding accuracy within a specific array doesn't actually fix any of the problems associated with making checks in game, it just obfuscates the big math and highlights the need for consistent DCs.


The fact that it redirects our attention to the DCs is probably a good thing but I seriously doubt there is a game out there with no bounds within its difficulty arrays at all; the system has got to assume certain number values are easy, hard and difficult to get with a check.

Magic items.


It's tricky, this 'cause I like that they're considered extras but powerroleplayer's right: they run the risk of basically wrecking the game experience because there is no inclusion at all in the system math. Even AD&D assumed that everyone got some magic treasure as they levelled, though it wasn't so firm as wealth by level.


Actually the wealth by level thing was born in the 2e DMG if I remember correctly with a section on replacing PCs which included a table of recommended wealth for said replacements. Either way, there were definitely assumptions about what kinds of magic were available to a party at a given level. I like that the system math makes no assumptions about magic, but I also question how viable this approach will be.

I think trained skills should just e ability checks with advantage...

It keeps in line with bounded accuracy and makes being trained truly feel important.
well the problem with that is it still doesn't really reflect advancement properly. There are things a novice cannot do that a master will do without thinking, and just slapping advantage on it and calling it training doesn't allow for that.
I think trained skills should just e ability checks with advantage... It keeps in line with bounded accuracy and makes being trained truly feel important.



I like the idea; however, I would suggest some changes to incorporate skill advancement. 

First, I would change it from an advantage mechanic to a xd20 take the highest mechanic.  A skilled character would get 2d20 (take the highest) for his or her attribute check.  As the character advances, he or she can opt to gain additional die to the check.  For example, a more specialized character can gain 3d20 (take the highest).

Second, I would tier the number of bonus die.  A character between level 1-5 would cap at two dice.  At 6-10, the character would cap at three dice.  Most of us can extrapolate out to level 20.  

Why? 

First, additional dice do provide a significant boost while preserving the upper and lower bounds.  DC 25 is currently the "Nearly Impossible" upper bound.  Under the current system, a first level character has, at least one that is correctly built, a 20% chance of beating a DC 25.  At level 10, the percentage could increase to 45%.  This, to me, is off.  By providing additional die, an unskilled character a 5% chance at meeting a DC 25.  A skilled character has a 9.75% of beating a DC 25.  Further skill training would bump this percentage by approximately 5%.  This makes meeting a DC 25 more likely by 5% per tier of training and is easier to balance from the game mechanics prespective. 

Second, it helps mitigate failure without introducing additional mechanics like skill mastery.  Assuming a +5 attribute bonus, a first level character would need an 11 to make a DC 16 check.  An unskilled character would beat this DC 50% of the time.  Under the current playtest, a skilled character meets this DC 65% of the time.  Under the proposed system, the character would make the roll 75% of the time.  Futher training would provide: 87.5% at 3 dice, 93.75% at 4 dice and 96.88 at 5 dice.  

Assuming a +5 attribute bonus, the numbers are:

DC 07 - 95% at 1 die, 99.75% at 2 dice, 99.99% at 3 dice, 99.99% at 4 dice.*
DC 10 - 80% at 1 die, 96.00% at 2 dice, 99.20% at 3 dice, 99.84% at 4 dice.
DC 13 - 65% at 1 die, 87.75% at 2 dice, 95.71% at 3 dice, 98.50% at 4 dice.
DC 16 - 50% at 1 die, 75.00% at 2 dice, 87.50% at 3 dice, 93.75% at 4 dice.  
DC 19 - 35% at 1 die, 57.75% at 2 dice, 72.54% at 3 dice, 82.15% at 4 dice.
DC 22 - 20% at 1 die, 36.00% at 2 dice, 48.80% at 3 dice, 59.04% at 4 dice.
DC 25 - 05% at 1 die, 09.75% at 2 dice, 14.26% at 3 dice, 18.55% at 4 dice.
* 3 dice or higher actually brings the percentage up 100% with a precision of 2 decimal points.

With the old system assuming a +5 attribute bonus and assuming that a 1 is an automatic failure, the numbers would be:

DC07 - 95% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC10 - 80% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC13 - 65% at +0, 80% at +3, 85% at +4, 90% at +5, 95% at +6, 95% at +7
DC16 - 50% at +0, 65% at +3, 70% at +4, 75% at +5, 80% at +6, 85% at +7
DC19 - 35% at +0, 50% at +3, 55% at +4, 60% at +5, 65% at +6, 70% at +7
DC22 - 20% at +0, 35% at +3, 40% at +4, 45% at +5, 50% at +6, 55% at +7
DC25 - 05% at +0, 20% at +3, 25% at +4, 30% at +5, 35% at +6, 45% at +7     

I don't like the straight bonus due it skewing the high end DCs.  The low and mid end numbers tend to do better overall with extra dice.  The DC07 line of the proposed system also shows the percent chance of rolling a one.
There's nothing wrong with the idea behind bounded accuracy.

