I *LIKE* skill dice!

139 posts / 0 new
Last post
I know that this is a playtest, and thus they are looking for feedback.   I read many threads from the people who dislike skill dice. The main complaint seems to be the feeling that, since all skills develop at the same rate, nobody is a specialist, and instead everybody is a generalist in their set of skills.   People miss having their +10 or +15 in a skill.

I disagree with them, and my player's rogue last night proved the point.

He had a skill focus in Sneak (minimum d20 roll is a 10) which proved to be FAR more useful and thematic.   He boldly knew that he would never roll less than a 10, and thus his minimum after mods was going to be a 14.  His expertise allowed him to make certain decisions in a far more meaningful way than having an extra +5 in a swingier roll ever would have.

It felt RIGHT.   It was more FUN.   And mechanically, it kept everybody within the same RANGE (i.e. I didn't have to somehow come up with reasons why low-level baddies were using mastercraft locks, just to give the skill expert a challenge).  I don't need anything more than that in a system, thanks!

Indeed, he could immediately see the point in pursuing the expert's path, and choosing various rogue-based skills to emphasize in a similar way.

And this is why I am hoping that others who see the value in the entire skill-die mechanic, and the interesting-and-useful ways it can be mingled with other mechanics, voice their opinions here.  

Let's give some positive feedback, so the developers know that it isn't all  doom-and-gloom.   I, personally, can do without huge modifiers, and I appreciate a more elegant system like the skill dice.

Speak up if you're enjoying them, just to make sure we can keep 'em! 
I LOVE the skill dice. I do hope, however, that an advanced rules module offers a way to train up your die size in different skills at different rates.
For the skill die to work as described above you NEED to play with feats as well. I have no problem with that as i always use both, just pointing out that it doesnt work as a stand alone system by current rules.
I think that many people overlook the combination of Skill Focus and Skill Supremacy, and how it really does make a character feel like an unfailing expert -- all the while keeping the DC's of tasks within a bounded-accuracy style system.

I like that the DC's never have to get too crazy, just to keep up with the experts. 
When I first saw skill dice I didn't really like them, but I'm warming up to the idea.

I like the concept of them.  Not had a chance to use them, though, so I have no comment there.
Agree with OP.
I LOVE the skill dice. I do hope, however, that an advanced rules module offers the rate to train up your die size in different skills at different rates. 



Same.
I do not like them. They've slowed my game down. They don't feel intuitive and they give my players more math to do.
I agree completely with the OP.  Skill dice are great! 

Sure, the player has to add the two rolls and the character's appropriate modifier, but he or she doesn't have to reference the character sheet to determine how much of a bonus that particular skill gets.  It works very well with bounded accuracy, where a higher level character only gets a better chance to succeed, and can still fail against the same challenges it faced at first level (assuming the Skill Focus feat wasn't added after 1st).  The skill die also somewhat mitigates the likelihood that a bad d20 roll will cause a character fail a low DC challenge, because it gives another die the player could roll a high score on (though at 1st level, that is not going to be very high).
I love them, too! 

Danny

They work well. I like them too.

A Brave Knight of WTF

I love them, too! 

I am not a fan.

What if, in our game, an apocalyptic event occurs, and the players are now forced to skulk in the shadows to avoid being found and eaten by the Ilithids that now rule the land.  

They are level 12 when this happens, and 4 levels later, they should probably have some experience sneaking around right?  Not in D&D Next.   

In D&D Next, they already picked their skills.... they already took their feats... they can no longer adapt to anything.  

Here's another scenario, I have a lvl 20 NPC known throughout the land as the "Cat Burglar"  he is a master thief.   
He has the Thief  Rogue Scheme, And even took Skill Supremacy with Sneak, and lets say he has a 20 dex.  He will be rolling 2d20 (Min 10, picking the highest) +5 and adding his skill die of 2d12 (picking the highest).  This is an absolute best case scenario.

That puts his range from 16 to 37.  Since he is rolling twice and choosing the highest result, lets put his average within the top 25% of that range, so something like a  31.  Pretty damn good right?

Well, if you compare that to a rogue with the same build and scheme... at level 3, with, lets say an 18 dex, they will be rolling 2d20(min 10 pick highest) +4 and 2d6 (pick highest).  Using the same math their average score will be something like 27.  

17 levels and that  is the separation.  A 1-7 point difference.  Statistically, the master thief at level 20 will be, on average, about 20% better than another rogue he outlevels by  666%

This is a failure to scale things appropriately, and this is giving skill dice the best possible scenario of direct comparison.  Take two trained characters without feats at the same levels and you will have the same range of difference, but the chance for the level 3 character to win in direct ompetition goes up considerably, since the total ranges will be more like 7-37 versus 6-30, with average scores landing in the middle of that range you have 22 versus 18, the same 20% advantage, but far more reliance on the random, which allows the lvl 3 character to win about 30% of the time.

Skill dice, ultimately, are very strange to me.

I don't understand why static, scaling bonuses aren't used instead.  Say they just gave you half of your total level (Rounded up) to a trained skill check.  Change skill mastery to be another static bonus that increases your "trained skill bonus" by half  again.

Using the same two rogues, we would have  the lvl 20 rogue rolling in the 30-40 range, and the lvl 3 rogue rolling in the 17-27 range.  Now you have a more appropriate comparison for two characters so distant in level and experience. 


In D&D Next, they already picked their skills.... they already took their feats... they can no longer adapt to anything.  



See the retraining rules; they can adapt. 


17 levels and that  is the separation.  A 1-7 point difference.  Statistically, the master thief at level 20 will be, on average, about 20% better than another rogue he outlevels by  666%



Bounded accuray. Being 666% higher in levels does not equal being 666% better in capability. You are operating under a paradigm that does not actually match the context of the game. 


