Core Mechanics: Randomization

Almost all role-playing games use dice to determine the outcome of actions. The ways that dice are used varies widely though and can have a large effect on the mathematics of a game system as well as how it feels at the table.

Dungeons & Dragons with 2d10

As an example, using 2d10 in place of 1d20 in Dungeons & Dragons would cause a ton of changes to the game despite the similar range of outcomes (1-20 vs 2-20). The average result of the roll stays about the same (10.5 for 1d20, 11 for 2d10), but the odds of rolling that average doubles when using 2d10 (5% of 1d20 rolls are an 11, 10% of 2d10 rolls are an 11). The odds of rolling values at the extreme edges of the range also drop (5% of 1d20 rolls are a 20, 1% of 2d10 rolls are a 20).

These changes to the probability curve have a big impact on the game. To demonstrate that, let’s look at a first level fighter’s chance to hit a range of enemies. Assuming the fighter has a +6 attack bonus:

• Goblin Sniper (AC 13): 70% with 1d20, 85% with 2d10
• Kobold Tunneler (AC 15): 60% with 1d20, 72% with 2d10
• Ogre (AC 18): 45% with 1d20, 45% with 2d10
• Troll (AC 21): 30% with 1d20, 21% with 2d10
• Flesh Golem (AC 24): 15% with 1d20, 6% with 2d10

As those numbers demonstrate, using 2d10 rather than 1d20 would make it easier to hit opponents that you outclass while making it harder to get lucky and hit an opponent that outclasses you. This would have a big effect on the game because lower level opponents would become unchallenging more quickly and higher level opponents would be too dangerous to introduce without some serious re-balancing of the system’s advancement math.

Using an Even Distribution

Rolling a single die has an even chance of generating any number in its range – your odds of rolling a 20 are the same as an 11 on a d20 for instance. This even distribution of values means that the probability of rolling a certain target number or higher changes in a linear way. Any +1 on a d20 results in a 5% increase in rolling a fixed target number (up to point where success is guaranteed).

For a game, even distributions offer a couple big advantages. First, its easy for players to quickly figure out the odds of hitting a certain target number – the math is always just dividing the number of successful results by the number of sides on the die. Second, it helps to make game balance easier because giving a bonus almost always has the same effect on the game’s probability. For example, if I want to create a power that gives a bonus to hit in exchange for less damage potential, I know that a -2 attack penalty means a 10% less chance to hit on a d20 and can then balance it with a 10% boost to damage.

Using a Curved Distribution

Rolling several dice and adding together the results causes a curved distribution where you are more likely to roll numbers near the average result than those at the extreme ends of the ranges. This means that the probability of rolling a certain number or higher is non-linear – a +1 bonus results in a different percentage increase in the probability of succeeding on a roll depending on what number you need to roll to succeed.

For a game, this distribution tends to better mimic reality where you are likely to perform near your average result rather than having a flat distribution over a wide range of results. A curve means that most of the time a roll will be about average for a character while still leaving a small chance that the character gets lucky (or unlucky) and ends up with a result far from average. The biggest drawbacks of a curved distribution is that their probabilities are harder to quickly estimate during play and they can make balance trickier especially around stacking bonuses (especially for rolls with large numbers of dice and steeper curves).

More Than a Random Number Generator

One neat trick in some games is to use specially colored dice or dice of different sizes in order to allow extra pieces of information to be generated by a roll rather than just a single number. For example, in the AGE system you roll 3d6 for most rolls, but one of the dice is a special color. This allows each roll to give you both its result to compare against a target number and a margin of success determined by just the value of the special die. Alternatively, the extra information might not have a mechanical effect on the game but could instead guide its story. For example, in a Cortex Plus game, you roll a pool of dice where each die comes from a trait of some sort. You then pick two dice to add together for your roll’s result. Because the dice are tied to traits, the dice that you pick can guide the story and explain how you succeed at your task.

Alternatives to Dice

While most role-playing games use dice for randomization, there are other options. A deck of playing cards is effectively a d13 where the suite of the card is a convenient way to add extra information. Leaving jokers in the deck could add an extra twist. Do: Pilgrims of the Flying Temple is an example of a game using something other than dice as its random element; it uses black and white stones drawn from a bag to determine the flow of the story.

Playing Without Randomization

Finally, while almost all role-playing games use some sort of random element, there are a lot of traditional games without any randomness. For example, neither checkers or chess use any random numbers. In a role-playing game, you could forgo randomization but still make the game interesting and fun.

For example, imagine a martial arts themed game where players have a range of fighting styles they know. Each style beats some styles but loses to others. During each round of a duel, the players involved simultaneously pick the style their character is using and that determines the winner. Effectively this is a game of paper-rock-scissors, but it potentially has far more than three choices.

What Are Your Favorites?

Which dice systems are your favorites? Have you played any games that use a different method or randomization or eschew random elements altogether?