The skill system just isn't implemented properly yet.  They haven't done the sort of tuning that you're requiring for it to not be a dealbreaker.  Which isn't surprising, since a lot of things are dealbreakers if they aren't finished being developed.



I will trust your word of the "skill system isn't done yet"...because the skill bonus/DC as you level up is what i am extremly pesimist about it...thought i am worried of how damage will be what scale in a way to compensate for the bounded accuracy, because so far all the damage increase on the game is based on....MOAR DICES!!! and that actually slow down the game on the table quite alot because damage calculation is usually what take more tiem in my experience on the table.
I think trained skills should just e ability checks with advantage... It keeps in line with bounded accuracy and makes being trained truly feel important.



I like the idea; however, I would suggest some changes to incorporate skill advancement. 

First, I would change it from an advantage mechanic to a xd20 take the highest mechanic.  A skilled character would get 2d20 (take the highest) for his or her attribute check.  As the character advances, he or she can opt to gain additional die to the check.  For example, a more specialized character can gain 3d20 (take the highest).

Second, I would tier the number of bonus die.  A character between level 1-5 would cap at two dice.  At 6-10, the character would cap at three dice.  Most of us can extrapolate out to level 20.  

Why? 

First, additional dice do provide a significant boost while preserving the upper and lower bounds.  DC 25 is currently the "Nearly Impossible" upper bound.  Under the current system, a first level character has, at least one that is correctly built, a 20% chance of beating a DC 25.  At level 10, the percentage could increase to 45%.  This, to me, is off.  By providing additional die, an unskilled character a 5% chance at meeting a DC 25.  A skilled character has a 9.75% of beating a DC 25.  Further skill training would bump this percentage by approximately 5%.  This makes meeting a DC 25 more likely by 5% per tier of training and is easier to balance from the game mechanics prespective. 

Second, it helps mitigate failure without introducing additional mechanics like skill mastery.  Assuming a +5 attribute bonus, a first level character would need an 11 to make a DC 16 check.  An unskilled character would beat this DC 50% of the time.  Under the current playtest, a skilled character meets this DC 65% of the time.  Under the proposed system, the character would make the roll 75% of the time.  Futher training would provide: 87.5% at 3 dice, 93.75% at 4 dice and 96.88 at 5 dice.  

Assuming a +5 attribute bonus, the numbers are:

DC 07 - 95% at 1 die, 99.75% at 2 dice, 99.99% at 3 dice, 99.99% at 4 dice.*
DC 10 - 80% at 1 die, 96.00% at 2 dice, 99.20% at 3 dice, 99.84% at 4 dice.
DC 13 - 65% at 1 die, 87.75% at 2 dice, 95.71% at 3 dice, 98.50% at 4 dice.
DC 16 - 50% at 1 die, 75.00% at 2 dice, 87.50% at 3 dice, 93.75% at 4 dice.  
DC 19 - 35% at 1 die, 57.75% at 2 dice, 72.54% at 3 dice, 82.15% at 4 dice.
DC 22 - 20% at 1 die, 36.00% at 2 dice, 48.80% at 3 dice, 59.04% at 4 dice.
DC 25 - 05% at 1 die, 09.75% at 2 dice, 14.26% at 3 dice, 18.55% at 4 dice.
* 3 dice or higher actually brings the percentage up 100% with a precision of 2 decimal points.

With the old system assuming a +5 attribute bonus and assuming that a 1 is an automatic failure, the numbers would be:

DC07 - 95% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC10 - 80% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC13 - 65% at +0, 80% at +3, 85% at +4, 90% at +5, 95% at +6, 95% at +7
DC16 - 50% at +0, 65% at +3, 70% at +4, 75% at +5, 80% at +6, 85% at +7
DC19 - 35% at +0, 50% at +3, 55% at +4, 60% at +5, 65% at +6, 70% at +7
DC22 - 20% at +0, 35% at +3, 40% at +4, 45% at +5, 50% at +6, 55% at +7
DC25 - 05% at +0, 20% at +3, 25% at +4, 30% at +5, 35% at +6, 45% at +7     

I don't like the straight bonus due it skewing the high end DCs.  The low and mid end numbers tend to do better overall with extra dice.  The DC07 line of the proposed system also shows the percent chance of rolling a one.



I can't find anywhere in the latest play test packet where anything but attacks have auto success on a 20 and auto failure on a 1.
 Would people prefer some sort of tweaked skills from another version of d20?

 Fear is the Mind Killer

 Fear is the Mind Killer  

I like the idea of simplified consolidated skills ala 4e, with training granting "roll 2d20 and take highest" rather than a +X bonus. Expert skill training feat would grant +3 bonus to the skill roll and an additional bonus equal to 1/4 level. This means someone who is an expert slowly increases the skill automatically as they level.
 Would people prefer some sort of tweaked skills from another version of d20?



I think everyone would like higher bounds.