I don't understand why static, scaling bonuses aren't used instead.  Say they just gave you half of your total level (Rounded up) to a trained skill check.  Change skill mastery to be another static bonus that increases your "trained skill bonus" by half  again.



They use the skill die because it does wonders for the math of a bounded system. The system you are describing would break the math of a bounded system. 





In D&D Next, they already picked their skills.... they already took their feats... they can no longer adapt to anything.  



See the retraining rules; they can adapt. 


17 levels and that  is the separation.  A 1-7 point difference.  Statistically, the master thief at level 20 will be, on average, about 20% better than another rogue he outlevels by  666%



Bounded accuray. Being 666% higher in levels does not equal being 666% better in capability. You are operating under a paradigm that does not actually match the context of the game. 


I don't understand why static, scaling bonuses aren't used instead.  Say they just gave you half of your total level (Rounded up) to a trained skill check.  Change skill mastery to be another static bonus that increases your "trained skill bonus" by half  again.



They use the skill die because it does wonders for the math of a bounded system. The system you are describing would break the math of a bounded system. 






1) Yes, you can retrain, and yes I kow the rules, but it is a restrictive methodology.  The player is required to spend a feat to gain access to a skill rather than trade out a skill.  Either way, the instant loss of a skill or ability that a character has used for 12 levels disrupts verisimilitude, and in terms of simulation, the new system does not fall in line with the way that skills are natural learned over time.  

Essentially D&D Next is saying that a 40 year old could learn spanish, but he would have to forget calculus first.  It is not intuitive.

2)  You are correct in suggesting that a 2nd level character is not 100% better than a 1st level character, but I never said anything of the like.  I am talking about progression.  Two schools of thought collide here.  I am of the belief that a level 20 rogue specializing in stealth should be unrivalled by a level 3 rogue, and that is not the case.  Power progression in terms of martial combat, spellcasting, health, etc.  all scale in ways that meet this criteria.  A level 20 fighter could never be rivalled in terms of combat prowess by a level 3 fighter, so why is it the opposite with skill progression (or lack thereof)?

3)  This is inherent to the problem I am addressing.  If DCs never really scale, and skill checks never really scale (or to be more accurate scale very little)  then you can avoid characters being blocked off from performing certain tasks that they aren't trained for.  I never saw this as an issue, and always considered it a piece of the game balance equation.  The way this mechanic is built, however, makes skills a tertiary aspect of the game rather than a robust system.  They are sidelined.  They progress at a snail's pace and reduce the impact of characters who might want  to specialize in non-combat skill usage.  Yes those characters will be better, but they will never truly reach a place of supremacy.  There is every possibility that an untrained lvl 1 character can beat a lvl 20 trained character in a test of skill vs. skill, and this does not feel right to me.

Anyway, everyone can feel about this however they want to feel, but  those were just my two cents.  I will be interested to see what the next pass over skills will do, and I hope that these issues I have about scaling are addressed. 
I do not like them. They've slowed my game down. They don't feel intuitive and they give my players more math to do.



How do they give your players more math to do? Before skill dice you added three numbers: your d20 roll, your ability modifier and your skill bonus. Now you still add three numbers: your d20 roll, your ability modifier and your skill die roll.


1) Yes, you can retrain, and yes I kow the rules, but it is a restrictive methodology.  The player is required to spend a feat to gain access to a skill rather than trade out a skill.  Either way, the instant loss of a skill or ability that a character has used for 12 levels disrupts verisimilitude, and in terms of simulation, the new system does not fall in line with the way that skills are natural learned over time.  

Essentially D&D Next is saying that a 40 year old could learn spanish, but he would have to forget calculus first.  It is not intuitive.

2)  You are correct in suggesting that a 2nd level character is not 100% better than a 1st level character, but I never said anything of the like.  I am talking about progression.  Two schools of thought collide here.  I am of the belief that a level 20 rogue specializing in stealth should be unrivalled by a level 3 rogue, and that is not the case.  Power progression in terms of martial combat, spellcasting, health, etc.  all scale in ways that meet this criteria.  A level 20 fighter could never be rivalled in terms of combat prowess by a level 3 fighter, so why is it the opposite with skill progression (or lack thereof)?

3)  This is inherent to the problem I am addressing.  If DCs never really scale, and skill checks never really scale (or to be more accurate scale very little)  then you can avoid characters being blocked off from performing certain tasks that they aren't trained for.  I never saw this as an issue, and always considered it a piece of the game balance equation.  The way this mechanic is built, however, makes skills a tertiary aspect of the game rather than a robust system.  They are sidelined.  They progress at a snail's pace and reduce the impact of characters who might want  to specialize in non-combat skill usage.  Yes those characters will be better, but they will never truly reach a place of supremacy.  There is every possibility that an untrained lvl 1 character can beat a lvl 20 trained character in a test of skill vs. skill, and this does not feel right to me.

Anyway, everyone can feel about this however they want to feel, but  those were just my two cents.  I will be interested to see what the next pass over skills will do, and I hope that these issues I have about scaling are addressed. 




Ok, so, I have responses to everything, and I don't agree with you on any single point. If you want to hear the reasons I will post them. But, I am not going to do so out of hand, because honestly we are talking about subjective appreciation of particular minute details. It is perfectly fair if you and me disagree. That is life. And, like you said, we each have our two cents. I mean, I definitely hope the vast majority of players side with me and really like skill dice, because now that they are here I think losing them might be a deal breaker (I would have to experience the event to know for sure). But, the truth is, for all the objective reason why I like them, it is subjective value that makes me like them. I am just of a different school of though, as you say. 



