Monday 11 November 2019

Theory post - Why We Have Dice Rolls, Game Mechanics, and Stats

Note: some games don't work like I'm about to describe here, and you engage primarilly with the game as a mechanical construct. I have no interest in them, and they aren't covered by this little essay. Diegesis as a concept is explained here.
In this essay, I'm probably repeating a lot of points that have been said elsewhere. Often I can't remember where I first read an idea. This is me explaining all these ideas in one convenient place.

So. I will explain in this essay why we have dice resolution and stats, at least in the sorts of games I consider worth playing.

To begin with, RPGs work on what I've seen called 'the conversation loop'. Here's how it works:

  1. The GM explains the situation that the PCs are in. Players might ask questions for clarification on the situation at this stage.
  2. The players say how their characters respond to that situation.
  3. The GM adjudicates and explains the outcome of that response; what happens, and what the new situation is.
  4. We loop back to 1.
It's very simple. GM gives a situation, players respond, GM adjudicates the results, repeat. This covers the bulk of play in my experience, and is the basic activity of tabletop rpgs. Here's an example:
GM: You are in front of a closed door in the dungeon. You can hear running water on the other side of it.
Player: I want to kick that door open!
GM: Sure! It's not locked, so they door swings open. On the other side, there's a room with a fountain in it.  
Easy, right?
This method gets used for all sorts of things; handling exploration, encountering hazards, conversing with NPCs, all sorts.
At this stage the game is entirely about engaging with diegetic things. The situation is a description of diegetic facts about the world, the players response is a diegetic response to that, and the outcome is likewise more diegetic facts.

Now, let's zoom in on "3. The GM adjudicates". This is where things get tricky, and GM skill really comes into play. Here, the GM has to assess the situation, what the players are doing, what they want to achieve, the PC's capabilities, etc, and decide what they think the outcome of the PCs actions should be. Obviously, there's a few things the GM needs keep in mind here:
  • The results need to make sense based on the situation. IE the response shouldn't be a total non-sequitur. 
  • The results need to allow the players agency (IE not render their choices meaningless).
  • The results should seem fair, or the players will disengage.
  • The results can't always be 'the players get what they want' since that would rapidly render the game boring.
There might be factors at play that the players don't know about (because finding all the relevant info before making a big decision is a skill, and sometimes players mess up), and sometimes the players have put themselves in a situation where they're fucked (because their actions have consequences, and again, sometimes they make bad choices). But those four points are important.

So then. Some actions are easy to adjudicate ("You open the door, on the other side, there is...") but others less so. Often, an action will be difficult to handle well through considering the diegetic facts and making a conclusion based entirely on in-fiction factors. For example:
  • The action's outcome is uncertain but relies on factors hard to model through the conversation loop. 
  • The action is an area which the players and GM have little IRL experience of to draw on, making it harder to engage with in purely diegetic terms or make a good adjudication.
  • The action has potentially very serious consequences (such as death or major losses), and so simply making a (possibly arbitrary seeming) decision will easily seem unfair.
  • A combination of some or all of the above: violence, for example, hits all three conditions except when it's very one-sided.
In these situations, we use dice rolls and game mechanics to make a decision on the GM's behalf when the outcome is otherwise in doubt and hard to adjudicate. This results in a conversation loop like this:

  1. The GM explains the situation that the PCs are in. Players might ask questions for clarification on the situation at this stage.
  2. The players say how their characters respond to that situation.
  3. The GM decides to use dice to resolve this action (for one of the reasons above), and rolls dice, or asks the players to roll.
  4. The result of the dice roll is interpreted. The GM adjudicates and explains the outcome of that response; what happens, and what the new situation is.
  5. We loop back to 1.
As an example: 
GM: There's a closed door in front of you. You can hear running water on the other side.
Player: I kick the door open!
[GM thinks: this door is latched shut on the other side, and the player might not be able to kick it down easily: they decide to use dice to resolve this] 
GM: Roll me a d20, if it's under your Strength score, you'll succeed.
Player: [rolls] A 16, so no.
GM: Okay, so after several solid kicks, the door hasn't budged. However, you have made quite a loud noise, which might be attracting attention...
Again, this is pretty simple, right? Situation, response, input from the dice, adjudicate results, repeat. 
The point here is that you use dice to resolve things when adjudicating in the normal conversation loop becomes difficult. It's a tool to abstract difficult or complex things, and to offload some responsibility to chance (which makes the outcome feel more 'fair' than if the GM just declared a bad thing happened). 
Let me repeat that.
Rolling dice is a tool for when it's hard for the GM to adjudicate the results.
Although the example I gave has the players rolling dice, this isn't the best example of how this tool gets used in my view. Something like a random encounter chance (for 'does a monster find you') or a reaction roll (for 'are these monsters going to attack you') is probably a better 'pure' example of this; they're tools for the GM to use when it could be hard to make a fair, consistent snap decision.
The other good example of a dice mechanic is Saving Throws. A saving throw is basically "Oh shit, your character might die because of this". You give the player a roll to see if they're fucked, since it feels less arbitrary and unfair when the GM leaves it to chance than when the GM just declares that you're dead. The various procedures of combat likewise use this; making chance a factor means it doesn't feel like the GM has it in for you when you get splatted by an ogre; it's not the GM's fault, you just got unlucky.