If you have a +8 (+5 stat, +3 training) at level 1 in a skill, you have the following percent chances to beat a skill check:

DC   Percent   Notes
1     100%
2     100%
3     100%
4     100%
5     100%
6     100%
7     100%     Trivial
9     100%
10   95%       Easy
11   90%
12   85%
13   80%       Moderate
14   75%
15   70%
16   65%       Hard
17   60%
18   55%
19   50%       Very Hard
20   45%
21   40%
22   35%       Formidable
23   30%
24   25%
25   20%       Nearly Impossible

So a level 1 character has a 20% chance at a nearly impossible skill check. That shouldn't happen. To fix it all they have to do is up the categories. Like this.

       +8          +0
DC   Percent   Percent   Notes
1     100%      100%
2     100%      95%
3     100%      90%
4     100%      85%
5     100%      80%       Trivial (untrained, untalented 75% chance, trained, talented ha 100%)
6     100%      75%
7     100%      70%
8     100%      65%
9     100%      60%
10   95%        55%       Easy (untrained, untalented 50% chance, trained, talented has 95%)
11   90%        50%
12   85%        45%
13   80%        40%
14   75%        35%
15   70%        30%       Moderate (untrained, untalented 30% chance, trained, talented has 70%)
16   65%        25%
17   60%        20%
18   55%        15%
19   50%        10%
20   45%        5%         Hard (untrained, untalented 5% chance, trained, talented has 45%)
21   40%        0%
22   35%        0%
23   30%        0%
24   25%        0%
25   20%        0%         Very Hard (trained and talented 20%)
26   15%        0%
27   10%        0%
28   5%          0%
29   0%          0%
30   0%          0%         Formidable (trained and talented by level 20 can have a +12 has a 15% chance)
31   0%          0%
32   0%          0%
33   0%          0%
34   0%          0%
35   0%          0%        Nearly Impossible (trained and talented by level 20 can have a +12 has a 0% chance)

This is what it should look like.
abanathie:

First off, shouldn't it be 40% max?  +7 skill +5 attribute = 12, 25-12=13, which is a 40% chance.  Second of all, I find it extremely odd that you're assuming a +5 attribute score at level one, as that's attainable only if you roll and that only if you roll exceptionally well and that only on a small subset of activities that the one stat you might ever get to 20 applies to.  Same goes for the +7, you're only ever going to get at most one skill to +7 by level 10.   Third, to my mind, it's you're system that "skews" the high end DCs downward, while straight bonuses shift the probabilities in an unskewed fashion.  Skewed suggests disproportionately effecting one end relative to the other.  Straight bonuses add 5 percentage points to all DCs, while extra dice add 25 percentage points in the middle and 4.75 on the ends.  Neither of them are adding even percentages, so that can't be it.  So how is it bonuses that are skewing things?


But far more importantly, why is it so important to you to preserve the upper and lower bounds?  And don't tell me "bounded accuracy," because that's not what bounded accuracy means.  A tenth level character should be able to do things that a 1st level character cannot, not just do the easy stuff more reliably.  It seems to me, a 10th level rogue who has put all his bonuses into lock picking should be able to pick locks that an ordinary first level character could not, no matter how well he rolled.  Even among things an untrained character should be able to do when extremely lucky, the true master should be able to do them more than 18.5% of the time.  No matter how many times I try to do a back flip, I will never ever succeed (unless I do it enough times that I get a training bonus and stronger legs, but that's another story).  But an expert tumbler should be able to do one every time.  True, D&D has never and will never reflect this completely because bonuses have always been dwarfed by the d20.  But that doesn't mean we should move even farther away from reflecting this in the name of some misinterpretation of an overvalued concept.  So it seems to me that the very benefits you seek to achieve with your alternative system are of negative value, so there's no reason to put up with the extra time costs of rolling additional dice (which get non-trivial by 4 dice when you may have to beg a d20 off a friend) and the total removal of any meaningful chance of failure at even hard DCs (without actually admitting that chance of failure has been removed such that we can just dispense with the die rolling the way we can when the bonus exceeds the DC).
First off, shouldn't it be 40% max?  +7 skill +5 attribute = 12, 25-12=13, which is a 40% chance.  Second of all, I find it extremely odd that you're assuming a +5 attribute score at level one, as that's attainable only if you roll and that only if you roll exceptionally well and that only on a small subset of activities that the one stat you might ever get to 20 applies to.  Same goes for the +7, you're only ever going to get at most one skill to +7 by level 10.   Third, to my mind, it's you're system that "skews" the high end DCs downward, while straight bonuses shift the probabilities in an unskewed fashion.  Skewed suggests disproportionately effecting one end relative to the other.  Straight bonuses add 5 percentage points to all DCs, while extra dice add 25 percentage points in the middle and 4.75 on the ends.  Neither of them are adding even percentages, so that can't be it.  So how is it bonuses that are skewing things?



First off who are you responding to?