Ok, so, I have responses to everything, and I don't agree with you on any single point. If you want to hear the reasons I will post them. But, I am not going to do so out of hand, because honestly we are talking about subjective appreciation of particular minute details. It is perfectly fair if you and me disagree. That is life. And, like you said, we each have our two cents. I mean, I definitely hope the vast majority of players side with me and really like skill dice, because now that they are here I think losing them might be a deal breaker (I would have to experience the event to know for sure). But, the truth is, for all the objective reason why I like them, it is subjective value that makes me like them. I am just of a different school of though, as you say. 




That is fair.  The whole thing is very subjective.  It is an opinion versus opinion discussion.  I would like to note, that I am not against the idea of skill dice, nor do I think that any previous system did a very good job of handling skills.  I think that they could work, but that they need retooling.  

For instance, I would love to see a parsed system, perhaps with ranks of proficiency.  Rather than automatically scalling the skill die for 4 skills, give players 4 skills to start (at a 1d4 skill die) and let them choose how to spread out the rest.  Each time they gained a skill, they could increase the die type for a skill they already have (iteratively: 1d4, 1d6, 1d8, 1d10, 1d12), or perhaps pick a new skill (which would start at 1d4).  This way, using my previous example of the wizard wanting to learn stealth, he could.  His training in another skill would be suspended, but otherwise he would be totally free to pick a new trained skill (albeit at only a 1d4, but you have to start somewhere).

I think this might make for a more flexible system... rather than having 4 maxed skills, I could see many players choosing to split ranks a few times.  Obviously this would muddy up any non-skill usages of skill dice, but its really just a wuick thought, not a proposal.

Anyway, thank you for your discretion and your honesty.  I hope things work out for the both of us.
 
While as a player I like the "feel" of skill dice, I also think they create a bit of a problem with "swingyness", as it makes skills even MORE random than they have been in other edditions (and it was a potential weakness then).

For instance, I would love to see a parsed system, perhaps with ranks of proficiency.  Rather than automatically scalling the skill die for 4 skills, give players 4 skills to start (at a 1d4 skill die) and let them choose how to spread out the rest.  Each time they gained a skill, they could increase the die type for a skill they already have (iteratively: 1d4, 1d6, 1d8, 1d10, 1d12), or perhaps pick a new skill (which would start at 1d4).  This way, using my previous example of the wizard wanting to learn stealth, he could.  His training in another skill would be suspended, but otherwise he would be totally free to pick a new trained skill (albeit at only a 1d4, but you have to start somewhere).
 



I like the idea of skills leveling (increacing in dice steps) as a character levels, and I'm more of a fan of skill dice than flat bonuses, because the contextual element that a random die result has. It  makes sense that a rogue be able to have his sneak and disable device skills at a level 3 (1d8) versus his bluff which might only be at a level 1 (1d4). I figure if every second level a character gained they could "gain a level" in their skills, up to a level 5 skill (1d12). This coupled with feats like skill focus and supremacy gives skill users a lot of versatility in their choices.

In a 20-level game that's a total of ten skill increases for any one character, and maybe the rogues could have a feature where every time they get to level up skills as a rogue they get to level up 2 skills instead of 1.

Just a thought or two.
Toronto Dungeon Master

While as a player I like the "feel" of skill dice, I also think they create a bit of a problem with "swingyness", as it makes skills even MORE random than they have been in other edditions (and it was a potential weakness then).


That is not really true. Statistically speaking, your mean roll is about what you would have with a +3 to +7 flat bonus spread. In fact, you usually have a slightly higher chance (by a few percent) of success than you would with a flat bonus. And, your chance of rolling any specific number in the middle of the bell curve is still 5% (just as it would be with a d20+flat modifier roll). That is not more "swingy," it is less. The only thing that is more "swingy" is your chance to roll a low/high number. You will sometimes hit numbers that you could not hit before (both in the low and high end of the scale). The thing is, that does not change your percentile chance of success when it comes to DCs in the middle of the bell curve. That middle range is no more “swingy” than it was before (and it is a little less). Meanwhile, in a bounded game, being able to still hit the lower and higher ends of the scale without drastically changing (and in fact minutely increasing) your chance of success when it comes to DCs in the middle of the bell curve is VERY beneficial to the game. In fact, that is one of the few ways you can keep the skill system bounded while still allowing a statistical change in the mean success rate that about equals what you would get with a +3 to +7 flat bonus.

Sorry, but this "it is more swingy" claim is starting to become one of my pet peeves. I don't mind people disliking the system for aesthetic reasons. But, propagating a disingenuous claim that is fairly bankrupt when it comes to the actual mathematical modeling of the system really bothers me. 


 

Personally I just preffer my math being mostly on my dice rather than mostly on my character sheet.
Toronto Dungeon Master
What Cyber-Dave said: a d20 roll + stuff is very swingy. Equal chance of rolling any possible result. However, a d20 roll + d6 roll + stuff averages toward the middle and tapers off at either extreme. Higher chance of getting a middle result, lower chance of hitting either extreme.

