Why Skill Dice?

Can anyone explain to me why we now have a skill die?

Despite initial misgivings about fighter martial damage/maneuver dice, that mechanic is growing on me.  If the system expands universally to other classes (skill trick dice for rogues, metamagic dice for wizards, miracle dice for clerics?), I can see it becoming part of D&D canon.

But I am having more trouble with the existing skill die mechanic.  I gather that rogues can "spend" the skill die, although exactly what that means is not always clear because it is not a limited commodity in the same sense as martial damage dice.  Other than that indirect use, I don't understand why it adds value to the game that a flat skill training bonus does not.

In particular, I don't see the need for even a shallow bell curve in skill checks.  With a +3, you're as likely to roll a 23 as you are to roll a 4, but with a d6 you are suddenly less likely to roll either result and more likely to roll in the middle of the range.  The larger your skill die becomes, the broader that middle swell becomes.  Why is this value?

Can anyone enlighten me?  What am I not seeing?
The way that I have been running it is similar to the MDM where you are given a number of skill die per turn/encounter/day and you can spend them to apply an applicable skill to a challenge.

So if we have a player using a shield to block on oncoming attack, they are given a skill challenge (d20+str/d20+str)  and if they have a skill that they can apply to that block then they can make it so that it's (d20+str+skill/d20+str). The opponent can still land the blow, but with the skill die thrown in it will give the player a boost to being able to perform.

So far we've set it so that they can apply the die if they fail and they don't have to give prior notice.

Another example is if we have someone trying to perform a Knowledge check, if they aren't able to hit the required skill check then they can throw in their skill die for a boost.

We use this feature more than almost any other new feature in the game and I'm fairly sure that my players are big fans of it. No complaints so far and it's been working out fairly well. 
Okay, you've given the skill die some life, but your post is in the vicinity of 99% house rule.  I mean, it is similar to the rules as written in that both you and the designers of the game are calling a game mechanic related to skills and dice "a skill die."

How are you determining the quantity of skill dice and the frequency of refresh?

Blocking with a shield does not require a skill check (and in fact is not an action at all, in the rules as written); how are you determining which ability score checks are permitted skill die rolls?

Can players roll multiple skill dice on the same check?

And my original question still stands -- how do the rules as written make the skill die valuable as opposed to a simple training bonus? 
Our method is 2 die per day per level (So at L4 you will have 8 total dice in a day) and then it's /2 dice per encounter and 1 die per action.

We also created a few feats that I think I posted here that gives you extra skill die per day based on your attributes.. but that's deffinately a house rule so far.

Our Rogue is at L12 and he can perform somewhere around 35 skill checks per day. It really helps us out and it livens up his character. Everyone else just has the base 2/level, but they still really enjoy the system.

I think we read that Blocking reduces half damage, if you succeed in the challenge it blocks it all. Makes our fighter a bit more rounded I think.

One dice per check


As for how they are valuable, I really do enjoy them (even before we wrote out the system above). They let players still have the chance of bad luck come into play (unless they took our table's steadfast skills feat where you can take a static mod of half the die roll) . It also is usually a larger mod than most static skills would have been to this level of gameplay, so even at L2 there is the rare chance of rolling a 25 and performing amazing feats of skill.

The old system had a similar setup, but that was a lot of stat farming to it, this doesn't care if you're a bullrushing son of a tank, if you have the skill Knowledge: History, you're aware of the history that you roll. 
As for how they are valuable, I really do enjoy them (even before we wrote out the system above). They let players still have the chance of bad luck come into play (unless they took our table's steadfast skills feat where you can take a static mod of half the die roll) .



But isn't that the purpose of a basic skill check?  What does the skill die add to this aspect of the game?

It also is usually a larger mod than most static skills would have been to this level of gameplay, so even at L2 there is the rare chance of rolling a 25 and performing amazing feats of skill.



This is what bothers me the most about the mechanic, and maybe I am just of a different mind than the developers and fans.

The way I see it, skills ought to provide predictable, linear success rates.  We randomize those success rates because that is what makes this pastime a game, but generally a certain level of skill ought to result in a certain reliability of success.

I don't understand the value of the additional sweep created by the skill die. 

The old system had a similar setup, but that was a lot of stat farming to it, this doesn't care if you're a bullrushing son of a tank, if you have the skill Knowledge: History, you're aware of the history that you roll. 



I'm not sure what you mean by "the old system," but I also don't see how the skill die helps with the problem you describe.
Trust me Nemo, you are not alone in this.I have seen a few threads like this on the forums.But as of right this moment the devs are mum on why they went with a skill die instead of the progression system that they had on the last packet (you know the one that made sense?).



I too fail to see how a random chance to be "better" makes more sense then a flat static bonus.
Thanks for the solidarity, Dragon.  Good to know I'm not the only one who sees this (or doesn't see it, as the case may be).

I'm not necessarily saying it is a bad idea, I just don't understand the need for it.
I would be suprised if the skill die wasent changed back to a static bonus within the next playtest that is going to drop soonish.Though there is always that small chance that my fear will come true and I'll have to muddle through it for longer then I would like to.
Check out this video

Mike Mearls and Jeremy Crawford talk about quite a bit regarding design, including this design. I had my reservations before watching this video about the skill die as opposed to a static number. After hearing their rationales, I have to say they just might be correct. I want to test to make sure, and am playing this Sunday.

Here is one part of the skill die I like after their explanation. With the skill die then the player only has to look at one number on their character sheet, the ability modifier. Also, if they roll their skill and come up with a number, its easier for others at the table to notice they didn't roll the skill die and add that result than knowing whatever their skill + mod for every skill is after a roll. It's far easier to catch that die just sitting there, never rolled. Then, all you have to do is roll it and add it.

Judging from the direction and design styles we've seen for Next, even if the skill die was here to stay I would be willing to bet the core rules would also include a static bonus variant for those of us who just flat prefer it.
Do you have an opinion on what campaign settings should be printed in D&D Next? If so, please cast your votes in this poll! Poll: What campaign settings do you want to see printed in D&D Next?
I don't really care if it's a die roll or a static bonus, I just don't want all of my skills to automatically increase at the same time. I want to be able to make my character the way I want, an expert in some skills and okay in others.
I don't really care if it's a die roll or a static bonus, I just don't want all of my skills to automatically increase at the same time. I want to be able to make my character the way I want, an expert in some skills and okay in others.


Bingo, Kyrion...

It a problem I see as a common one in the rules... you choose race, class, background, and specialty at character creation and your whole career and progression through it is pretty much laid out before you with very little choice, and any other character that made the same initial choices as you will be substancially the same at every point in their career.

-KW
Check out this video
Mike Mearls and Jeremy Crawford talk about quite a bit regarding design, including this design. I had my reservations before watching this video about the skill die as opposed to a static number. After hearing their rationales, I have to say they just might be correct. I want to test to make sure, and am playing this Sunday.

