Alea Iacta Est

The Dice giveth and the Dice taketh away!

Humans really like the idea of fate and destiny. It is ingrained so deeply into their culture, that there isn’t even a god in their pantheons to deal with it. The gods themselves are subject to fate. Gods influence the fate of humans, but the gods themselves often have a fate they can’t escape. Greek mythology is especially dramatic about this. But why are we talking about fate today?

Because humans don’t understand statistics. Heck, most humans even shy away from using fractions. I could rant on about why this is and how to fix it, but the point is, humans use fate in it’s many iterations (luck, fortune, destiny, prophecies) to explain what happened. If humans don’t understand why something happened to them, they rather get irrational, make shit up, burn someone they dislike. Humans can get really pissed if they suspect that their lives might be unimportant to the cosmos as a whole.

I’ll talk about how important statistics are another day, but today I’m talking about how useless statistics are in creating a good dice system:

Why Even Bother?

Why do we use randomizersin an RPG? It is after all possible to design RPGs without them (Amber the diceless rpg). Still the majority of RPGs and boardgames use some form of randomizer (be it cards, dice, spinners, whatever).

Because humans like the uncertainty of randomness. It tickles your brain in the exact right spots. Randomness sells booster packs, loot crates, casinos, lottery. People talk about that crit/fumble they roled last game night. It is exiting.

Amber works via simply comparing your ability score to your opponent’s. You can influence the outcome by changing the challenge to a more favorable stat comparison (I try to magic him dead instead of slay him) or by (de)buffing (teaming up, using tools).

The dice are on the table because players lean forward and follow the roll of the dice with their eyes in expectation. And still every wannabe indie and his mother are trying to use statistics to define what aspects make a dice system good.

Like how the dicepool or the D20 are the superior system because of statistics: OBVIOUSLY the dicepool is so much better because it has a normal distribution. And OBVIOUSLY the D20 is so much better because the GM knows every +1 means 5% of difference. And statistics are important, don‘t get me wrong.

What we want to know at the end is, if the dice mechanic helps in creating an enjoyable experience. Statistics can’t do that. You can determine the chance of success for any dice system. And you can make it follow any pattern of frequency of success. But statistics can’t tell you if your system is any good. Enter Emotional Dice.

Emotional Dice

Emotional dice deal not with what statistic says about our dice, but with what players expect of our dice. Humans suck at math and can‘t really deal with statistics. Lefaret explain the Goat-Zonk-thing please.

It is my pleasure! The Monty Hall problem is set in a game show. You have the choice between three closed doors (1, 2, 3). Behind one of the doors is a sweet sweet price McGuffin. Behind the other two are goats or other useless stuff. You choose one door. Let‘s say you choose door 3. Now the game host opens one of the two doors you didn‘t choose, revealing one of the goats. In our example number 2. He offers that you stay with your initially picked door (3) or you may switch to the other door (1). Whichever door you choose is then opened and you get what’s behind it.

The mathematical question is: is it better to stay, to switch or doesn‘t it matter. And there was plenty of debate about that. If you want to know what the best way of getting your box of dice is and why, go educate yourself. The answer is you should switch, because the chance of the other door is 2/3 or 66% success chance and your door is just 1/3 or 33%.

Statistics is often unintuitive. I reccomend the Book The Drunkard‘s Walk on more examples of how humans misjudge stuff like that.

Thanks Lefaret. See: for us as game designers dice are giving the players expectations. And humans are flawed in what they expect of dice. Otherwise casino and lottery wouldn‘t work. So when we design dice it tells you MUCH more when you look at what the players expect & feel instead of raw numbers. Now we can actually learn a LOT from videogame design in that regard (watch this, learn something). What matters is emotional impact of randomness. Let‘s look at two dice mechanics of successful games and try to analyse their emotional impact:

D&D 3.0/3.5

D&D established the famous D20 system. If you don’t know the D20 system please show me the rock you are living under.

The system works by rolling a twelve-cornered dice numbered from 1 to 20 and adding/substracting your skill rank, your ability modifier and special modifications (magic, situation, tool, etc.). The number you get is compared to a difficulty, if it is smaller you fail, if not, you succeed.

