The Dragon Die

Sarah Darkmagic - Posted on 07 November 2010

One piece of advice I’ve received from multiple people I trust is that I should learn about games beyond D&D. Not only will they help me learn how to become a better GM but this knowledge is essential if I ever want to become a game designer/developer. That advice along with my general love of learning pushes me to expand my RPG knowledge. With that in mind the other night, I picked up the Game Master Guide that came in my Dragon Age RPG boxed set. This book is full of great GM advice, useful in other games as well. I'll discuss many of them in a later blog post but for now I want to discuss the Dragon Die mechanic in the game.

Unlike D&D, where skill checks are made using a d20, Dragon Age uses 3d6 for the roll. This change has some mathematical arguments in favor of it, many of which Chris Sims discussed in his article, Dice Heresy, on [1]. This change to multiple dice does more than change the mathematics of the roll. For example it opens the possibility of adding a rich dice mechanic [2]. Dragon Age RPG does this through the introduction of the dragon die. I may be oversimplifying this, but skill checks are called tests in Dragon Age. Some tests are binary in nature, all that matters is whether the character succeeds or fails. For a success, the overall roll has to cross a particular threshold. Other times, however, we care about a degree of success. For instance, when convincing a merchant to lower his price, we care not only about whether or not he lowers the price, but the degree to which he does so.

In the Dragon Age RPG, one of the 3 dice will be a different color and is called the dragon die. This die determines the degree of success for non-binary tests. So, in the bargaining example, the character attempts to negotiate the price, either against a fixed threshold or against an opposed roll. If he succeeds in his test, the GM would determine the degree of success by looking at the result on the dragon die. Perhaps the discount is a multiple of the value on the die or the merchant's feelings about the character change. The GM only has to worry about 6 possible values and can group these values as she sees fit.

Now some examples of degrees of success exist in D&D 4e but they are often tied to the overall result. Monster knowledge checks result in different amounts of information depending on the result of the die roll. The horizontal distance jumped is calculated by dividing the die roll by 10 (5 with a running start). Many areas where we have degrees are used to tell us how badly we fail rather than how well we succeed. For instance, failing a swim check means you stop movement. If you fail by 4 or less means you are able to tread water whereas a failure by 5 or more means you sink a square.

Introducing degrees of success to D&D 4e also brings with it a number of complexities. For some checks, we might not have the granularity we want for a particular die roll. For more difficult challenges, we have fewer likely outcomes above the threshold, at least for some players. Yet these tests are the ones where we are most likely to want to provide different degrees of success.

We also have the difficulty of determining what intervals to tie these results to. The range of outcomes for a particular skill varies greatly between players. If we tie a particular degree of success to a higher threshold, it’s possible for that success to just not be possible for some of the players at the table. Now, for some narrative elements this makes sense. Perhaps a player shouldn’t be able to jump across a chasm without falling prone unless they can beat a hard DC + 5.

Finally, a nice part of the dragon die is that they allow us to randomize what a success means, allowing us to add interesting complications. Instead of dealing solely with degrees of success, we can use the result of the dragon die to pick randomly between 6 outcomes of equal value. Each of them could add a different complication to the result with none of them being harder or easier than the others. In the heat of the game, I’m often terrible at this, not wanting to screw my players even though I know at the end of the day the game will be better if they have interesting choices to make.

Of course, this points to a potential downside of this mechanic. If I roll 6s on my other two dice and have decent bonuses to the roll, I may be a bit perturbed when I find out I still only barely clear the chasm and am left hanging off the edge because my dragon die has a 1. Especially when the same result with a 6 on the dragon die instead would have left me sitting pretty. Howeer, I think this “unfairness” would have bothered me more as a newer player and as someone who has never GMed than it does now.

Besides helping in a basic test such as the one we've been discussing, the dragon die helps in Dragon Age's version of skill challenges. Called "advanced tests," these are a succession of basic tests. The dragon dice are tallied for each basic test and the test is completed when the sum surpasses the threshold for the test. Contrast this with the often arbitrary nature of 4e skill challenges which tend to fit the formula of x success before y failures.

During skill challenges, players often get upset when multiple exceptionally high rolls don’t get them any closer to overall success than the rolls that barely succeed. They may have rolled high enough 3 times in a roll to convince the king to give over his first born son, but they still have some number of successes to rack up before he’ll give them the support they need. Now the conventional wisdom is that the DM should consider ending the skill challenge early since it doesn't make sense in the narrative to continue the negotiation. However, this decision is rather arbitrary and/or subjective and may make some players uneasy. By using the dragon die result instead, players are better rewarded for high die rolls in a way they understand and predict.

So what are your thoughts of the dragon die mechanic, degrees of success or anything else I discussed here?

1. While the original article currently is unavailable due to some changes on the site, you can find a cached version through Google.

2. Now I am by no means the first person to talk about this, it's just that I'm only now at the point of learning about these other mechanics. For instance, Rob Donoghue has at least two great articles about this, Rich Dice: Force, Finesse and Fortune and Rich Dice Extravaganza.