Here is one part of the skill die I like after their explanation. With the skill die then the player only has to look at one number on their character sheet, the ability modifier. Also, if they roll their skill and come up with a number, its easier for others at the table to notice they didn't roll the skill die and add that result than knowing whatever their skill + mod for every skill is after a roll. It's far easier to catch that die just sitting there, never rolled. Then, all you have to do is roll it and add it.



Oh  man... I saw the video and they really missed the mark on all the reasons they tried the skill dice for.

a) Mearls and Crawford say they believe that the skill  it easier to new players. In my case that turned out to be the opposite. The system returns to the classic noob problem of "which die do I roll now?", I used to have in AD&D. It slows the game down and makes a new player always feel frustrated - as it did in my table. What is brilliant in the d20 system is the simple overarching mechanic that is "When you are trying something roll a d20" - It creates an easy mental cue for the newbie. It is easier to explain it to and then you explain modifiers as you go along, rather than educate him on a different subsystem.

b) You still need to look at your character sheet - this time to see for your skills and skill dice! In our experience it doesn't make an important cognitive difference. On a sidenote I don't see this as a very important design goal. I see myself looking my CS even when I have played a character for years, just to refresh my memory on what skills I have and how can they apply to the situation at hand. 

c) Other player's remembering the bonus is not really the problem in this one. What players tend to forget are the bonuses that come from spells, or some other condition. Ability and Skill bonuses are usually the first that come to mind because they are standard.

d) From what a die size (i.e a random variable value field) is not something that a player can intuitevely perceive as "you are good at something". While when you say "you have a +3 in Climb" every new player can relate to the fact that his character is a better climber than average.

e) It is not very evident to see and explain what the Skill Die's randomness represents in the real world. Now I've seen some people say it means how your skill expertise performs rather than all the variable factors around it, but IMHO makes the narrative side of being a DM more difficult. For example, a PC tries to hide and rolls a 16 on his d20 and a on his Skill Dice. If his skill performance is a 1 and the random factors was the 16 you'll have to find a good reason to reflect that in the description. That makes it a bit lame to accomodate this every time. A simple d20 gives you more narrative freedom.

f) It creates a monster/player difference in gameplay that  doesn't make for a coherent game system.  A dragon has a +5 according to it's "Keen Senses". Normally that would be a... what d8? Why do monsters have static modifiers and not skill dice? 

The only positive side I saw from playtesting it, was that it made rolls more random and therefore more interesting - but it's not something as notable as all the burden it creates. 
I am neither a fan nor a critic of the skill die feature, personally.  I see the value in it, as mentioned by others, and I also understand a lot of the misgivings.  I did have an idea that is something of a hybrid, an inspiration taken from another system I play.

So, in previous packets, every even level you would increase a static bonus by 1 (I was always of the opinion that rogues should have gotten a skill increase every level due to the larger skill set, but whatever =)).  What if, instead of a skill die progression, you would have a "number of skill dice" progression.  So, at the highest level you would roll 2 or 3 skill dice and keep the highest result. 

However, instead of a static progression of the die type for all skills (as it is in the current packet), every even level (or every level), you get to increase a die type for one skill.  So, you would start out with your trained skills at d4, and as you level you can increase to d6->d8->d10->d12.  Also, you can use that skill increase to get a new trained skill instead of increasing an already existing one.  As you level, you get more of these dice to roll on a single skill check.

So, you could have a skill set that looks like:
Climb - d8
Search - d6
Track - d4
Profession (Underwater Basket Weaving) - d12

And so when you're trapped underwater and you really, really need to weave a basket to send a message to the surface, at level 20 you could roll 1d20+Dex+3d12 (keep highest).

The benefit of this is that you can now specialize, and even be better at something faster than its current progression, and you would also wind up with much more stability of rolls as a result of the roll several, take highest.  The rogue's Skill Mastery feature could add an extra die to each skill dice roll.

I figure if people are unsatisfied with the current system, and an idea like the Skill Die is one that the designers want to stick with, then this might be a good compromise.
I don't really care if it's a die roll or a static bonus, I just don't want all of my skills to automatically increase at the same time. I want to be able to make my character the way I want, an expert in some skills and okay in others.


Bingo, Kyrion...

It a problem I see as a common one in the rules... you choose race, class, background, and specialty at character creation and your whole career and progression through it is pretty much laid out before you with very little choice, and any other character that made the same initial choices as you will be substancially the same at every point in their career.

-KW



I agree 100% - this must be adressed. What if a character decides to abandon his peasant roots and become a knight? Shouldn't she be able to learn some Heraldry instead? 

Let me just say that in a static bonus system is easy to fix this. You just adopt the 3.5 rank system and say that you can add the +1 to another skill. Skill Dice will complicate this I'm afraid...
But I am having more trouble with the existing skill die mechanic.  I gather that rogues can "spend" the skill die, although exactly what that means is not always clear because it is not a limited commodity in the same sense as martial damage dice.



It took me a while to grok the "spending" of skill dice, as well.  But, after looking it over, it's pretty easy: when you "spend" a skill die, you don't roll it in the ability check.

For example: Vanish says, "When you use your action to hide, you can spend your skill die to move up  to your speed before you hide."

Normally, a character must have some kind of concealment or cover to hide.  A rogue with this skill trick can spend his skill die to be able to move 30' to a suitable hiding spot before rolling the check.  However, the rogue would not get to add the skill die to his ability check for that hide.

Why the tradeoff?  Skill Tricks are like expanded uses of skills that only Rogues can do.  As a tradeoff for this breadth of capability, sometimes they lose the benefit of their expertise in that skill.  In the Vanish example, the rogue would normally get his skill die on any use of the Sneak skill.  However, the skill trick keeps him from doing that.  Narratively, he's had to rush to get into cover, and he's not had the time to properly hide.  If he fails, perhaps the one who spotted him saw him as he was running to cover.  Maybe the tip of his sword is sticking out from behind the crate.  Maybe he's really tired, and the exertion of the rush to cover has hiim breathing more heavily than normal.

The same thing could be accomplished with a 3E-style system, as well, by changing "Spend your skill die" to "Forgo your skill bonus."  And, I fully expect a 3E-style skills module in the core rulebooks.
Good responses, all.  I'm glad to see people thinking about this mechanic.  I was going to respond to TheLyons personally, but man.of.tomorrow hit everything I was going to say, and more eloquently.

It took me a while to grok the "spending" of skill dice, as well.  But, after looking it over, it's pretty easy: when you "spend" a skill die, you don't roll it in the ability check.



This issue is being addressed in greater detail in another thread, to which I see you have also responded; let's continue that discussion there as it is not entirely germane to this topic.
Have you guys playtested the mechanic yet? I have the static style and loved it. I haven't tested this style yet so I can't give my honest opinion, yet. My last playtest session was nothing but role playing. That's what the players wanted to do so I let them run with it.