Second off, yeah +7 max, let me adjust my last post to account for it. Your math is off. You can beat the check by rolling the DC. So you would do this: 100 - ((25-12-1) * 5) = 40% chance of beating a DC 25 check. We are looking at the outliers. The ones that can break the system. By level 20 you will have +10 to spread around with a max of +7 on any skill. So with trained skills starting at +3 and getting 3-4 of them regardless of class, they only need +4 each so you would end up with 2 trained skills at +7 and one trained skill at +5. Spread that around a party of five characters with each having different skill sets and you have at least one master at most skills. With some skills being more important than others.

But far more importantly, why is it so important to you to preserve the upper and lower bounds?  And don't tell me "bounded accuracy," because that's not what bounded accuracy means.  A tenth level character should be able to do things that a 1st level character cannot, not just do the easy stuff more reliably.  It seems to me, a 10th level rogue who has put all his bonuses into lock picking should be able to pick locks that an ordinary first level character could not, no matter how well he rolled.  Even among things an untrained character should be able to do when extremely lucky, the true master should be able to do them more than 18.5% of the time.  No matter how many times I try to do a back flip, I will never ever succeed (unless I do it enough times that I get a training bonus and stronger legs, but that's another story).  But an expert tumbler should be able to do one every time.  True, D&D has never and will never reflect this completely because bonuses have always been dwarfed by the d20.  But that doesn't mean we should move even farther away from reflecting this in the name of some misinterpretation of an overvalued concept.  So it seems to me that the very benefits you seek to achieve with your alternative system are of negative value, so there's no reason to put up with the extra time costs of rolling additional dice (which get non-trivial by 4 dice when you may have to beg a d20 off a friend) and the total removal of any meaningful chance of failure at even hard DCs (without actually admitting that chance of failure has been removed such that we can just dispense with the die rolling the way we can when the bonus exceeds the DC).



I can tell from this part that you aren't responding to me at all. Since in my suggested chart a high level Rogue that mastered lock picking would be just a a few points shy of being able to pull off nearly impossible challenges while the trained and talented 1st level character would have no chance.

Edit: Posted before your edit.
Don't based you calculation around characters with max ability at level 1.  The curent playtest is based on a character starting with a 16-18 on the primary stat.

so a skilled level 1 PCs is running +7 best on their checks is

DC 10 90%
DC 13 75%
DC 16 60%
DC 19 45%
DC 22 30%
DC 25 15%

An unskilled untalented character with a +0 is running

DC 10 55%
DC 13 40%
DC 16 25%
DC 19 10%
DC 22 0%
DC 25 0%

So a DC 19 should be something a ordinary person suceeds 1 out of 10 times but a talented and skilled person can do about half the time (like bust through a wooden door with a average lock)

A DC 22 has to be something ordinary folk have NO CHANCE at but the talented expert has a little less that 1/3 chance at succed (breaking manacles, and busting metal locks).

Basically, increase all DCs in the DM Guidelines by 3 and the system makes sense.

Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds. Constitution Based Class for Next!

I don't agree with most of the complaints here. To the exclusion of one, I do not like the fact that party's always get multiple checks (because of multiple charachters) to achieve something. I house rule that only allows one chacrachter an attempt to persuade, if other characters have the skill also they can aid (giving advantage). Allowing 4 charachters to roll makes ALL challenges too easy.

Aside from that 1st level characters  having 10% chance to do something nearly impossible is perfect. It shoudl almost never happen and if it does, sweet epic moment for the players. If something is impossible, I just say its impossible.

I don't see any positives to having giant ranges, what for? Just to make low level characters weaker? Is that the goal? I feel pretty good with the ranges and the difficulties as they are now (minus skill mastery). A bigger problem in my mind is that too many monsters have too low skill bonuses.

My mind is a deal-breaker.

Being trained in a skill means you have a better chance to succeed.  A level 1 character with a +3 atttibute bonus will have a hard time doing things he's not trained to do. So that's working.

A PC who is built to be the good at his skill should have an easy time at a hard DC and also get a real shot at doing something crazy or nearly impossible. So a guy with a +3 attribute and skill training has a 55% chance to succeed at a hard task and a 10% chance of doing something nearly impossible.  So that's working, IMO.

Now, a guy with a +5 atttibute, skill training, and spending all his skill-ups on one particular area is telling you, "I  want to be the freaking king of doing this!" He's got a total of +12 to do his one specific skill. It comes at the cost of being fairly average at everything else.  He'd walking through hard DCs without aatting an eye, and he has a decent shot at what would be suicide for anyone else to consider. It seems to me, this is fine.

If I want the to be the best lockpicker in the world, I want my DM to include crazy locks in the game that nobody else has a shot at.  I want to break into Bahama's car and steal his stereo.  But to do that, I'm still sweating most of my other checks.  Escape manacles? I'm no better than the next guy.  Decipher a scroll? Maybe I shouldn't even bother.

I don't think this is failing.  In practice, it seems to be pretty balanced. 

Now, skill mastery is another story but the basic system seems fine.
 