The crit and fumble system differs from edition to edition 3.0/3.5 goes like this: For attack rolls a 20 is an auto hit, a 1 an auto miss but in order to do double damage you have to “confirm” a crit by regularly hitting your opponent again. Some weapons got higher crit chances (like 19&20) in which case you had to regulary hit with your 19 in order to get to confirm your crit.

4th edition rules: if you hit regulary with your 20 you crit, if you wouldn’t normally hit with a 20, a 20 is just an auto hit. 5th edition finally said: YOU JUST FUCKING CRIT ALREADY! And the reason for that is that the confirmation roll is an utterly disappointing mechanic: You roll a 20 YAY, but OH NO, the confirmation roll takes your crit away from you. I‘ve recently stumbled about something similar in a Dev Diary on Asylum.

D&D is a hero game, so they dropped that. For a horror game you might want to have a mechanic like that. Though I don‘t like horror and am thus unfamiliar with them. Maybe it just creates frustration across the board regardless of the tone of game.

But let‘s look at the default D20 system. You roll a D20+Mod vs DC. Well duh. But how does this shape the game. Well unlike dicepools you can keep adding bonuses for a longer time. It is not uncommon to start with a +6 in something you‘re good at in 3.0/3.5. And you grow quickly, more than +1 per level. Such growth isn‘t quite possible in dicepool systems because you can‘t expect the player to empty buckets of dice onto the table.

There also is one new addition to the D20 system in 5th edition: Advantage and Disadvantage. If the odds are in your favor (the situation is dope) you roll two D20 and discard the lower. If you have a bad feeling about this you roll two D20 and discard the higher.

Mainly this does away with tables upon tables of adding and subtracting modifiers. Look at this tables, would you?

Plus cover, plus concealment, plus buffs and debuffs etc.

I hate the situation in which a player goes „24, oh wait I have higher ground: 25… Wait I forgot about my shaken: 23!“. D&D 5 streamlines this in favor of speed of action resolution. And you couldn‘t just move to a „you get +2 if you have advantage and -2 if you have disadvantage.“ Because that would not be obscure enough. That would be always 10%. Players would argue „is the ground being a bit slippery really just as bad as being literary on fire?“ And no, of course it isn‘t. The system gives a shit about realistic granularity in that regard. But 2D20 are obscure to players. It isn‘t immediately at the top of your mind that the blind, burning, prone, flanked, surprised, fleeing coward is at the same disadvantage as the guard standing in the rain.

The mechanic is also great because it feels to the player like they lost/won due to the second dice. If you roll low and high that damn second dice saved your ass/insured your doom. Statistically this is bollocks but as EC said in above video the job of the game developer is to create an illusion for the player and for this we have to accept that human perception is broken.

Meanwhile every Joe Shmoe indie dev is going to tell you that (Dis)Advantage is the worst mechanic ever because it‘s statistical impact is changing depending on the DC in the situation as if that matters. When was the last time you heared a player complain about that fact? You can wait for the four riders till you hear that from any player. What matters is emotional impact not statistical impact. It is far more important that the situation feels good for the player than it actually being good for the player.

Shadowrun 4

Shadowrun 4.0 uses a relatively boring d6 Dicepool. Attribute+Skill+Stuff in D6s are rolled. 5&6 are successes. The number of successes is compared against the DC.

It is not unusual in Shadowrun to roll more D6s then your hands can hold.

The interesting part comes in combat. The attacker rolls to hit vs the defender and every surplus success the attacker gets, raises the base damage. Then the defender rolls damage resistance to reduce damage and then notes the remaining damage.

By this rising and falling damage the situation appears more threatening than it actually is. The GM can threaten you with 10 damage (which would kill your character) and the character can survive because you reduce it afterall. This is however at the cost of resolution speed and can make combat annoyingly slow.

Principles Behind Emotional Dice

Before we move on to analysing what principles lay behind emotional dice, let me stress that I think the most important thing you can do is try a lot of Dice mechanics. Use statistics to analyse them if you want but also try to get a feel for them by trying them actually out, not just reading about them.

If you need a place to start: Try Fate, Burning Wheel, FFG StarWars, 7th Sea, Splittermond, if you speak German.