My point is, it we haven't tested the mechanic ourselves, we don't know if it works well or not. Also, even if we don't like the system, at the least it can be included as an optional rule in the core rules, even if it doesn't suit our fancy.
Do you have an opinion on what campaign settings should be printed in D&D Next? If so, please cast your votes in this poll! Poll: What campaign settings do you want to see printed in D&D Next?
Other than that indirect use, I don't understand why it adds value to the game that a flat skill training bonus does not.


Because rolling is fun.

d6 is more fun than +3.
D&D Next = D&D: Quantum Edition
Suggestion
Basic game: No skills. Fighters get to add skill dice on STR checks, wizards INT checks, etc. Skill die progresses according to a table.
Standard game: Much like now. Basic skills list, everyone picks 4 at character creation, and each skill improves by die step every so often.
Advanced game: As Landale3's suggestions. Skills list along with "skill points" acquired every so often. With a skill point, you can either add a new skill or improve an existing skill by one die step. More advanced breakdown of DC tasks for players who want more depth in the system.

Not sure I agree with the idea that everyone should get Skill Mastery. At d12, you are really only getting +6.5 to your check every roll. In 3E, you could get to that point at level 3-4, but here they specifically want to keep everyone--people with the skill and those with just a high modifier--in the same playing field. If the average rose any higher (rogues excluded), I think issues of bounded accuracy might creep back. I would love to see some numbers.
Have you guys playtested the mechanic yet?  



I have ve tried it in two sessions so far with an inexperienced player at the table. This is where I saw that it failed the 'intuitive mechanic' test. Now it may be kept or not, but if their stated goal is to make D&D easier for new players, IMHO it doesn't fit the bill. 

 Also, even if we don't like the system, at the least it can be included as an optional rule in the core rules, even if it doesn't suit our fancy.



On a sidenote, there was another similar rule to the skill dice in the 3rd edition DMG that I really liked. 
Instead of having a base Armor Class of 10 you got to roll a d20 on every attack someone made against you - so your armor class was (d20+DEX+Armor etc.) It changed the game entirely in a very interesting way but in the end it slowed down combat so much we had to drop it.  

Skill Dice work better, mathematically speaking, in a bounded system—especially when it comes to skill checks (which don't auto fail/succeed on 1s or 20s). A lot of the feedback came back saying that auto successes felt boring/sucked the excitement out of a skill check. Skill dice, for the most part, leave you with the exact same binary chance of success or failure over an 8 number spread. The higher end DCs stay impossible without a sufficiently large skill dice. With the lower end DCs, however, there is a small percentile chance that you will roll a 1 with your d20 and your skill dice, making your lowest possible roll 2+stat mod for the entire game. This means that with simple checks (and you are only supposed to ask people to make such checks when there is a good reason, as outside of an adventure where Murphy can sometimes poke your pooch to actual effect those checks are supposed to be handled by just allowing the PC to successfully perform the check) there is still some excitement to the check (and a reason to make players roll in certain adventure based situations). Likewise, the skill die also gives you a chance to roll a higher roll than you would have been able to with a static modifier when a dire situation that calls for a high risk check comes around.


What skill dice do for the game, in a bounded accuracy system, is brilliant. Let say you had a +4 bonus from training with X skill (given a flat modifier system). With a 1d8 skill die your percentile chance of rolling 5-24 is roughly the same (give or take a percent with some numbers). In fact, your chance to roll 9-21 is exactly the same. You actually have a lower chance of rolling between 5-8. There is still, however, a small chance that you will roll between 2-4. This keeps your statistical probability modeled in such a way that you feel skilled (as your probability of rolling those low numbers is very low, and lower in some cases than it would have been with a flat modifier), and also keeps the check more dynamic/exciting (as there is a small unlikely chance that Murphy will strike right when it effects you most). On the other side of the system, rolling a 22-24 is slightly less likely (which keeps the hard skill checks feeling hard), but the system allows you a very small chance of rolling as high as 28 (which means that in emergency situations characters might actually try the unlikely).


The skill die, along with bounded accuracy, is one of the most innovative aspects of DDN. Right now, bounded accuracy and the skill die are its two MAJOR selling points as far as I am concerned. I REALLY hope they drop the "1 auto hits, 20 auto crits, and attack bonuses are static" rule system that is currently in place. What they should do is change attack rolls to d20+mod dice rolls akin to the skill die system. A +3 bonus to attack would become a +1d6 bonus to attack. Meanwhile, rolling a 1 would not be an automatic miss and rolling a 20 would only be a crit if you also hit your target. This would work SO much better with the bounded accuracy chassis they are trying to build. 

Basically, bell-curves which still allow for outlier results work better in a bounded system. And, the bounded system is the one true innovation of this edition. 


I have to say, I also like keeping the mathematical mods low. I like most results coming from dice. My past experiences at game tables mirror what Mearls talks about in the video linked above. Of course, I realize that such evidence is anecdotal. It seems that at least one poster in this thread has quite different experiences--antithetical experience, even. But, there it is. I find it easier to remember which dice I need to add to a pool than which host of small modifiers I need to add to a pool. I find that I double check my character sheet more often when I need to add a large static modifier than when I need to add X number of dice. I also find rolling dice more fun than adding large static modifiers (due to the effects of the bell curve math described above). 


That being said, the math behind a flat mod/a die of the appropriate size is SO close that I see no reason why both options cannot be offered. DMs could decide what fits their game best. Either rule system would be an optional module. Neither module would break the chassis of the game (though different players would have different opinions as to which works better). Everyone can go home happy knowing that their preference is supported. That is the point of this edition, no?


P.S. Something else I just thought of: one of the biggest complaints in regards to both the 3e and 4e skill system is that characters were, very often, either so good with a skill that there was no point to roll a skill check or so bad that there was no point to roll a skill check. Basically, very often the range of success/failure left characters with either a 0% chance of success or a 0% chance of failure. As a result, characters often didn’t want to try skill checks unless they were well trained in said skill (because the potential fallout of failure was too high to risk a particular check). One of the ways in which DDN first tried to solve this problem was with a bounded skill system. IE: the max DC was a 25, and the first round of benefits in regards to skill use came in the model of “you automatically pass DC of X or below). That system received a LOT of complaints. Automatically passing skills didn’t feel exciting on the PC end and frustrated many storytellers on the DM end. Meanwhile, characters could end up with very unreasonable chances of success once again. In fact, it was very possible for people to end up with almost a 100% chance of success in nearly all skill checks. The skill die system with an increased range of DCs (max is now 35) fixed ALL those problems in one fell swoop. The bell curve keeps people from, for the most part, automatically succeeding on checks and keeps skill checks relevant. It also makes it possible to hit higher DCs with a low rate of probability, which allows DMs to once in a while throw out the “you are unlikely to pass this check, but if you do it will reap great rewards” side scenarios. Brilliant!