Orzel that would bump the upper limit to DC 28? I would round that up to 30. 30 or 35 should be the pper limit IMHO as a 10th level whatever should be able to do something a levle 1 character in normal circumstance could not dd regardless of how good they rolled.

No prolems if magic items or mundane means grants a large bonus to a skill check (+10 bonus on knowledge in the Great Library, +10 craft:ship in shipyard etc).

 Fear is the Mind Killer

 Fear is the Mind Killer  

Orzel that would bump the upper limit to DC 28? I would round that up to 30. 30 or 35 should be the pper limit IMHO as a 10th level whatever should be able to do something a levle 1 character in normal circumstance could not dd regardless of how good they rolled.

No prolems if magic items or mundane means grants a large bonus to a skill check (+10 bonus on knowledge in the Great Library, +10 craft:ship in shipyard etc).




DC 28 is Perfect

Max at level 1 is +8 with is a 5%
Max at level 10 is +12 whish is 25%

Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds. Constitution Based Class for Next!

I think trained skills should just e ability checks with advantage... It keeps in line with bounded accuracy and makes being trained truly feel important.



I like the idea; however, I would suggest some changes to incorporate skill advancement. 

First, I would change it from an advantage mechanic to a xd20 take the highest mechanic.  A skilled character would get 2d20 (take the highest) for his or her attribute check.  As the character advances, he or she can opt to gain additional die to the check.  For example, a more specialized character can gain 3d20 (take the highest).

Second, I would tier the number of bonus die.  A character between level 1-5 would cap at two dice.  At 6-10, the character would cap at three dice.  Most of us can extrapolate out to level 20.  

Why? 

First, additional dice do provide a significant boost while preserving the upper and lower bounds.  DC 25 is currently the "Nearly Impossible" upper bound.  Under the current system, a first level character has, at least one that is correctly built, a 20% chance of beating a DC 25.  At level 10, the percentage could increase to 45%.  This, to me, is off.  By providing additional die, an unskilled character a 5% chance at meeting a DC 25.  A skilled character has a 9.75% of beating a DC 25.  Further skill training would bump this percentage by approximately 5%.  This makes meeting a DC 25 more likely by 5% per tier of training and is easier to balance from the game mechanics prespective. 

Second, it helps mitigate failure without introducing additional mechanics like skill mastery.  Assuming a +5 attribute bonus, a first level character would need an 11 to make a DC 16 check.  An unskilled character would beat this DC 50% of the time.  Under the current playtest, a skilled character meets this DC 65% of the time.  Under the proposed system, the character would make the roll 75% of the time.  Futher training would provide: 87.5% at 3 dice, 93.75% at 4 dice and 96.88 at 5 dice.  

Assuming a +5 attribute bonus, the numbers are:

DC 07 - 95% at 1 die, 99.75% at 2 dice, 99.99% at 3 dice, 99.99% at 4 dice.*
DC 10 - 80% at 1 die, 96.00% at 2 dice, 99.20% at 3 dice, 99.84% at 4 dice.
DC 13 - 65% at 1 die, 87.75% at 2 dice, 95.71% at 3 dice, 98.50% at 4 dice.
DC 16 - 50% at 1 die, 75.00% at 2 dice, 87.50% at 3 dice, 93.75% at 4 dice.  
DC 19 - 35% at 1 die, 57.75% at 2 dice, 72.54% at 3 dice, 82.15% at 4 dice.
DC 22 - 20% at 1 die, 36.00% at 2 dice, 48.80% at 3 dice, 59.04% at 4 dice.
DC 25 - 05% at 1 die, 09.75% at 2 dice, 14.26% at 3 dice, 18.55% at 4 dice.
* 3 dice or higher actually brings the percentage up 100% with a precision of 2 decimal points.

With the old system assuming a +5 attribute bonus and assuming that a 1 is an automatic failure, the numbers would be:

DC07 - 95% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC10 - 80% at +0, 95% at +3, 95% at +4, 95% at +5, 95% at +6, 95% at +7
DC13 - 65% at +0, 80% at +3, 85% at +4, 90% at +5, 95% at +6, 95% at +7
DC16 - 50% at +0, 65% at +3, 70% at +4, 75% at +5, 80% at +6, 85% at +7
DC19 - 35% at +0, 50% at +3, 55% at +4, 60% at +5, 65% at +6, 70% at +7
DC22 - 20% at +0, 35% at +3, 40% at +4, 45% at +5, 50% at +6, 55% at +7
DC25 - 05% at +0, 20% at +3, 25% at +4, 30% at +5, 35% at +6, 45% at +7     

I don't like the straight bonus due it skewing the high end DCs.  The low and mid end numbers tend to do better overall with extra dice.  The DC07 line of the proposed system also shows the percent chance of rolling a one.



I can't find anywhere in the latest play test packet where anything but attacks have auto success on a 20 and auto failure on a 1.