This does the job of saying "My character can do this above-average task with exceptional consistency" while still allowing the slight chance of extreme failure or extreme success.
d20+d6 gives you an equal 5% chance across 7~21

So you have an equal 5% chance of rolling a 9 as you do a 13 or 17, etc

It goes down to
6 or 22 - 4.17%
5 or 23 - 3.33%
4 or 24 - 2.5%
3 or 25 - 1.67%
2 or 26 - 0.83%

The weird thing is that the dice size pushes that flat 5% higher, and makes it tighter but makes the possible combinations wider
d4 = 5~21
d6 = 7~21
d8 = 9~21
d10 = 11~21
d12 = 13~21

This roughly translates into the better you get, the less consistent you become

d4 = 5~21 (2~24) Distance from Min/Max = 3 (7.5% of not being average)
d6 = 7~21 (2~26) Distance from Min/Max = 5 (12.5% of not being average)
d8 = 9~21 (2~28) Distance from Min/Max = 7 (17.5% of not being average)
d10 = 11~21 (2~30) Distance from Min/Max = 9 (22.5% of not being average)
d12 = 13~21 (2~32) Distance from Min/Max = 11 (27.5% of not being average)
d20+d6 gives you an equal 5% chance across 7~21

So you have an equal 5% chance of rolling a 9 as you do a 13 or 17, etc

It goes down to
6 or 22 - 4.17%
5 or 23 - 3.33%
4 or 24 - 2.5%
3 or 25 - 1.67%
2 or 26 - 0.83%

The weird thing is that the dice size pushes that flat 5% higher, and makes it tighter but makes the possible combinations wider
d4 = 5~21
d6 = 7~21
d8 = 9~21
d10 = 11~21
d12 = 13~21

This roughly translates into the better you get, the less consistent you become

d4 = 5~21 (2~24) Distance from Min/Max = 3 (7.5% of not being average)
d6 = 7~21 (2~26) Distance from Min/Max = 5 (12.5% of not being average)
d8 = 9~21 (2~28) Distance from Min/Max = 7 (17.5% of not being average)
d10 = 11~21 (2~30) Distance from Min/Max = 9 (22.5% of not being average)
d12 = 13~21 (2~32) Distance from Min/Max = 11 (27.5% of not being average)



Whatcha got there? Numbers?

Thanks for the math.  It actually struck me as strange that they didn't go for more combined dice to mitigate this effect.  With a max of 3d4 rather than 1d12, you have the same max, a higher min and a much more consistent check system at high level.  My problem with the system is still more about character customization and flexibility though.  Perhaps there will be a better setup in advanced.  I know that a lot of people would just take maxed skills in 3.5 (Class Skill Points + Int Mod = number of maxed skills at any given level) but I would prefer to have choices.  Broader training with less expertise or narrow training with more expertise.  Currently the system only allows for the latter.
I know that a lot of people would just take maxed skills in 3.5 (Class Skill Points + Int Mod = number of maxed skills at any given level) but I would prefer to have choices.  Broader training with less expertise or narrow training with more expertise.  Currently the system only allows for the latter.


Since most of my constructive posts get buried by rehashed arguments, I like to take every opportunity to toot my own horn and provide my own suggestions for advanced rules modules.
I am not a fan.

What if, in our game, an apocalyptic event occurs, and the players are now forced to skulk in the shadows to avoid being found and eaten by the Ilithids that now rule the land.  

They are level 12 when this happens, and 4 levels later, they should probably have some experience sneaking around right?  Not in D&D Next.   

In D&D Next, they already picked their skills.... they already took their feats... they can no longer adapt to anything.



This isn't a problem with skill die.  It's a problem with skill training.  And really, it's a problem because they haven't finished with feats and levels above 10 get no additional chance to train skills.  If in the finished game you get to train a new feat every two levels (or whatever), then your apocalypse sneakers will have plenty of chances to train their newly appropriate skills. 


Here's another scenario, I have a lvl 20 NPC known throughout the land as the "Cat Burglar"  he is a master thief.   
He has the Thief  Rogue Scheme, And even took Skill Supremacy with Sneak, and lets say he has a 20 dex.  He will be rolling 2d20 (Min 10, picking the highest) +5 and adding his skill die of 2d12 (picking the highest).  This is an absolute best case scenario.

That puts his range from 16 to 37.  Since he is rolling twice and choosing the highest result, lets put his average within the top 25% of that range, so something like a  31.  Pretty damn good right?

Well, if you compare that to a rogue with the same build and scheme... at level 3, with, lets say an 18 dex, they will be rolling 2d20(min 10 pick highest) +4 and 2d6 (pick highest).  Using the same math their average score will be something like 27.  

17 levels and that  is the separation.  A 1-7 point difference.  Statistically, the master thief at level 20 will be, on average, about 20% better than another rogue he outlevels by  666%

This is a failure to scale things appropriately, and this is giving skill dice the best possible scenario of direct comparison.  Take two trained characters without feats at the same levels and you will have the same range of difference, but the chance for the level 3 character to win in direct ompetition goes up considerably, since the total ranges will be more like 7-37 versus 6-30, with average scores landing in the middle of that range you have 22 versus 18, the same 20% advantage, but far more reliance on the random, which allows the lvl 3 character to win about 30% of the time.



This isn't really a problem with skill die either.  If it's a problem, its with the ability cap.  In your example, you picked your level 3 guy at a point that was already 90% of the way there by giving him 18 dex. The bigger skill die are giving an average of +2.5 to his checks, which in a world where a bonus of 5 is the difference between a Formidable DC and a Nearly Impossible DC is actually a big deal.

As has been pointed out, giving a static bonus that scales with level simply erases all the benefits of a bounded accuracy system.  Trained characters would be so good at what they do, for them to literally be challeneged at all by a task would mean that that same task is absolutely impossible for non-trained characters to complete. 

I know that a lot of people would just take maxed skills in 3.5 (Class Skill Points + Int Mod = number of maxed skills at any given level) but I would prefer to have choices.  Broader training with less expertise or narrow training with more expertise.  Currently the system only allows for the latter.


Since most of my constructive posts get buried by rehashed arguments, I like to take every opportunity to toot my own horn and provide my own suggestions for advanced rules modules.