Have you guys playtested the mechanic yet?



Yeah.  It doesn't break anything, at least not at low level.  That's not really my concern.  I just don't understand what it adds to the game other than "fun" as Mand12 puts it.

Personally, I don't find it "fun," beyond the visceral enjoyment of rolling multiple dice.  As I said, I find the sweep distracting and confusing.

And I have to figure that when my skill die hits d12, it's going to be a lot worse.  How do you set fair DCs for d20+d12?  I don't have the math on hand but that sounds like a nightmare.

This thread is not about the power curve, and as much as it makes me nervous I'm still not getting into that.  What I'd really like to see are the probability curves showing the difference between d20+d10 and d20+d12, because I am hazarding a guess that the latter is /not/ a full-on improvement.  My math is rusty, but every time that second die gets bigger, your results become more and more average, right?



If you are looking for what it adds, read my last two posts. I go into it in detail. And no, a larger die does not result in more average results. You would get more average results if it added more small dice. That is not what happens. The size of your second die increases. That increases the probability of rolling any given number (as there are more numbers you can roll which will achieve the total required to pass any given DC), and increases the size of the outlier range that you can hit. You can figure out the bellcurve for 1d20 + 1d12 at anydice.com . Just type: output 1d20 + 1d12 in the function box. Basically, you will have a mean roll of a 17 with a standard deviation of 6.72. That means that any number between 10.28 and 23.72 will come up fairly frequently. Your maximum roll will be a 32 and your minimum roll will be a 2. All of this is not counting stat bonuses or any other class/skill based mechanic. It is pretty easy to figure out what sort of skill checks players are likely to make...
I was writing my previous post while you were posting, Dave, so I deleted my post when I saw you made an effort to answer the questions I hadn't yet asked.  I still have a lot of reservations but I think they can be summed up fairly easily:

Setting DCs shouldn't require knowing a standard deviation.

The skill die works fine for DCs in the middle of the range or at the extreme ends, but set a DC just outside the +/-1 SD range (7-9 or 24-26) and you have a 10-13.3% difference in success rate from a standard single-die roll.

I just can't see these probabilistic gyrations we are going through as intuitive or good.
I've ran three sessions with the mechanic in place so far and everyone understands it pretty intuitively.

They tend to set their skill die aside near their d20 to remember it's there. Most of what they try comes from pure ability score checks, but occasionally they get to add the skill die to fantastic results.

The big intention is to take the focus OFF the skill system. The skill system is NOT the task resolution system. That was a big problem in 4E. I'm playing in a campaign and every time there's a skill challenge, everyone is craning to figure out how to use their skills in a relevant way in things that are wholly outside their area of expertise.

The skills come in when you're doing something you're good at. Linking it to class benefits like Skill Tricks and Parry makes it even more interesting, as it provides progression for class abilities as well.

Also, the metaphor is pretty easy to understand.

Let's try a knowledge skill. This is easy. Everyone knows something about the world around them. Some have done studies on it. They've read books and consulted sages. If the skill die comes up a 1, they've probably only read a little about the specific thing they're trying to remember in a massive lexicon. If it comes up a 12, they've written an extensive research paper on it. Remember that Knowledge Warfare means that you know about famous historical battles. Just having it as a skill means you know a little more about it than most; a low skill die roll means you've heard some tales of it, a high skill die roll means you've memorized troop movements, death counts, and who stood out as exemplary in the battle.

Now let's go with the stealth roll. If your skill die comes up low, it means that you've gotten a bit lazy with it, but even you on your worst day is a little better than those untrained in stealth on your best day. If it comes up high, you're paying so much attention that you minimize the blades of grass that bend from your step.

Now let's go with climb. A low roll means that you're not paying too much attention to your footing because you're focusing too much on your hand techniques. A high roll means that you're in the mountain mode, and all of your movements are being carefully orchestrated from hammering in pitons to jumping to difficult ledges.

Remember, almost any check under 25 is within the possibility of the untrained. Checks of DC 25 should be rare. 20 means 'pretty hard.' Skill dice are not there to enable you to accomplish a specific task; your ability modifier is there to mark how good you are at doing just about anything. Skill dice are there to reinforce what your specific character is highly trained in and excels at. When characters roll it, they should feel that their area of expertise is coming into play. Fighters are good at parrying. Rogues are good at being sly. Anyone can be really good at underwater basket weaving if their character was an underwater basket weaver before picking up the sword/staff/mace/whatever. Skills are here to personalize characters, not enable characters to even succeed.
I was writing my previous post while you were posting, Dave, so I deleted my post when I saw you made an effort to answer the questions I hadn't yet asked.  I still have a lot of reservations but I think they can be summed up fairly easily:

Setting DCs shouldn't require knowing a standard deviation.

You don't. Knowing a standard deviation (which is really not that big a deal, considering the tools we have online to ease up the amount of math actually performed) merely allows you to choose the numbers you use with a little more finesse. DMs who are a little more advanced at our trade will likely check means and standard deviations. If you merely read the skill description for particular skill DCs, however, that is enough. The descriptions given are very clear, and accurately model the actual math of the system. So, all one needs to do is read the skill descriptions and then pick DCs based on them.


If you are ok with doing a little more math, you can quickly guesstimate a mean by taking 10 + 1/2 the maximum number rolled given your skill die. You can then quickly guesstimate the standard deviation by taking +/- 6 from that number. It will not be exact. It will, however, be very close. It is really not hard to figure out what numbers someone is likely to hit. I say that as a DM.



The skill die works fine for DCs in the middle of the range or at the extreme ends, but set a DC just outside the +/-1 SD range (7-9 or 24-26) and you have a 10-13.3% difference in success rate from a standard single-die roll.

I just can't see these probabilistic gyrations we are going through as intuitive or good.




On the contrary, they are both intuitive and good for the game. The numbers on the lower end (your 7-9) are "10-13.3%" (I am not double checking your numbers right now, so I am just repeating the numbers you gave me) easier to roll over than the numbers in the middle. This helps model the fact that easier DCs represent tasks that are easier for a character to succeed at. The numbers on the higher end are, likewise, a little harder to hit. That helps model the fact that those DCs are the hardest DCs a character of that training is expected to make. It is that fact, that bell curve, that solves the "there is just no point to roll" issue that arose in 3e and 4e while still modeling the fact that the tasks assigned to those numbers are easy/hard to perform. Moving the probability rate of success with any given DC to a normal distributionmakes skill checks far more organic and natural. They should have never used a linear modeling system. That is exactly what broke 3e/4e skill mechanics in the first place! It makes no real sense (from the point of view of statistical math) that it is as easy to roll a specific low number (which is supposed to be hard to fail) as a specific median number. That is just not how things tend to actually be modeled, statistically speaking. And, that modeling created massive problems in 3e and 4e. It was that modeling, in part, that made the treadmill a requirement. 