Interesting mechanic. As a player and GM, I think I like the Dragon Die much more as a success or complication randomizer than as a degree-of-success determiner.
To take your example in paragraph 8 a step further, it sounds like even a *worse* overall dice roll (provided it *scraped* above the success mark) with a 6 on the Dragon Die would yield a better in-game result than a stellar overall dice roll with a 5 on the Dragon Die. Am I getting that right? If skill bonuses in Dragon Age work anything like those in D&D then, ooo, that would burn me up. I'd want the skill bonus that I spent resources acquiring (and my players' skill bonuses) to directly affect degree of success if degree of success is in question.
I think the Story/Fluff Randomizer angle, though, is Made Of Win, as the kids are saying these days.

I like this idea. When you roll 3d6, you have 3-18 (15 numbers) as possible results, but a more bell shaped curve as to what the average result is, as opposed to a d20 where every number has an equal chance of coming up. If you want to make it more "D&D" like, of course, you'd roll 4d6 (4-24, 20 numbers). I'm a little unclear as to the dragon die mechanic. Could you give an in depth example? It looks like you could fail on the dragon die, and still succeed, but just barely? So the overall number determines success, but with extra weight to the dragon die?

The example they give in the book is jumping a chasm. The overall result determines if the character makes it across. If the total 3d6 is above the appropriate threshold, he succeeds. From there, the Dragon Die is used to determine the degree of success.

A 1 could mean hanging from the opposite edge with the possibility of falling, a 2-3 could mean a crash landing that leaves the character prone, a 4-5 could mean the character landed on his feet, and a 6 could mean the character landed so skillfully that he flies over the chasm and lands on his feet two yards beyond the edge. (Game Master's Guide, p 17)

Just a quick comment to Brian. 4d6 is actually a resulting 21 numbers since both your high and low bracketing numbers (4 and 24 in this instance) both count so you end up with 21 numbers. Still very close to a d20 though. I can easily see myself running a game using 4d6-3 as a skill check mechanic on some of the more fun checks and giving the random 6 'dragon die results' as well as a 7th exceptional result from a '21' roll.

I could also see this as a way to reward players even if they fail. A resulting 6 on the dragon die in a failure could be perceived as a victorious fight in a losing war. It adds another way to randomize results in failures too. I am definitely going to add this into my game. Thank you so much for bringing this up. It is the first time I have heard of any mechanics from Dragon Age RPG. I think I might have to go check that one out =)

I haven't ever tried Dragon Age, but the whole cursing which die got which number brought back memories of D6 Star Wars; nothing got players annoyed as much as a good roll spoiled by a bad wild die.

Just want to jump in here on one of those who was in on Chris Sims' original dice heresy bruhaha to remind folks of something. Now, I don't know how Dragon Age handles anything as I've neither played it nor read it (yet). But the 3d6 mechanic has two mathetmatical issues over the d20 mechanic in general. Call this just a fun FYI for those not in the know.

First, the chance of success (meeting or exceeding a target number) is not intuitive. The chance of rolling an 11 or better is 50%, whereas the chance of a 12 or better is 37%. The chance of a 13 or better is 25%. The chance of a 15 or better is less than 10%. Thus, DMs have a hard time setting reasonable and balanced target numbers on the fly without using some sort of chart or table.

Second, because the chance of success is non-linear, the effect of any modification to the roll or change in the target number is very skewed. If you need to roll a 12 and have a +1 bonus, that bonus improves the chances by 13%. If you need a 15, the +1 bonus improves the chances by 7%. In general, if you are very likely to succeed or very likey to fail, bonuses and penalties have no impact at all. If the chance of success/failure is closer to 50%, bonuses and penalties have a very magnified impact. While this does make a certain amount of logical sense, it also means that the value of a bonus or penalty is very hard to guage and balance. A DM who gives a +2 bonus to someone shooting for a 10 or 11 is improving their odds by 25%. The same bonus to someone who is shooting for a 15 is 15%.

As I said, this is just a fun little FYI for those of you who like the 3d6 idea. A DM needs to be aware of the math involved when setting good challenges and bell curves are not intuitive compared to a linear progression (1d20) where every modification has the same impact and chance of success follows from the distance up the scale (halfway up (10 or 11) is about 50%).

@TheRedDM - The dragon die still counts towards the rolls total and is only used as a determining factor for degrees of success when there actually is a success.

I think the dragon die mechanic works very well. I liked it so much that I have changed to Dragon Age. The settings, monsters, and classes are a little thin but I basically meshed the AGE engine from Dragon Age with my DnD worlds, races, classes, etc. I just restat DnD monsters to dragon age structure. I have custom stunt tables as well. For my group the D6 dice factor really helps since my group is new and get many of the RPG dice confused.

Great article, but non-uniform distributions for uniformly distributed success rates are generally very problematic. The system would be more granular if only one die determined success and multiple dice determined degrees of success. Basically the dragon dice system is inverted from where it would be most practical.

I began to publicly address this issue about three months ago:

It's to Table-top gaming's detriment that very few designers have any background in statistics, otherwise this issue wouldn't be so ubiquitous.

Send feedback using the contact form or through twitter, @sarahdarkmagic.

Resources for FAQs



Syndicate content