I’ve been thinking about dice mechanics as we develop Forthright Open Roleplay. The first pen-and-paper RPGs I played were GURPS 3rd Edition, Advanced Dungeons & Dragons 2nd Edition and some of the original White Wolf games. So I tried out 3d6, d20 and dice pool mechanics all within a year of joining the hobby. The first games I made as a kid used notecards and a d6 from my Garfield boardgame to add a little something to the Choose Your Own Adventure books I read.
That “little something” was random chance, even if I was too young to understand all the implications. Dice provide an element of randomness to an RPG for situations where the outcome is uncertain. The question is just what types of uncertainty provide the best feel for a particular game.
Some players strongly prefer either 3d6 or 1d20 when it comes to task resolution. The bell curve provided by 2d6 or 3d6 provides more consistency than the linear distribution of a d20 or d100. That consistency in results is why ability scores are often generated with a bell curve: it is more likely that the results will fall around the middle, making the extremes of 3 and 18 stand out more.
The chances of rolling an 11 or higher on a d20 are 50 percent. The chances of rolling 11 or higher on 3d6 are 50 percent. So, why not use a d20 for generating ability scores or 3d6 to determine the success of an attack for a game of D&D? Because raw percentages don’t give an accurate picture of how dice impact gameplay.
Some people argue that if the odds of success are 50 percent, it doesn’t really matter what method is used to determine the result. A coin flip could do it just fine. You could try each resolution method 10,000 times and probably produce something very close to the 50/50 split we would expect to see.
But no one rolls that many times in a session. There might be three or four important rolls per player during a combat. And for some players, the expected average result of a bell curve provides a kind of mental cushion to make the game feel more under their control. For others, the lack of an expected outcome makes risk-taking more meaningful.
There is no perfect die mechanic for an RPG, but the types of dice a game uses and how it chooses to use them can be combined to generate a specific feeling over time. Players notice when these mechanics reinforce the themes of a game. Call of Cthulhu is known for how deadly it can be for player characters, so it’s no surprise that the mechanics use a linear distribution (d100) to determine success.
Skill systems expose sore spots in an RPG when it comes to randomness. Games that include numerous skills that are resolved with linear distribution encourage players to pile on the bonuses. That’s because players expect high levels of competence in professional work — a beginning baker who ruins 45 percent of his bread would be viewed as incompetent rather than amateur.
A bell curve fits how many people view day-to-day work as something that has a predictable outcome (i.e, someone hired for a job can likely perform it adequately most of the time). But that lowers the odds of extraordinary outcomes that change gameplay. And rolling dice for what is expected to be a non-random result can get boring fast.
Failure is a major part of a roleplaying game, because setbacks force players to approach problems from a new angle. Mechanics that lessen the chance of defeat run the risk of pushing players even further into niches. The end result is each player is the master of one thing and sits around bored until that task comes up.
Modern games place greater emphasis on using the same mechanic for all tasks to simplify gameplay, but that can expose edge cases where randomness promotes incoherent outcomes rather than interesting ones. That’s why many systems remind players and gamemasters to only roll when there is significant danger or penalties for failure. If the character is a baker, just say yes when the player wants to make a cake.
Of course, most RPGs don’t actually use just one resolution method or the other for everything. D&D has linear distribution for attack rolls, but the multiple dice of damage that are rolled over an entire combat form a bell curve that determines when someone is defeated. GURPS can combine bell curve dice rolls with particularly deadly combat rules. Games with spendable tokens explicitly provide a structure so players can decide when the odds should be changed in their favor. A Hero Point can turn the death of a player character into just a flesh wound without having to change the dice mechanics at all.
That’s why I disagree with those who say either bell curves or linear distribution are always better for all games. The right mixture of dice and rules make a game that feels “just right” for the players who enjoy it. And that’s really the goal of all this design talk. Mathematical balance is good; player enjoyment is better.
How do your favorite games mix-and-match dice and mechanics to make it fun for you? Let us know in the comments below.