The thing about the skill die that makes it so smart, so great, is that it uses a bell curve to make easier tasks statistically easy without making them automatic and makes hard tasks statistically hard without making them impossible. Meanwhile, the range of numbers that you can actually expect to occur commonly will be about +/-6 from the mean roll. 

Check out this video

Mike Mearls and Jeremy Crawford talk about quite a bit regarding design, including this design. I had my reservations before watching this video about the skill die as opposed to a static number. After hearing their rationales, I have to say they just might be correct. I want to test to make sure, and am playing this Sunday.

Here is one part of the skill die I like after their explanation. With the skill die then the player only has to look at one number on their character sheet, the ability modifier. Also, if they roll their skill and come up with a number, its easier for others at the table to notice they didn't roll the skill die and add that result than knowing whatever their skill + mod for every skill is after a roll. It's far easier to catch that die just sitting there, never rolled. Then, all you have to do is roll it and add it.

Judging from the direction and design styles we've seen for Next, even if the skill die was here to stay I would be willing to bet the core rules would also include a static bonus variant for those of us who just flat prefer it.



Thank you for the link. I do not agree that it is easier to take 2 dice add them together and add your ability score, than to add a skill mod an ability score and a die...its 3 numbers you add up in both cases. It could even be argued that seeing as you have 2 unknowns and 1 static that its slower with slow calculating players. Neither is it any easier to notice if people have a skill just because they have a die on the table, our table is pretty crowded with dice at is is ;)

THE DIE ITSELF
I've run with the skill die (in a level 11 game), my players are fairly competent players. It was very smooth, I had to remind them to remember to add their mod as well as they forgot it in all the excitement about the new die system.

 As a GM it doesn't really matter to me, as Cyber Dave pointed out, its minute details that distinguishes it in terms of broader statistics from a static buff. As long as it is kept within the bonded accuracy I couldn't really care less if a player has a sneak of d8 or +4.

As a player I feel the same, BUT I hate the 4e approach of "you are now trained in stealth, bluff and knowledge arcana...you will increase equally in all of these throughout the game no matter if you want to or not". Individual skill die increases to individual skills please.

SPENDING skill die
 Needs less fluffy writing. If you spend something you obviously cant use it again before youve earned it back, but there is no way to earn it back.
 This is the area however where the Die shines over a static buff, its easier mentally to forego your dX than to take a -X to your static skill. Would be lovely to see other classes than the rogue being able to spend their skill die though, EG get rid of Identify and Read magic and let the wizard spend his skill die to do this.
I was writing my previous post while you were posting, Dave, so I deleted my post when I saw you made an effort to answer the questions I hadn't yet asked.  I still have a lot of reservations but I think they can be summed up fairly easily:

Setting DCs shouldn't require knowing a standard deviation.

You don't. Knowing a standard deviation (which is really not that big a deal, considering the tools we have online to ease up the amount of math actually performed) merely allows you to choose the numbers you use with a little more finesse. DMs who are a little more advanced at our trade will likely check means and standard deviations. If you merely read the skill description for particular skill DCs, however, that is enough. The descriptions given are very clear, and accurately model the actual math of the system. So, all one needs to do is read the skill descriptions and then pick DCs based on them.


If you are ok with doing a little more math, you can quickly guesstimate a mean by taking 10 + 1/2 the maximum number rolled given your skill die. You can then quickly guesstimate the standard deviation by taking +/- 6 from that number. It will not be exact. It will, however, be very close. It is really not hard to figure out what numbers someone is likely to hit. I say that as a DM.



The skill die works fine for DCs in the middle of the range or at the extreme ends, but set a DC just outside the +/-1 SD range (7-9 or 24-26) and you have a 10-13.3% difference in success rate from a standard single-die roll.

I just can't see these probabilistic gyrations we are going through as intuitive or good.




On the contrary, they are both intuitive and good for the game. The numbers on the lower end (your 7-9) are "10-13.3%" (I am not double checking your numbers right now, so I am just repeating the numbers you gave me) easier to roll than the numbers in the middle. This helps model the fact that easier DCs represent tasks that are easier for a character to succeed at. The numbers on the higher end are, likewise, a little harder to hit. That helps model the fact that those DCs are the hardest DCs a character of that training is expected to make. It is that fact, that bell curve, that solves the "there is just no point to roll" issue that arose in 3e and 4e while still modeling the fact that the tasks assigned to those numbers are easy/hard to perform. Moving the probability rate of success with any given DC to a normal distributionmakes skill checks far more organic and natural. They should have never used a linear modeling system. That is exactly what broke 3e/4e skill mechanics in the first place! It makes no real sense (from the point of view of statistical math) that it is as easy to roll a specific low number (which is supposed to be hard to fail) as a specific median number. That is just not how things tend to actually be modeled, statistically speaking. And, that modeling created massive problems in 3e and 4e. It was that modeling, in part, that made the treadmill a requirement. 


The thing about the skill die that makes it so smart, so great, is that it uses a bell curve to make easier tasks statistically easy without making them automatic and makes hard tasks statistically hard without making them impossible. Meanwhile, the range of numbers that you can actually expect to occur commonly will be about +/-6 from the mean roll. 




I love reading Cyber-Dave's number crunching explanations.   I hope that WotC does that too so that when the system is finalized, we don't have to think about the probabilities of success, standard deviations, etc.   This should be done "under the hood" certainly before all is said and done.   

I've found in our games (as player and DM) that the skill die does make player decision making faster and easier.   If I'm trained, I feel as if I should try to do something if the situation is right, and as a player, I don't worry so much about my chances to succeed or fail since the skill die is a 2nd random die roll that piggy backs on my d20 random die roll.    

The one thing I don't like is how Rogues have to roll the skill die 2x and pick the highest for Skill Mastery.   That slows down the game and is a little more messy.  I'd rather see Skill Mastery grant a d20 roll no lower than a "10".  That would almost guarantee success for a 1st level rogue attempting a moderate tasks of DC 15...but DC 18 would be a little more tricky.   Personally, I don't mind a rogue auto-succeeding in a task.  Even with the first package, we house ruled that even when the rogue had an auto-success situation, he still had to roll the d20.  A "1" resulted in a complication.    Even that 5% chance of muffing it, made my players tense.


A Brave Knight of WTF - "Wielder of the Sword of Balance"

 

Rhenny's Blog:  http://community.wizards.com/user/1497701/blog

 

 

I have mixed feelings about skill dice. My biggest concern with them is that they might not add enough reliability. Even a highly skilled person, with a d12 skill die, can still get a check result of 2 (1 on the d20 +1 on the d12). I just can't see such a highly skilled individual failing that spectacularly that often. That said, I think the old packet made skills a bit too reliable, especially in the case of rogues. I'll just have to playtest it some more. I'm not sure whether I like them or not yet.

