The roll of a dice is not a casual thing. It is the fulfilment of a contract between the game and the players.
In many cases a perfunctory toss of some dice is given no consideration by players and game designers alike. However, the dice resolution mechanic is the underpinning of the entire endeavor of a role-playing game it cannot be brushed aside as trivial. Every rule mechanism rests on the consequence and accuracy of what the dice roll symbolizes. The dice represent the structure of reality in meaningful ways.
When we are asking a player to roll a dice we are asking a question, and using the dice to return an answer. I have seen many role-playing systems where the question is not clear and the answer even less so. In these games the dice roll represents many things at once. The gamers who devised these systems, and trust me they are gamers in every sense of the word, are excited by a “clever” dice mechanic that seems fun. They enjoy interacting with the dice more than the imagining what the die roll means.
In this role-playing game the action roll asks a question…”how well does your character perform this action?” The player rolls dice and adds two of them together to answer the question. As they get better scores, they can roll dice with more sides. They roll dice with one less side than the score so that the results of the rolls cluster around the character’s score in that endeavor. We start with a peaked distribution of results, and drift towards a shifted bell curve once we add a 3dx, take the best/worst two mechanic. This is my first point. The distribution of results is not linear. The game is structured to answer the question “how well do you perform this task” by returning a plausible set of results.
The world’s most popular role-playing game uses a simple mechanism also. A d20 is rolled and a set of modifiers are added. This gives us a list of results where every outcome has a 5% chance. As a natural “20” is a universal critical success (except on skill checks) and a “1” is a universal failure, these two results remain fixed regardless of how the table of results is shifted by the bonuses. At the table a natural “20” is highly celebrated and generates lots of fun. In the fifth edition, the designers did two excellent things: they recognized that the shifts from the bonuses were too large in previous editions and constrained them so that the mechanism didn’t break down too soon, and they introduced a roll two, take one mechanic. I wonder if I had any part in this, as I had been espousing the roll three – take two mechanism on the internet for decades. The roll two take one mechanism has some odd flaws when it comes to probabilities making the natural “20” or natural “1” the most common result. Distribution curves bulge in the middles; they aren’t slopes that favor the extremes.
So, why does any of that matter? If it warps reality it warps the story. In a world where something extreme happens 5% of the time, regardless of ability or talent; the common man is induced to attempt things that are not plausible. The GM is now forced to constrain absurdity so that Gary Kasparov is not beaten by Doug in a game of chess. You might want to point out that reality is not fantasy. There is a difference between a plausible fantasy and an implausible fantasy. Implausible fantasy is not consistent within its own framework. Absurdity can be fun; Monty Python is brilliant, but even that is not carelessly crafted. This core distinction seems to escape many. This game is not written for them. They will not understand the motivation behind the endeavor, but I do hope they can still enjoy some of the other mechanisms and elements I have created.
So, my first criticism of a dice mechanic is that it must return a plausible set of results. Our most popular RPG doesn’t quite deliver on that point. However, it delivers the fun of the critical roll, and it does a good job in framing the question when it comes to what I would consider an action roll: attack rolls and skill checks. It frames these questions in a very similar way to ours: “how well does my character perform this task?” This creates a direct link between the toss of the dice and the internal story. They roll the dice and conjure an image in their mind of what happened. The narrative is missing the other side of the coin, however.
Reality intrudes upon talent. Even monkeys fall out of trees. When a player rolls the dice and the result tells them how well their character has performed an action, like shooting a bow. This should be immutable. A specific die roll should mean one specific thing. However, hitting the target is sometimes dependent on more than just skill. After the arrow leaves the bow it can be acted upon by things out of a character’s control. The target can move in an unexpected way, the wind could gust or lull, another object may intervene. This can be modelled by rolling not only the action, but also the difficulty. Rather than accepting a static difficulty target, a randomized difficulty target represents the challenge of the situation. It allows for extreme results, while skewing towards plausibility and narrative consistency.
When a character jumps onto a crate and the GM rolls a difficulty of ten, the player knows that the crate is slipping or breaking because the difficulty threshold is unusually high. They roll the dice to perform the task and can interpret the results immediately. This is the power of clarity. Rolling the dice on an action roll always means one thing: how well does the character perform the task. The threshold difficulty always means one thing… “what is the barrier to any level of success?” The difficulties can change to represent what is happening independent of the character. The separation of these concepts is essential to being able to build off the rolls in the other rules of the game.
Imagine a standard dartboard. Excellent successes, like the treble-twenty, are adjacent to poor results like the “1”. The skill of the throw indicates how well grouped and clustered the throw is, but it is subject to the vagaries of the target – a shifting difficulty that imposes a randomization that can thwart the expert and elevate the novice. Over time the expert prevails, but the immediate odds are able to shift for inspection.
This brings us to the next pillar of consideration: the inspection level. A single event has a wider swing of probabilities than a set of events. A game system must be able to recognize and assist the GM with these differences. Am I measuring an isolated event or a series of events? When measuring a series of events we need to roll more dice. In our system we can roll additional dice and count the best two. This gives us a shifted bell-curve of results that is consistent with an imagined reality.
The less-popular role-playing games often invoke rules where players arrange or manipulate dice. This draws their attention away from the shared story and has them engaging with bits of plastic, wood, or metal and strategically thinking about artificial mechanisms. I am all for strategic thinking, but such thinking should exist in the shared world, not external to it in some concocted schema that has no relevance to the characters of the story and does not support the reality that is being imagined. No matter how clever these mechanisms are they are nothing but a distraction… a side game that takes focus on the action at the table. Soon the dice become the game.
We have a second type of roll in our game: the effect roll. While it is mechanically similar to the action roll, it is asking a different question...”how much am I able to alter the world with my action?” All other role-playing games have effect rolls too. Damage is a roll that is ubiquitous.