A 5 is close to a 6. Statistical bollocks but emotional truth. A player expects a high number to be good. I’ve shown above how D&D criticals fucked that up and how Shadowrun uses it. How often a system fucks up this is generally not as important as how bad it fucks it up when it does. The human brain is an emotional memory chip. That means hundreds of unemotional slightly off rolls don’t matter. One stolen crit however will stick with your player and haunt him and bring him to hunt you down and call down otherworldly powers upon you. So don’t design such shit.

Principal 2: Complexity, Depth, Speed

As always in design you pay for depth with complexity. Make sure whatever you add is worth it’s increase in complexity. Good Design is not when there’s nothing left to add but when there’s nothing left to cut. As a rule of thumb, the more information the outcome of a dice gives you, the longer it may take to resolve, but you always want the fastest resolution possible.

D20 resolves yes/no questions, it gives you one info and thus has to be as fast as it is. The Shadowrun dicepool also determines damage (degree of success) and takes longer. The FFG StarWars dicepool tells you not only degree of success but also if there are random story complications or advantages to be build upon. You basically roll two dicepools at the same time.


People like different stuff. Whatever you do to make your stuff likeable isn’t going to cut it for some people. Some people will always just want their D20, their D10 dicepool, or whatever they are used to. Usually it’s the first dice system they played when they got into the hobby. Humans hate change and love nostalgia. On top of that it takes practice to learn a new dice system. It feels faster and simpler (Principle 2) if the player is already used to it, because of this practice.

Along with this goes that there is no “ultimate dice system”. Every system has certain strengths and weaknesses. Depending on what focus your game takes, some dice systems will fit well, others won’t. And no game can focus on everything.

Try to stay away from multiple dice mechanics within the same game in the hopes of optimizing the dice for each specific situation. That is usually a very bad depth/complexity ratio (look at Ad&d if you want to see a system which does that).

E.g. Dis/Advantage: This Concept discourages preperation of the situation. The best you will get out of a situation is advantage. Your players get nothing more out of creating multiple advantageous aspects of a situation than just creating one advantage for themselves. For D&D that‘s fine, because the game is about kicking down the door of the dungeon and killing the dragon that is inside.

Principle 4: Smoke and Mirrors

To be honest, I’m not 100% certain ab out this one, but I’l present it anyway:

What is most probable of the three? Which is least?

With a dicepool of 5 D6s: To roll 1 success

With a dicepool of 5 D6s: To roll 2 successes

With a dicepool of 6 D6s: To roll 2 successes

Answer: It’s the same probability, 32.9218106995885%

As shown with the above example of the D&D 5e (dis)advantage mechanic, most Players can’t directly read dice as hard numbers. This can be dangerous, because the GM has to have an understanding of the working behind the curtain.

This obscurity can be used to help with suspension of disbelief. When you put a hard number to something people start wondering if that number is really that plausible. Human brains usually don’t question stuff they don’t understand, because they don’t like to admit to not knowing stuff.

How Does This Help Me?

Emotional Dice forces you to state your design goal. While statistics play a crucial role in analyzing and building dice systems you want to know what the dice do to the players at the table. Often when I read about Dice mechanic Design online it boils down to talk about statistics. The Design Goal of the Dice is usually inherently connected to those statistics often in the terms of easy to read, intuitive design, display of realistic chances. However I propose that Dice should be there to fuel the experience you want to create. Realism or intuition can be a goal. But they don’t have to be. Chess is not realistic. Dice Pools are not intuitive. But both are there to create an experience.

If you are designing a dice mechanic for your system ask yourself these to questions:

  1. What does this aspect of my dice mechanic symbolize? What does it stand for?
  2. What is my game about? What feelings am I trying to evoke?
  3. Are those two in synchrony?

You should explore other games and their dice mechanics. Would your dice mechanic make sense in their games and why (not) so? Unlike statistics there aren’t any hard answers regarding emotional dice, after all they are at the heart of it about feelings.

There are more aspects to consider when talking about randomization in RPGs, especially statistics for which I recommend Torben Mogensen and Chaos engineering from the below reading list. But for today I’m done.

Further Reading:

An analysis of Dice Mechanics by Living Myth RPG

Ask Angry: I’ll Keep My d20 by The Angry GM

Common Dice Mechanics by Aramis

Designing for Emotion by The Game Kitchen

Dice-Rolls in Role-Playing Games by Torben Mogensen

Several articles on Dice by ChaosEngineering

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.