Just a thought for players of older editions...
The plus five to attribute is based on whether it is possible to 18 (which is it possible when you roll) in any of the attribute creation systems currently in the game.  I tend to lean to worse and best case scenarios.
Orzel that would bump the upper limit to DC 28? I would round that up to 30. 30 or 35 should be the pper limit IMHO as a 10th level whatever should be able to do something a levle 1 character in normal circumstance could not dd regardless of how good they rolled.

No prolems if magic items or mundane means grants a large bonus to a skill check (+10 bonus on knowledge in the Great Library, +10 craft:ship in shipyard etc).




DC 28 is Perfect

Max at level 1 is +8 with is a 5%
Max at level 10 is +12 whish is 25%




 DC28 still gets wrecked by expertise dice and they seem determined to shoe horn that mechanic onto the non spellcasting classes. I liked that mechanic on the fighter, not sure if I want it used on multiple classes.

 Fear is the Mind Killer

 Fear is the Mind Killer  

abanathie:

First off, shouldn't it be 40% max?  +7 skill +5 attribute = 12, 25-12=13, which is a 40% chance.  Second of all, I find it extremely odd that you're assuming a +5 attribute score at level one, as that's attainable only if you roll and that only if you roll exceptionally well and that only on a small subset of activities that the one stat you might ever get to 20 applies to.  Same goes for the +7, you're only ever going to get at most one skill to +7 by level 10.   Third, to my mind, it's you're system that "skews" the high end DCs downward, while straight bonuses shift the probabilities in an unskewed fashion.  Skewed suggests disproportionately effecting one end relative to the other.  Straight bonuses add 5 percentage points to all DCs, while extra dice add 25 percentage points in the middle and 4.75 on the ends.  Neither of them are adding even percentages, so that can't be it.  So how is it bonuses that are skewing things?


But far more importantly, why is it so important to you to preserve the upper and lower bounds?  And don't tell me "bounded accuracy," because that's not what bounded accuracy means.  A tenth level character should be able to do things that a 1st level character cannot, not just do the easy stuff more reliably.  It seems to me, a 10th level rogue who has put all his bonuses into lock picking should be able to pick locks that an ordinary first level character could not, no matter how well he rolled.  Even among things an untrained character should be able to do when extremely lucky, the true master should be able to do them more than 18.5% of the time.  No matter how many times I try to do a back flip, I will never ever succeed (unless I do it enough times that I get a training bonus and stronger legs, but that's another story).  But an expert tumbler should be able to do one every time.  True, D&D has never and will never reflect this completely because bonuses have always been dwarfed by the d20.  But that doesn't mean we should move even farther away from reflecting this in the name of some misinterpretation of an overvalued concept.  So it seems to me that the very benefits you seek to achieve with your alternative system are of negative value, so there's no reason to put up with the extra time costs of rolling additional dice (which get non-trivial by 4 dice when you may have to beg a d20 off a friend) and the total removal of any meaningful chance of failure at even hard DCs (without actually admitting that chance of failure has been removed such that we can just dispense with the die rolling the way we can when the bonus exceeds the DC).



You're right; I mistyped the last number (40 vs 45).  However, 40% on a nearly impossible task seems excessively high to me (even at level 10).  So, I would like a flatter system that actually caps at the max.  If you want to balloon the numbers at the high end, that is one way of going with it; it's just not good for me.  In addition, I was attempting to address two issues at one time: capping the top level and providing a mitigation effect at the low end.
All of issues with Bounded Accuracy on skill checks would be solved by using the whole ability score rather than the modifier.  Make a hard check something like DC 30, where someone with strength 20 could succeed on a 10 and someone with strength 9 could never possibly succeed.

The metagame is not the game.

All of issues with Bounded Accuracy on skill checks would be solved by using the whole ability score rather than the modifier.  Make a hard check something like DC 30, where someone with strength 20 could succeed on a 10 and someone with strength 9 could never possibly succeed.



No, it just shifts the numbers upwards.  You still have to deal with skill points and specialization on some level.
Just my 2cp: I like bounded accuracy and the designers/developers have announced that they will be working on hammering out skills a bit more next, since they feel that are getting close to having many of the other core elements locked down. 
Orzel that would bump the upper limit to DC 28? I would round that up to 30. 30 or 35 should be the pper limit IMHO as a 10th level whatever should be able to do something a levle 1 character in normal circumstance could not dd regardless of how good they rolled.

No prolems if magic items or mundane means grants a large bonus to a skill check (+10 bonus on knowledge in the Great Library, +10 craft:ship in shipyard etc).




DC 28 is Perfect

Max at level 1 is +8 with is a 5%
Max at level 10 is +12 whish is 25%




 DC28 still gets wrecked by expertise dice and they seem determined to shoe horn that mechanic onto the non spellcasting classes. I liked that mechanic on the fighter, not sure if I want it used on multiple classes.



ED warps the process. The chacters become too reliable again

Max at level 1 become +10.5 with is a 15%
Max at level 10 is about +19 which is 60%

Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds. Constitution Based Class for Next!

All of issues with Bounded Accuracy on skill checks would be solved by using the whole ability score rather than the modifier.  Make a hard check something like DC 30, where someone with strength 20 could succeed on a 10 and someone with strength 9 could never possibly succeed.