Huh, I suggested something very similar in 2 other threads.  Yours are far better written, mine were just offhanded.  I also think that breaking things down into a point-buy system is handy.  I hadn't really considered the need to further tax the first rank based on average bonus, but that does make sense.  I also have said the same about bonus racial/class skills just being kept at max.  These changes seem totally reasonable and realistic to me.  Lets hope they get implemented.  Then I can just complain about bounded accuracy making skills feel cheap rather than my primary concern about railroaded progression. =) 
I am not a fan.

What if, in our game, an apocalyptic event occurs, and the players are now forced to skulk in the shadows to avoid being found and eaten by the Ilithids that now rule the land.  

They are level 12 when this happens, and 4 levels later, they should probably have some experience sneaking around right?  Not in D&D Next.   

In D&D Next, they already picked their skills.... they already took their feats... they can no longer adapt to anything.



This isn't a problem with skill die.  It's a problem with skill training.  And really, it's a problem because they haven't finished with feats and levels above 10 get no additional chance to train skills.  If in the finished game you get to train a new feat every two levels (or whatever), then your apocalypse sneakers will have plenty of chances to train their newly appropriate skills. 


Here's another scenario, I have a lvl 20 NPC known throughout the land as the "Cat Burglar"  he is a master thief.   
He has the Thief  Rogue Scheme, And even took Skill Supremacy with Sneak, and lets say he has a 20 dex.  He will be rolling 2d20 (Min 10, picking the highest) +5 and adding his skill die of 2d12 (picking the highest).  This is an absolute best case scenario.

That puts his range from 16 to 37.  Since he is rolling twice and choosing the highest result, lets put his average within the top 25% of that range, so something like a  31.  Pretty damn good right?

Well, if you compare that to a rogue with the same build and scheme... at level 3, with, lets say an 18 dex, they will be rolling 2d20(min 10 pick highest) +4 and 2d6 (pick highest).  Using the same math their average score will be something like 27.  

17 levels and that  is the separation.  A 1-7 point difference.  Statistically, the master thief at level 20 will be, on average, about 20% better than another rogue he outlevels by  666%

This is a failure to scale things appropriately, and this is giving skill dice the best possible scenario of direct comparison.  Take two trained characters without feats at the same levels and you will have the same range of difference, but the chance for the level 3 character to win in direct ompetition goes up considerably, since the total ranges will be more like 7-37 versus 6-30, with average scores landing in the middle of that range you have 22 versus 18, the same 20% advantage, but far more reliance on the random, which allows the lvl 3 character to win about 30% of the time.



This isn't really a problem with skill die either.  If it's a problem, its with the ability cap.  In your example, you picked your level 3 guy at a point that was already 90% of the way there by giving him 18 dex. The bigger skill die are giving an average of +2.5 to his checks, which in a world where a bonus of 5 is the difference between a Formidable DC and a Nearly Impossible DC is actually a big deal.

As has been pointed out, giving a static bonus that scales with level simply erases all the benefits of a bounded accuracy system.  Trained characters would be so good at what they do, for them to literally be challeneged at all by a task would mean that that same task is absolutely impossible for non-trained characters to complete. 




Yeah... the benefits of bounded accuracy.  Still looking for those.  And yes, trained characters would be so good that non-trained characters wouldn't be able to compete.  Its why I don't play quarterback in the NFL...its why I am not a Pro Golfer.  That is the point.  That is the problem.

/shrug

As a DM that has had to design skill challenges that actually challenge a party and don't devolve into "Put your dice away.  He wins, the rest of you lose.", I love bounded accuracy. 
I propose a gradually increasing potency to skill dice as an added incentive to levelling up. (I think) this works with the bonded accuracy, of which I am a fan.

Skill Dice Level Chart
Every skill has an attributed “Skill Level”, which determines the type of die that is associated with that skill’s use.
Level 1 - d4
Level 2 - d6
Level 3 - d8
Level 4 - d10
Level 5 - d12

Skill dice levels can increase every second level (levels 2, 4, 6, 8, 10, 12, 14, 16, 18, 20), when a character picks one skill to step up by one level. These skill levels, in addition to the opportunity to improve skill use with feats, should make the feeling of developing frequently used skills and talents come to the surface of the player’s experience.

Rogue Skill Development
Rogues, being the class that focuses on skills more than any other, gains an additional skill level up at levels 6, 12 and 18.

These are just preliminary ideas and I'm sure that (not being a math guy) there are numbers problems, but this would be the kind of improvement in character skill that I would have fun playing with. I'm sure there are tweaks that would appease you number-oriented folks (Level 4 skill dice = 2d6 rather than 1d10 maybe) but I came up with this ten minutes ago.
Toronto Dungeon Master
d20+d6 gives you an equal 5% chance across 7~21

So you have an equal 5% chance of rolling a 9 as you do a 13 or 17, etc

It goes down to
6 or 22 - 4.17%
5 or 23 - 3.33%
4 or 24 - 2.5%
3 or 25 - 1.67%
2 or 26 - 0.83%

The weird thing is that the dice size pushes that flat 5% higher, and makes it tighter but makes the possible combinations wider
d4 = 5~21
d6 = 7~21
d8 = 9~21
d10 = 11~21
d12 = 13~21

This roughly translates into the better you get, the less consistent you become

d4 = 5~21 (2~24) Distance from Min/Max = 3 (7.5% of not being average)
d6 = 7~21 (2~26) Distance from Min/Max = 5 (12.5% of not being average)
d8 = 9~21 (2~28) Distance from Min/Max = 7 (17.5% of not being average)
d10 = 11~21 (2~30) Distance from Min/Max = 9 (22.5% of not being average)
d12 = 13~21 (2~32) Distance from Min/Max = 11 (27.5% of not being average)