The bad thing is, players are only getting their skill dice on a few specific skills, and the DCs of actions shouldn't assume that players have a skill die. But they do. The DCs of actions were dramatically increased in this latest packet, and it's now impossible to succeed on very hard tasks unless you have a 20 ability score and roll a 20 on the d20, or if you have a skill die. Formidable tasks are flat out impossible unless you have a skill die. This needs to change. They need to go back to 25 being the highest DC.

The other problem I have is that all skills progress at the same rate. You have only a few skills, all of which are chosen at 1st level, and then that's it, you never gain any more skills (unless you take a feat). And then, they all progress at the same pace. You can't be more skilled in one area than another. That may be okay for more basic games, but I'd like to have the option to improve skills individually, and I think players should also gain additional skills as they increase in level without having to spend a feat.

As for rogues, in particular, I like the general idea behind their skill tricks, giving them cool abilities, but I don't like the current mechanics of letting them roll their skill dice twice or having to "spend" them to use tricks. Having to spend your skill dice to do a trick means you become less capable of the attempted action. For example, the charming presence trick lets you give up your skill dice on a diplomacy type check to charm the target if you succeed. But because you gave up your skill die, you're far less likely to succeed. I don't like that. Rogues shouldn't have to sacrifice competence in order to be able to do cool tricks.

My other problem with the rogue tricks (which is unrelated to skill dice) is that there are quite a few things that IMO shouldn't be rogue-only. Only rogues can have Use Magic Device, read lips or be master linguists? These used to be things any character could get if they were willing to invest in the appropriate skills.

If they can address these issues, I can probably get behind the skill dice idea. 
Knowing a standard deviation (which is really not that big a deal



This is where you lost me, Dave.  I've got a college degree; I took college-level statistics and calculus, and I would have a tough time doing this at the table (as evidenced by my terrible, terrible math in this thread -- and I had a spreadsheet!).  Everything you said after this point was aimed at a far, far more left-brained gamer than I will ever be.

Introducing a 10% error in DCs around the 33% success-rate mark is just terrible game design, in my opinion.  I appreciate that the system has mathematical elegance, but it fails at what I'd call the crux of the enterprise.

I love reading Cyber-Dave's number crunching explanations.   I hope that WotC does that too so that when the system is finalized, we don't have to think about the probabilities of success, standard deviations, etc.   This should be done "under the hood" certainly before all is said and done.



Alas.  This is what I was most afraid of.  DMs have to be able to set and tweak DCs on the fly; they cannot rely on a table.  This is practically the most important mechanical thing the DM does.

I'm sorry to say that this thread is rapidly confirming the negative suspicions I had about this mechanic.
I'd rather see Skill Mastery grant a d20 roll no lower than a "10".


They already tried that.  The feedback was overwhelmingly negative.

You need to know how to run DCs, and the way they set it up is pretty easy and elegant.

10- Modest chance of failure, but generally easily doable. Most can handle it if they're adept at the general area of proficiency associated with it (Ability Modifier)

15- Moderately difficult task. Anyone can pull it off, but it's a bit trickier and has a higher chance of failing.

20- Difficult task. It's achievable by anyone in a dire pinch, but most who aren't trained will fail.

25- Very difficult. Requires some kind of special training or extraordinary ability. Reserved for deviously powerful traps, incredibly paranoid barons, and the most esoteric of lore.

30- Impossible. Something that should not be able to be done. Those who can pull it off will be legendary for their efforts.

35- Godlike. Moving an immovable rod. Feats that seem to break the laws of common sense. Learning the dark secret of a deity of secrets by piecing together random bits of lore about them and making a grand discovery. Disarming the trap on the door to godhood.

At the table, most DCs should fall between 10 and 20. DC's of 25 are for rare occasions at high levels. DC's of 30 are for astronomically stupid plans that just might work to everyone's surprise. DC's of 35 are for world-impacting checks. 

First of all, I just noticed that one of the things I wrote was confusing. I said that the lower range numbers are easier to roll. That is not what I mean. I meant they are easier to succeed at because they are harder to roll. The statistical probability of rolling them makes them easier to succeed at as you are for more likely to roll over their value. Likewise, as it is harder to roll the higher end numbers, it is far easier to fail on those checks (as you are far more likely to roll under them). My overall point, however, was accurate. I just miswrote/Freudian typed roll instead of succeed at. 



This is where you lost me, Dave.  I've got a college degree; I took college-level statistics and calculus, and I would have a tough time doing this at the table (as evidenced by my terrible, terrible math in this thread -- and I had a spreadsheet!).  Everything you said after this point was aimed at a far, far more left-brained gamer than I will ever be.


Then I really don't understand why you are having so much trouble. This stuff is a lot easier than college level math. Hell, I am an arts student! I am writing my PhD in English Literature, not statistics. I just don't see what the big deal is. Maybe I am just left-brained despite my arts degrees. I don't know. Let me try and explain it in a more clear way.


When rolling 1d20 + 6 your chance of rolling 2-6 is 0. Your chance of rolling a 7 is 5%. When rolling 1d20 + 1d12 your chance of rolling between 2-6 is 6.25%. Your chance of rolling a 7 is 2.5%. Your chance of rolling 7 or lower is 8.75%. So, you are 3.75% more likely to fail a DC 8 skill check (as you need to roll equal to or higher than to succeed) than you would be with a flat math system. Your chance of rolling an 8 or lower in a flat math system is 10%. Your chance of rolling an 8 or lower with the skill die system is 11.7%. You only have a 1.7% higher chance of failing with a skill die. Your chance of rolling a 9 or lower with a flat math system is 15%. Your chance of rolling a 9 or lower with a skill die is 15.03%. You are only 0.3% more likely to fail with a skill die. Your chance of rolling a 10 or lower with a flat math system is 20%. Your chance of rolling a 10 or lower with a skill die is 18.78%. You are 1.22% more likely to succeed. Your chance of rolling an 11 or lower with a flat math system is 25%. Your chance of rolling an 11 or lower with a skill die system is 22.95%. You are 2.05% more likely to succeed. Your chance to roll a 12 or lower with the flat math system is 30%. Your chance of rolling a 12 or lower with the skill die system is 27.53. You are 2.47% more likely to succeed. Now, between 13-21 both systems will advance at a rate of +5% more chance of failure per higher number. At any of those DCs the skill die system gives you a 2.47% higher chance of success. You have a 75% of rolling 21 or lower with a flat math system. You have a 72.53% of rolling 21 or lower with a skill die system. After that things start sliding again, with the flat math system progressing at a fixed 5% of failure per higher DC number, and the skill die giving you a better and better chance of success the higher you go. A flat math system gives you a 0% of rolling more than 26. Meanwhile, you have an 8.75% of rolling between 27-32.