No, it just shifts the numbers upwards.  You still have to deal with skill points and specialization on some level.

For ability score checks, use the whole ability score, with DC range 10-35.  For attack rolls and skill checks, where the majority contribution should be coming from other sources, use your ability modifier and fix the contribution from other sources so that your total modifier stays in the +2 to +20 range.

The important thing is that the range of bonuses and DC stays comparable to the variation in the dice roll, so that very qualified people can succeed with reasonable chance in situations where unqualified people cannot.

The metagame is not the game.

All of issues with Bounded Accuracy on skill checks would be solved by using the whole ability score rather than the modifier.  Make a hard check something like DC 30, where someone with strength 20 could succeed on a 10 and someone with strength 9 could never possibly succeed.

No, it just shifts the numbers upwards.  You still have to deal with skill points and specialization on some level.

For ability score checks, use the whole ability score, with DC range 10-35.  For attack rolls and skill checks, where the majority contribution should be coming from other sources, use your ability modifier and fix the contribution from other sources so that your total modifier stays in the +2 to +20 range.

The important thing is that the range of bonuses and DC stays comparable to the variation in the dice roll, so that very qualified people can succeed with reasonable chance in situations where unqualified people cannot.




Well, you would have to explain to me the skill progression mechanic to me.  I think I got it; however, it'll be better with clarification from you.
There's nothing wrong with the idea behind bounded accuracy.

The skill system just isn't implemented properly yet.  They haven't done the sort of tuning that you're requiring for it to not be a dealbreaker.  Which isn't surprising, since a lot of things are dealbreakers if they aren't finished being developed.



I will trust your word of the "skill system isn't done yet"...because the skill bonus/DC as you level up is what i am extremly pesimist about it...


It's also incredibly simple to change if things aren't working out.  The skill DC tiers are separated by 2 now, right?  Well, if that's too narrow given the available bonuses, separate them by 3.  Tadaa, scaling adjusted.  While giving guidance that you can choose numbers between, of course.

The point is that the gripe is about one of the minor balance number-change tweaks that are completely and wholly irrelevant at this stage.  That part comes later.
D&D Next = D&D: Quantum Edition
There's nothing wrong with the idea behind bounded accuracy.

The skill system just isn't implemented properly yet.  They haven't done the sort of tuning that you're requiring for it to not be a dealbreaker.  Which isn't surprising, since a lot of things are dealbreakers if they aren't finished being developed.



I will trust your word of the "skill system isn't done yet"...because the skill bonus/DC as you level up is what i am extremly pesimist about it...


It's also incredibly simple to change if things aren't working out.  The skill DC tiers are separated by 2 now, right?  Well, if that's too narrow given the available bonuses, separate them by 3.  Tadaa, scaling adjusted.  While giving guidance that you can choose numbers between, of course.

The point is that the gripe is about one of the minor balance number-change tweaks that are completely and wholly irrelevant at this stage.  That part comes later.



You need to increase them numbers alot more...we are talking about each number just being a 5% less chance of failing when we are rolling d20s...that's the problem...d freaking 20!

Hell, even the very controvertial +3 strength gloves would only give you at most 10% less chance to fail....
Or you compact the bonuses a lot more.  The point is the scale isn't even relevant to the testing at this point, and focusing on it and making judgments about systems because they're not properly tuned is counterproductive.
D&D Next = D&D: Quantum Edition
Or you compact the bonuses a lot more.  The point is the scale isn't even relevant to the testing at this point, and focusing on it and making judgments about systems because they're not properly tuned is counterproductive.



Compact them would only increase the problem...the only way i can see skills checks tow work with bounded accuracy is if we rolled 1d6 or 1d8 instead of a 1d20.  Like i have said before....as of now, a fighter with the same stregth at level of level 1 and one of level 10 would struggle almost as much to do the same task...having +1 or +2 more than the level 1 is insignificant over the course of 10 levels when it only means a decrease on failure chance by 5% and 10%

Rolling a 1d6 would make each increase on the check 16.666% less chance to fail and on a 1d8 it would 12.50% This is why bounded accuracy works on other system that use small dices for those kind of things...

I know it's silly but I don't like the DM guidance on the topic either. If Nearly Impossible (DC 25) really means "Tasks of this difficulty are so challenging that only demigods and their peers can succeed without assistance," then why is it so easy to make that roll?


Even raising the DCs by 3 and making Nearly Impossible a DC of 28 means a lowish level character still can succeed at the skill without assistance. Assuming a 16 to the stat, we're lookingat a +8 to the skill by lvl 4. By lvl 10 we've got a poor chance to succeed but the whole point of "Nearly Impossible" is only demigods and their peers can succeed without assistance.


Really, that means the hardest checks should require either a spell, maneuver or a magic item to pull off. I'm thinking we should up the DC increments by as much as 5 to reflect all this. Perhaps more.