Yea, that is not exactly how it works. You are ignoring the way the percentile values work in a binary pass/fail system. The higher the skill die, the larger your chance of success against any particular DC in all cases. Your success rate does not become more inconsistent. I can post more math on this tomorrow if required, but I am a little too tired to do so right now. 

d20+d6 gives you an equal 5% chance across 7~21

So you have an equal 5% chance of rolling a 9 as you do a 13 or 17, etc

It goes down to
6 or 22 - 4.17%
5 or 23 - 3.33%
4 or 24 - 2.5%
3 or 25 - 1.67%
2 or 26 - 0.83%

The weird thing is that the dice size pushes that flat 5% higher, and makes it tighter but makes the possible combinations wider
d4 = 5~21
d6 = 7~21
d8 = 9~21
d10 = 11~21
d12 = 13~21

This roughly translates into the better you get, the less consistent you become

d4 = 5~21 (2~24) Distance from Min/Max = 3 (7.5% of not being average)
d6 = 7~21 (2~26) Distance from Min/Max = 5 (12.5% of not being average)
d8 = 9~21 (2~28) Distance from Min/Max = 7 (17.5% of not being average)
d10 = 11~21 (2~30) Distance from Min/Max = 9 (22.5% of not being average)
d12 = 13~21 (2~32) Distance from Min/Max = 11 (27.5% of not being average)




Yea, that is not exactly how it works. You are ignoring the way the percentile values work in a binary pass/fail system. The higher the skill die, the larger your chance of success against any particular DC in all cases. Your success rate does not become more inconsistent. I can post more math on this tomorrow if required, but I am a little too tired to do so right now. 




He never said that success rates became more inconsistent.  Only that the total score rolled became less likely to land in the average middle range.  Obviously, since the max keeps going up, your chances of success keep going up against any set DC.  You could say this abouot a lot of things.  If you only rolled a 1d4 for skill checks, your range would drop and you would get more consistent outcomes.  If you roll a d20 and a d12 then your range is larger and your totals become more inconsistent.  Consistency in terms of success or failure will continue to rise simply as chance of success rises (which it always will at the top end of the range advances and the minimum stays where it is)

He never said that success rates became more inconsistent.  Only that the total score rolled became less likely to land in the average middle range.  Obviously, since the max keeps going up, your chances of success keep going up against any set DC.  You could say this abouot a lot of things.  If you only rolled a 1d4 for skill checks, your range would drop and you would get more consistent outcomes.  If you roll a d20 and a d12 then your range is larger and your totals become more inconsistent.  Consistency in terms of success or failure will continue to rise simply as chance of success rises (which it always will at the top end of the range advances and the minimum stays where it is)



I realize that. It just doesn't really matter. It doesn’t actually make you less consistent at anything that matters. All that matters is your consistency when it comes to passing skill checks. That was my point. When you focus on consistency when it comes to rolling any single number in the fringes of your range you ignore the fact that the constancy when it comes to passing any given skill check only grows and your consistency in not rolling one particular low number also only grows. Meanwhile, while your standard deviation will increase a little, it won't increase drastically. The difference in standard deviation when you roll 1d20+1d4 and 1d20+1d12 is only 0.85. That isn't exactly a strikingly large number.  

When it comes to a bounded system, the skill die does nothing but good things mathematically speaking.  



. That is not more "swingy," it is less. The only thing that is more "swingy" is your chance to roll a low/high number. You will sometimes hit numbers that you could not hit before (both in the low and high end of the scale). The thing is, that does not change your percentile chance of success when it comes to DCs in the middle of the bell curve. That middle range is no more “swingy” than it was before (and it is a little less). Meanwhile, in a bounded game, being able to still hit the lower and higher ends of the scale without drastically changing (and in fact minutely increasing) your chance of success when it comes to DCs in the middle of the bell curve is VERY beneficial to the game. In fact, that is one of the few ways you can keep the skill system bounded while still allowing a statistical change in the mean success rate that about equals what you would get with a +3 to +7 flat bonus....



I wasn't being disengenuous, just bad at applied maths, as I was only thinking of the high and low end. Now that you have pointed out the flaw in my thinking, I happily renounce my initial position.

Cheers for showing me where I was going wrong Smile


. That is not more "swingy," it is less. The only thing that is more "swingy" is your chance to roll a low/high number. You will sometimes hit numbers that you could not hit before (both in the low and high end of the scale). The thing is, that does not change your percentile chance of success when it comes to DCs in the middle of the bell curve. That middle range is no more “swingy” than it was before (and it is a little less). Meanwhile, in a bounded game, being able to still hit the lower and higher ends of the scale without drastically changing (and in fact minutely increasing) your chance of success when it comes to DCs in the middle of the bell curve is VERY beneficial to the game. In fact, that is one of the few ways you can keep the skill system bounded while still allowing a statistical change in the mean success rate that about equals what you would get with a +3 to +7 flat bonus....



I wasn't being disengenuous, just bad at applied maths, as I was only thinking of the high and low end. Now that you have pointed out the flaw in my thinking, I happily renounce my initial position.

Cheers for showing me where I was going wrong


Actually, it depends on your interpretation or definition of 'more swingy', if you mean 'the standard deviation is greater', then you are completely correct, and it is 'more swingy'

Personally, I'd like to see characters have the ability to be better at some skills than they are at others, as well.  However, I don't want to deal with different die sizes for different skills.  I like a more binary option, such as the one that gives you a skill die if the character is trained in a skill, and does not give a skill die when untrained.  I think the feats Skill Focus and Skill Supremacy do that very well, though I would put the benefits of Skill Supremacy under Skill Focus and vice-versa, as I feel that Skill Focus is more powerful as they are now, since it gives a guaranteed 11 + modifier while Supremacy still allows for a 2 + modifier (thought that chance is only 1/400).