I am saying this is great for the system. 6.25% is not statistically probable. But, that chance of rolling between 2-6 keeps players on their toes EVEN when performing an easy task in a high stress environment (as, otherwise, no skill check would be called for at all). That ensures that the environment actually feels stressful. While, when comparing a flat math system to a skill die system, it is a little harder for the character to succeed on checks up to DC 9, it becomes easier to succeed on higher skill DCs. The difference on most skill checks will not be that great, but it grows between 21-26. With a flat math system you only have a 5% of making a DC 26 skill check; with a skill die system you have a 11.67%.  Meanwhile, while a flat math system gives you no chance to succeed on a DC higher than 26, you still have a small chance to succeed on the DC range 27-32 with a skill die. That means that in emergency situations players are more likely to try and make high DC skill checks. This fixes the “don’t bother rolling” problems created in 3e/4e (where there just didn’t seem to be a point to roll if a DC was too easy/too hard). It keeps the bounds of the system working better.


Meanwhile, if we look at the internal math in relation only to itself, it works very logically—far more logically than linear skill system math does. The chance of rolling the lower numbers is very low. With each number, the percentile increase in failure grows by a very small margin. For example, that increase between a DC 5 and a DC 10 (again, as always in this post, assuming a d12 skill die) is only 14.58%. This makes it very easy to succeed on lower DC numbers. Then, in the middle range the percentile increase in failure grows by a very large margin. That increase between a DC 15 and a DC 20 is 25%. It grows by over the double the amount seen in my last example! While the extra chance of failure between numbers in the higher ranges grows smaller once again, so does the extra chance of success. Meanwhile, you have all of the middle range numbers adding to your chance of failure (instead of your chance of success, as in the case of trying to pass a low DC skill check). As a result, it is very hard to pass the high DC numbers, and becomes harder and harder the higher up you go (even if it is still easier than it would have been with a fixed modifier system).


The end result is this: it is very hard to fail a low DC skill check, but it is still possible; it is very hard to pass a high DC skill check, but it is still possible; the chance of passing a middle range skill check is about the same as it was before—specifically, you are 2.47% more likely to succeed on any DC between 13-21. All of those statements are good, except for the fact that it is 2.47% easier to pass middle range skill checks, which is neither good nor bad (and is largely negligible).



Introducing a 10% error in DCs around the 33% success-rate mark is just terrible game design, in my opinion.  I appreciate that the system has mathematical elegance, but it fails at what I'd call the crux of the enterprise.

Alas.  This is what I was most afraid of.  DMs have to be able to set and tweak DCs on the fly; they cannot rely on a table.  This is practically the most important mechanical thing the DM does.

I'm sorry to say that this thread is rapidly confirming the negative suspicions I had about this mechanic.


The skill descriptions succeed at the crux of the enterprise. You don’t need to know any of this math. All you need to do is read the skill descriptions. They explain all of this. The one thing they could do to make things more clear is give a few examples of what sort of characters could roll any given skill check as their mean value. And, figuring things out on the fly is really not all that hard. All you need to do in your head is say: “his skill die is a d8 and he has no stat bonus. I guesstimate that his mean will be about 14, and he is likely to roll between 8-20 most of the time. His relatively easy skill checks should be about 8. His relatively hard skill checks should be about 20. His relatively average skill checks should be about 14.” I just did all that without a calculator and without exact math. Double checking my math, the exact mean is actually 15 (I was off by 1), and the exact standard deviation is 6.20 (I was off by 0.20). My guesstimating on the fly would have broken nothing and took me all of a couple seconds. Meanwhile, I can compare those numbers to the numbers in the book to see if his character is effective or not at the task in question (by comparing how his relative capabilities relate to actual capabilities described in the rules).


I am not seeing a problem. All I am seeing is a very simple rule set that utilizes some complex under the hood math to elegantly solve most of the big problems in the 3e/4e skill system. But like I said, the math is close enough to using fixed modifiers that you could easily do so without breaking the game. I don’t think it would run nearly as elegantly as it does now, but there is no reason not to provide a modular “use fixed averages” rule for those DMs who prefer that. 

You need to know how to run DCs, and the way they set it up is pretty easy and elegant.

10- Modest chance of failure, but generally easily doable. Most can handle it if they're adept at the general area of proficiency associated with it (Ability Modifier)

15- Moderately difficult task. Anyone can pull it off, but it's a bit trickier and has a higher chance of failing.

20- Difficult task. It's achievable by anyone in a dire pinch, but most who aren't trained will fail.

25- Very difficult. Requires some kind of special training or extraordinary ability. Reserved for deviously powerful traps, incredibly paranoid barons, and the most esoteric of lore.

30- Impossible. Something that should not be able to be done. Those who can pull it off will be legendary for their efforts.

35- Godlike. Moving an immovable rod. Feats that seem to break the laws of common sense. Learning the dark secret of a deity of secrets by piecing together random bits of lore about them and making a grand discovery. Disarming the trap on the door to godhood.

At the table, most DCs should fall between 10 and 20. DC's of 25 are for rare occasions at high levels. DC's of 30 are for astronomically stupid plans that just might work to everyone's surprise. DC's of 35 are for world-impacting checks. 


Exactly. Only, there is also DC 5, for those checks that almost everyone will pass, but some less adept characters might fail during stressful situations. It is pretty simple and straight forward, and it is in the DM guide rules right now. This is all the over the hood knowledge anyone really needs. Most of what I talked about in my last few posts isn't stuff anyone needs to know. It is just an explanation of how/why it works so well. 

So, needless to say, the skill die gets much love from my corner.  



I love reading Cyber-Dave's number crunching explanations.   I hope that WotC does that too so that when the system is finalized, we don't have to think about the probabilities of success, standard deviations, etc.   This should be done "under the hood" certainly before all is said and done.   




Thank you.  Just keep in mind I meant "easier to roll over" and not "easier to roll" (in regards to the low DC numbers). 

The one thing I don't like is how Rogues have to roll the skill die 2x and pick the highest for Skill Mastery.   That slows down the game and is a little more messy.  I'd rather see Skill Mastery grant a d20 roll no lower than a "10".  That would almost guarantee success for a 1st level rogue attempting a moderate tasks of DC 15...but DC 18 would be a little more tricky.   Personally, I don't mind a rogue auto-succeeding in a task.  Even with the first package, we house ruled that even when the rogue had an auto-success situation, he still had to roll the d20.  A "1" resulted in a complication.    Even that 5% chance of muffing it, made my players tense.



I actually think it works very well as is myself, but to each their own.
Auto-success is boring as hell. It's more intriguing to be incredibly more LIKELY to succeed than it is to succeed without even trying. Also, being more likely to succeed at higher level tasks is more satisfying than automatically succeeding at lower level ones but being just as good as everyone else at more difficult things.
I like that skill dice will make low rolls and high rolls more uncommon. With one bad roll, a player will blame the d20. With two bad rolls, I think the player will feel the universe conspiring against him.