The question then becomes; which dice do you roll, and what do the results mean? As a general rule, the more likely a thing is, the greater the chance that the dice will say 'this succeeds'. 
You also want to keep the dice rolls fairly consistent, with recognisable patterns. This means that:  

  1. your players won't be wildly confused whenever the dice get involved, speeding up the process of rolling and interpreting.
  2. your players can make informed decisions about their chance of success when dice get involved. 

IE: Consistently asking for roll-under-Strength (perhaps with a bonus or penalty) when the players try to break something will mean the players know it's best to let strong PCs do the vandalism, and you won't have to explain the dice roll afresh each time it comes up.

This, then, is why we have codified mechanics. They provide a rough framework that everybody is familiar with for when it's necessary to roll dice to adjudicate an action. They mean everybody is up to speed with how the rolls will work, and its easy to predict your rough odds of success.
It's worth noting here that, if you rely on dice adjudication too much, players will stop interacting with things diegetically, and start to think in terms of game mechanics. The same problem occurs, very much amplified, if your dice mechanics are too complex and prescriptive; you risk the mechanics taking over the game. The tail starts to wag the dog, as it were.
It's also worth remembering that dice adjudication is a tool for the GM. You don't have to invoke it, even if you did before in similar situations. You certainly don't have to use the exact same mechanic if it doesn't fit the situation; that would result in the game mechanics superseding the diegetic factors. For this reason, I (at least) prefer relatively streamlined and modular mechanics. Ones you can bolt-on as needed, hack on the fly, or even flat ignore, without the game coming apart.

This, then, is why we have dice rolls.
Why do we have character stats? There are three main reasons.
The first is that they define diegetic facts about the PCs. It defines in concrete terms things like 'Alice is a Thief and can do the sorts of things thieves do' or 'Carl can cope with injury much less than other PCs' or 'Dan is carrying this equipment' or 'Erica is not very clever'.
The second is that it gives a concrete framework for things that wouldn't have any real-world equivalent to draw on for adjudication purposes; typically magic. Magic could be capable of fucking anything (it's fucking magic), but character stats and mechanics give it limits. This ties in with the first point; it defines quite strictly the diegetic facts of what magic can do. You can cast this many spells and they can do this when you cast them.
The third is that character stats let you make more consistent rulings. If you know that Bob the Fighter has +1 Constitution, you can give him that +1 to rolls where his constitution is a factor, and it feels fair and consistent. If everybody has a numerical value for their armour, you can consistently give them the same chances to suffer a serious injury, and use that same value to assess things like how badly their armour weighs them down. 
If you aren't bothering with dice at all, points one and two matter; they still define facts about the PCs, that should be taken into account when adjudicating the results of PC actions. Sometimes, these diegetic facts are tight constraints (IE: "you cannot cast fireball, bob, you're a fucking fighter"). Other times, they'll just be descriptive (IE: "Well, Alice is a thief, so sure, it makes sense that she'd be able to sneak past the guards without being noticed.")
Point three is critical, though. Defining character traits lets you make more consistent rulings when you bring in the dice to adjudicate, letting your players make more informed choices about their chances of success at different actions. 

In summary, here are the three key points:
  • The game is about the conversation loop of 'situation, response, adjudicate result, repeat'.
  • The dice exist to allow you to adjudicate in situations where this would otherwise be hard.
  • Relatively consistent dice mechanics allow you to use dice rolls in a way that is fair, and lets players make informed choices about risk.
  • Character stats help inform the 'adjudicate' part of the conversation loop, and give you a framework for consistent dice mechanics.


  1. I like this. It lays out the basic argument and explains it well. I'll definitely be referring back to this in the future.

  2. Replies
    1. I'm not sure I'm familiar with the Lumply Principle. I've probably read it, but I don't remember. Remind me?

    2. Vincent Baker (of Apocalypse World fame) describes RPG as a shared conversation and has the Lumpley Principle as : "System (including but not limited to 'the rules') is defined as the means by which the group agrees to imagined events during play."
      It says very tersely what you are skillfully developping here (with the difference that the Lumpley principle includes GMless RPGs)

    3. ah, yes.
      I think I stole the idea of 'the conversation loop' from pbta, so that makes sense.

  3. Very nicely compiled and laid out. Thanks for another useful resource!

  4. Excellent! This is where I will point people who ask what all that weird nonsense about elves is if I fail to explain it properly.

  5. Thank you for such a well written article.It’s full of insightful information and entertaining descriptions.Your point of view is the best among many. D20

  6. Hi! Daisy-chaining off this thought process here, reading this and Necropraxis's got me thinking of the mythic/ritual elements of dice and chance etc.

    Let me know what you think, if you’re so inclined

  7. Okay that sounds good but I come from a pre-D&D wargaming environment and in my mind rolling dice is always a way of the DM displaying fairness (I have thoughts about the evolution of wargaming rules into role playing rules and the mutating role of the adjudicator but I won't bore you with them).
    Secondarily it is a way of injecting uncertainty into the story.
    I may use dice to adjudicate a result but the the reason I choose to use dice is almost never because it is hard for me to adjudicate the results.
    You actually state 4 key points at the end of your article If you drop the second point then I agree entirely.