  That said, I also think characters should get feats more often, though that might require reducing the effect of many feats to compensate.  IIRC, this goes agaisnt the stated goals of the developers, but I felt that 3.X characters aquired them too slowly, and that the rate was just about right in 4E (though I always wanted more feats as a player).  However, with faster feat acquisition, players would be more likely to take feats like Superior Skill Training, Skill Focus, and Skill Supremacy, especially if they are felt to be on par with the other feats, which they don't really feel to me at present (see Polearm Training, esp.)

I wasn't being disengenuous, just bad at applied maths, as I was only thinking of the high and low end. Now that you have pointed out the flaw in my thinking, I happily renounce my initial position.

Cheers for showing me where I was going wrong

I did not mean to imply that you specifically were being disingenuous, only that the argument is disingenuous (in its sources and nature). And no problem.


Actually, it depends on your interpretation or definition of 'more swingy', if you mean 'the standard deviation is greater', then you are completely correct, and it is 'more swingy'





Here is a great example of why I call that argument disingenuous. On the surface, this argument is true. If you define "more swingy" as "having a larger standard deviation" then you can call the skill die system "more swingy." The standard deviation with a flat modifier is a flat 5.77. Meanwhile, with a skill die, the standard deviation starts at 5.87 and grows to 6.72. But, the difference we are talking about is not a drastic difference (0.95 at most). So, even when we focus on the standard deviation, the change is minuscule. The range of numbers you are statistically probable to roll is greater by less than a single number on either end; it is only 1.9 numbers in terms of the total range!  Moreover, focusing on that standard deviation serves only one purpose: obfuscating all the percentages that matter, such as your chances of success and failure. Those never become “more swingy.” In fact, in many cases the skill die makes them less “swingy” (as your percentile probability of success is actually a few points higher). The only time they are “more swingy” is when the number rolled in very high or very low—there is a higher chance that you will fail a very easy skill check or pass a very hard skill check (though in both cases the chances are very small) as you can roll numbers that you simply could not roll before. When you roll 1d20+1d12 you can still roll between 2-7 and 28-32 (which you could not with a flat +7 modifier). Again, though, that is EXACTLY what you want a bounded system to allow! 


So, claiming that they are de-facto “more swingy” due to something like their standard deviation only serves to hide the way they actually operate. In terms of the bounded system, mathematically speaking, the skill die is NOTHING but good for the game. In all the ways in which you do not want a bounded system to be “more swingy” the skill die is not “more swingy.”


The only real complaint against the skill die amounts to this: “I aesthetically don’t like bounded accuracy when it is applies to skill systems.” That, however, is not the same as “the skill die is more swingy.” What you are actually saying is “I don’t like ever passing very easy or very hard skill checks; I think all skill checks should be trapped into binary success/failures that exist in a middle range window; I think that characters should never be able to fail very easy skill checks or pass hard ones.” Now, I am not going to argue those claims with anyone. They are aesthetic preferences. I STRONGLY disagree with all of them. I think the bounded system is both far more realistic and far better for the game. But aesthetic preferences, ultimately, are aesthetic. Meanwhile, in every way that matters when it comes to a bounded system, namely the percentile probability of success when it comes to numbers in the middle range, the skill die is not more swingy as far as your success rates are concerned. 


He had a skill focus in Sneak (minimum d20 roll is a 10) which proved to be FAR more useful and thematic.   He boldly knew that he would never roll less than a 10, and thus his minimum after mods was going to be a 14.  His expertise allowed him to make certain decisions in a far more meaningful way than having an extra +5 in a swingier roll ever would have.

It felt RIGHT.   It was more FUN.   And mechanically, it kept everybody within the same RANGE (i.e. I didn't have to somehow come up with reasons why low-level baddies were using mastercraft locks, just to give the skill expert a challenge).  I don't need anything more than that in a system, thanks!


I love how your example of "skill dice are good" involves no mention of skill dice, just of skill focus, which does exactly the opposite of skill dice, in that that it ensures minimum competancy, which is exactly what skill dice don't do. I seriously cannot believe I'm the first to bring this up, that's so absurd. That's exactly why skill dice are horrible, because high level characters who haven't spent 25% of their feats on a skill, can't just look at basic checks and go "I'm too good for that", if there were more ways to get similar effects, that would fix a major issue I have with the uh, "non-combat task resolution system which doesn't have a name other than ability checks".


Actually, it depends on your interpretation or definition of 'more swingy', if you mean 'the standard deviation is greater', then you are completely correct, and it is 'more swingy'


 

Here is a great example of why I call that argument disingenuous. On the surface, this argument is true. If you define "more swingy" as "having a larger standard deviation" then you can call the skill die system "more swingy." The standard deviation with a flat modifier is a flat 5.77. Meanwhile, with a skill die, the standard deviation starts at 5.87 and grows to 6.72. But, the difference we are talking about is not a drastic difference (0.95 at most). So, even when we focus on the standard deviation, the change is minuscule. The range of numbers you are statistically probable to roll is greater by less than a single number.  Moreover, focusing on that standard deviation serves only one purpose: obfuscating all the percentages that matter, such as your chances of success and failure. Those never become “more swingy.” In fact, in many cases the skill die makes them less “swingy” (as your percentile probability of success is actually a few points higher). The only time they are “more swingy” is when the number rolled in very high or very low—there is a higher chance that you will fail a very easy skill check or pass a very hard skill check (though in both cases the chances are very small) as you can roll numbers that you simply could not roll before. When you roll 1d20+1d12 you can still roll between 2-7 and 28-32 (which you could not with a flat +7 modifier). Again, though, that is EXACTLY what you want a bounded system to allow! 