Orzel that would bump the upper limit to DC 28? I would round that up to 30. 30 or 35 should be the pper limit IMHO as a 10th level whatever should be able to do something a levle 1 character in normal circumstance could not dd regardless of how good they rolled.

No prolems if magic items or mundane means grants a large bonus to a skill check (+10 bonus on knowledge in the Great Library, +10 craft:ship in shipyard etc).




DC 28 is Perfect

Max at level 1 is +8 with is a 5%
Max at level 10 is +12 whish is 25%




 DC28 still gets wrecked by expertise dice and they seem determined to shoe horn that mechanic onto the non spellcasting classes. I liked that mechanic on the fighter, not sure if I want it used on multiple classes.



ED warps the process. The chacters become too reliable again

Max at level 1 become +10.5 with is a 15%
Max at level 10 is about +19 which is 60%



Crunching the math, the actual percentage is closer to 79.875% for a skill mastery (ED to skill checks; assuming a 10th level character with 3d10 ED).  Averages are great for a certain type of picture but aren't accurate for this particular picture.
Actually I've been trying to think of why Bounded Accuracy itches me the wrong way, and I'm guessing it's because of this: by reducing what players get at each level, and by narrowing down both DCs and scaling, you effectively lower the relevance of levels.  Yes a level 10 PC has tons more abilities, HP and slightly greater accuracy, but in the grand scale of things, Bounded Accuracy makes everything... well, not so grand.  Because by level 10, you're basically something like a level 5 PC in every other edition prior to Next, but with more HP.

It's not that Bounded Accuracy can't work, it's just that while Bounded Accuracy is supposed to be slowing the scaling, the fact of the matter is that Bounded Accuracy both flattened and hyperinflated the scaling.


  • It's flatter because the checks don't actually change; in an effort to keep low level monsters relevant, all creatures beyond low level are rendered irrelevant (see: level 1 PCs fighting level 6 creatures in a system where the maximum core level is 10, and compare level 4 PC capabilities vs. level 10+ monster capabilities).

  • It's hyperinflated because in previous editions, assuming the math allowed you to be *roughly* 50% successful in any endeavor appropriate to your level, a task that's supposed to be reserved for demigods and their peers are reserved for demigods and their peers (let's see if a level 1 PC in 0E-4E pull off a DC that's meant for a level 20 PC).  In D&D Next, because the math is flattened (not made shallower), each point of increase is so relevant that the predetermined static DCs are quickly rendered useless


The way I see it, the developers currently do not know enough of their own system to make the "right" calls on significant aspects of their system, namely:


  • skill DCs relative to PC capabilities as their levels increase

  • monster stats, again relative to PC capabilities as their levels increase (because seriously the high level monsters are either always too strong -- from forcing the DMs to place hyperinflated stats to keep them relevant at high levels compared to the saving throws they must make relative to PC capabilities -- or easily rendered too weak as seen in the playtests


I'd actually have to agree with kadim on this, but with a different angle to it: if PCs are meant to be demigods or peers to the demigods by level 10, then have the DCs and PC capabilities scaled appropriately.  If PCs are meant to be "mundane" all the way to level 10, then the DCs should certainly be raised so high that not even natural 20s with max stats would reach it, and instead the PCs ought to find alternative ways to make the seemingly impossible, possible.

The crux of my issue with Bounded Accuracy is this: Bounded Accuracy should limit scaling on the universal level, not render everything except PCs on a flat scale.

And again my main problem with current D&D Next design, which includes Bounded Accuracy, is that PC levels are inconsistently both irrelevant and too relevant; irrelevant because of the existence of dead levels and static DCs, too relevant because when the bumps do come in, compared to the rest of the "system" they're like tsunamis of capabilities. 
Show

You are Red/Blue!
Take The Magic Dual Colour Test - Beta today!
Created with Rum and Monkey's Personality Test Generator.

You are both rational and emotional. You value creation and discovery, and feel strongly about what you create. At best, you're innovative and intuitive. At worst, you're scattered and unpredictable.

D&D Home Page - What Monster Are You? - D&D Compendium

57047238 wrote:
If you're crossing the street and see a city bus barreling straight toward you with 'GIVE ME YOUR WALLET!' painted across its windshield, you probably won't be reaching for your wallet.
I Don't Always Play Strikers...But When I Do, I Prefer Vampire Stay Thirsty, My Friends
This is what I believe is the spirit of D&D 4E, and my deal breaker for D&D Next: equal opportunities, with distinct specializations, in areas where conflict happens the most often, without having to worry about heavy micromanagement or system mastery. What I hope to be my most useful contributions to the D&D Community: DM Idea: Collaborative Mapping, Classless 4E (homebrew system, that hopefully helps in D&D Next development), Gamma World 7E random character generator (by yours truly), and the Concept of Perfect Imbalance (for D&D Next and other TRPGs in development) Pre-3E D&D should be recognized for what they were: simulation wargames where people could tell stories with The Best Answer to "Why 4E?" Fun vs. Engaging