Regardless of Cyber-Dave's excellent analysis, I do believe Nemo_The_Lost has a point about the difficulty of creating fair DCs on the fly. In previous systems, if a player could only succeed by rolling a 10, and the DC increases by 2, then now the player can only succeed by rolling an 8. That player is only 80% as likely now to succeed, or 40% down from 50% total chance. If the roll needed dropped to a mere 5, and the DC still increased 2, then they now need a 3. That's a much more significant loss: 60% as likely to succeed, or 15% down from 25% total chance. No graphs or tables required; the impact can be eyeballed.

I don't believe with a skill dice system I will be able to judge DCs and modifiers on the fly. I will have to lean on the bounded accuracy crutch and cross my fingers more often. Raising the DC by 2 will be more difficult at the low/high fringes and less difficult in the mid-range, but I have no idea how to calculate that without pulling up a graph.

But that's OK. As long as the bounded Easy/Medium/Hard DC paradigm works, then I'll be happy.
I like that skill dice will make low rolls and high rolls more uncommon. With one bad roll, a player will blame the d20. With two bad rolls, I think the player will feel the universe conspiring against him.

Regardless of Cyber-Dave's excellent analysis, I do believe Nemo_The_Lost has a point about the difficulty of creating fair DCs on the fly. In previous systems, if a player could only succeed by rolling a 10, and the DC increases by 2, then now the player can only succeed by rolling an 8. That player is only 80% as likely now to succeed, or 40% down from 50% total chance. If the roll needed dropped to a mere 5, and the DC still increased 2, then they now need a 3. That's a much more significant loss: 60% as likely to succeed, or 15% down from 25% total chance. No graphs or tables required; the impact can be eyeballed.

I don't believe with a skill dice system I will be able to judge DCs and modifiers on the fly. I will have to lean on the bounded accuracy crutch and cross my fingers more often. Raising the DC by 2 will be more difficult at the low/high fringes and less difficult in the mid-range, but I have no idea how to calculate that without pulling up a graph.

But that's OK. As long as the bounded Easy/Medium/Hard DC paradigm works, then I'll be happy.




Here is the thing though: you can eyeball it just like you did before. I it just so happens, in the middle range +/- one number still changes your percentile probability by 5% (no matter what your skill die is) just like it did before. That is an artifact of using a d20. Moreover, the mean of each skill check tends to be equal to the average of a d20 roll (10.5) + the average your skill die roll, which is about half its maximum value (+ 0.5 to be exact). So, if he has a d4 skill die you can eyeball that will roll 13 50% of the time. +/- one number will still make him +/- 5% more or less likely to pass a check. With a d10 skill die the 50% statistic would be about 16. Again, +/- one number will still make him +/- 5% more or less likely to pass a check.

The only thing that is harder to eyeball is what happens with fringe numbers. But, you don’t need to eyeball those numbers. All you need to know is that if you place a number on the low fringe a character will almost always pass (but there is a chance they will fail as long as the DC is higher than 2+ their stat mod). Meanwhile, if you place a number on the higher fringe, a character will almost always fail (though there is a chance they will succeed). You should then plot those DCs accordingly (as in, don’t place high number fringe DCs as binary pass or the quest fails DCs. Use them sparingly and for dramatic effect. Make them side DCs that grant significant benefits, or “critical pass DCs” that grant extra benefits if you roll them when you are making a skill check that otherwise grants success on a lower number). Meanwhile, you can always estimate where the fringe starts by adding/subtracting 6 from the 50%/ mean value. So, the fringe for someone rolling a d10 sill die would be any number lower than 10 or higher than 22.

I don't like the skill dice system the way it is. It does not ever take into account that in gaining 19 levels, a character might get better at some skills. The skill die only says that they have a better chance of getting lucky at some skills. They also have a better chance of the die srewing them over too.

It is entirely possible for a level 20 character skilled in climbing with a +5 stat bonus to roll a 1 on each die and fail to climb a ladder in a stressful situtaion. I game with people who can make those rolls!

We need some points that we can place into skills as we level. They can put in a limit as to how many you can place into one area, stat or skill, but we need a sense of our characters actually getting better at certain tasks/skills as they level. If they need to lower the skill die some to balance it, then I am fine with that as long as it doesn't screw over people who didn't put points into a skill they have expertise in.
The baseline DCs should be matched with corresponding descriptors like they are in the playtest package, so we don't have to set DCs.    A difficult check is a difficult check.   The skill die should not even be part of it.   I'm not going to change the design of my adventure based on which PCs are in it and which skills they are trained in.   If I set a spot check at Moderate, it will be easier for a trained PC.  That's all.   That's why the numbers should be crunched before the chart is published.   Unlike 4e, DCs will not scale with level.  That's the beauty of this skill system.   It is much easier.

A Brave Knight of WTF - "Wielder of the Sword of Balance"

 

Rhenny's Blog:  http://community.wizards.com/user/1497701/blog

 

 

Regardless of Cyber-Dave's excellent analysis, I do believe Nemo_The_Lost has a point about the difficulty of creating fair DCs on the fly. In previous systems, if a player could only succeed by rolling a 10, and the DC increases by 2, then now the player can only succeed by rolling an 8. That player is only 80% as likely now to succeed, or 40% down from 50% total chance. If the roll needed dropped to a mere 5, and the DC still increased 2, then they now need a 3. That's a much more significant loss: 60% as likely to succeed, or 15% down from 25% total chance. No graphs or tables required; the impact can be eyeballed.



I just had to quote this because you made the exact same mistake I did earlier in that if you start with a DC of 10 and it gets more difficult by 2 then you need a 12, not an 8 -- higher DCs are more difficult.

Solidarity, my cruciferous brother.

I know we both know this; it's been a core mechanic forever, which raises the fascinating question of why we both made the mistake.

I'm sticking to my guns that the answer is that "WTF PROBABILITY MATH" is enough of a reason to put the skill die out of its misery.  Again, I'm not arguing that it is not a better sysem overall, I just think it suffers from a critical flaw: math is hard.

Here is the thing though: you can eyeball it just like you did before. I it just so happens, in the middle range +/- one number still changes your percentile probability by 5% (no matter what your skill die is) just like it did before.



Have to call you on the carpet on this one, Dave -- yeah, that plateau is straight 5%s all the way across -- unfortunately, the larger your skill die, the less relevant to the actual linear numeric scale that is.  At a d12, the percent chance per linear numeric increase (from 1-32) is closer to 3%.

The only place you see a match between the probability curve and the linear progression is, again, at about the 33% and 66% marks (speaking linearly), which is /still/ misleading because that's where you see the largest difference (~10%) between the linear and probabilistic scales.