Okay, firstly, while "swingy" is a bit vague, you know very well that " your percentile probability of success is actually a few points higher" is completely irrelavent, because "higher chance of success on appropriate task" has nothing to do with swingyness.

The extra "swinginess" is very much about not being able to look at easy tasks and not worrying about them, because there is always the chance of failure, even at level 20, unless you've taken skill focus, which is exteremely specific, and expensive considering how specialized it is.
 

So, claiming that they are de-facto “more swingy” due to something like their standard deviation only serves to hide the way they actually operate. In terms of the bounded system, mathematically speaking, the skill die is NOTHING but good for the game. In all the ways in which you do not want a bounded system to be “more swingy” the skill die is not “more swingy.”

The only real complaint against the skill die amounts to this: “I aesthetically don’t like bounded accuracy when it is applies to skill systems.” That, however, is not the same as “the skill die is more swingy.” What you are actually saying is “I don’t like ever passing very easy or very hard skill checks; I think all skill checks should be trapped into binary success/failures that exist in a middle range window; I think that characters should never be able to fail very easy skill checks or pass hard ones.” Now, I am not going to argue those claims with anyone. They are aesthetic preferences. I STRONGLY disagree with all of them. I think the bounded system is both far more realistic and far better for the game. But aesthetic preferences, ultimately, are aesthetic. Meanwhile, in every way that matters when it comes to a bounded system, namely the percentile probability of success when it comes to numbers in the middle range, the skill die is not more swingy as far as your success rates are concerned. 





Why? why do only middle range tasks matter? In your 20th level group, there's going to be a mage with 10 str and no training who has trouble making DC 10 str checks, so there could easily be DC 10 str checks, and the 20th level Fighter with 20 str and an appropriate skill could fail them, which given what a 20th level Fighter is supposed to represent, and can do in combat* is totally absurd. Yes, it's rare, yes the problem is mostly in people's head, but this is role-playing dude, EVERYTHING is mostly in people's head. Minimum competency matters, and you've have to explain to me why that extra point of Standard Deviation is needed in skills when it's not in attacks or spells.

Especially compared to combat abilities, I mean, in a combat, yoy generally have everybody rolling multiple times, whereas in a non-combat encounter, there's often much less rolling, with everything coming down to 2, maybe four rolls. I know that's not great DMing, but that's how they tend to get used, it's why WotC tried to get people to do something more dynamic in skill challenges (which didn't really work) because having a non-combat encounter revolves around so few rolls instead of a whole list of them already increases the "Swingyness" of an encounter, because as you know, rolling a d20 three times has a larger Standard Devation then rolling it 16 times, so skills as used in a large percentage of games are aleady way swingier than attack rolls, or saving throws, so yes, anything that increases that, however much that is, is a step in the wrong direction, and using an extra dice instead of a flat bonus is a step in that wrong direction.

Not to mention bounded accuracy seperate from skill dice works totally different as currently written for skills compared with combat abilities. When you go from 1st to 10th level as Fighter, yes your abilities to hit an Orc Leader guarding a bag of gold only slightly increases, but the effect of that hit totally changes, similarly a first level Wizard casting Charm Person doesn't have a much lower chance of succeeding than a 10th level Wizard casting Dominate person on the same Orc Leader, but the the effect is quite different. However, if you use a skill to attempt to deal with that encounter, the change of success likelyhood doesn't change much less than the other two options, but the effect of the skill doesn't change at all. Unless skills scale in some way more than they do now, they can only become less and less useful and relevant as you level up, so no, I'm not sold on bounded accuracy for skills, and it's not "an asthetic preference", it's a quite obvious balance issue.


* By which I mean at this point all characters have combat abilities that are so far past "realistic" as to not be able to see such a concept anymore, so the fact that you even mention that shows you've either never played much above 10th level, or weren't paying any attention when it happened.
I know that this is a playtest, and thus they are looking for feedback.   I read many threads from the people who dislike skill dice. The main complaint seems to be the feeling that, since all skills develop at the same rate, nobody is a specialist, and instead everybody is a generalist in their set of skills.   People miss having their +10 or +15 in a skill.

I disagree with them, and my player's rogue last night proved the point.

He had a skill focus in Sneak (minimum d20 roll is a 10) which proved to be FAR more useful and thematic.   He boldly knew that he would never roll less than a 10, and thus his minimum after mods was going to be a 14.  His expertise allowed him to make certain decisions in a far more meaningful way than having an extra +5 in a swingier roll ever would have.

It felt RIGHT.   It was more FUN.   And mechanically, it kept everybody within the same RANGE (i.e. I didn't have to somehow come up with reasons why low-level baddies were using mastercraft locks, just to give the skill expert a challenge).  I don't need anything more than that in a system, thanks!

Indeed, he could immediately see the point in pursuing the expert's path, and choosing various rogue-based skills to emphasize in a similar way.

And this is why I am hoping that others who see the value in the entire skill-die mechanic, and the interesting-and-useful ways it can be mingled with other mechanics, voice their opinions here.  

Let's give some positive feedback, so the developers know that it isn't all  doom-and-gloom.   I, personally, can do without huge modifiers, and I appreciate a more elegant system like the skill dice.

Speak up if you're enjoying them, just to make sure we can keep 'em! 




I didn't read through the whole entire thread, but I agree with your post. I've enjoyed playing using the skill dice and I personally think it is pretty well implemented.