Dice and Emergence

By | November 2, 2013

Hello once again!

One of the issues we’ve wrestled with here at Room 209 Gaming is the type of resolution mechanic to use for various systems and subsystems.  While I myself have always been partial to a unified d20 mechanic, Ray and Sarah have encouraged me to consider other die mechanics such as multi-d6 for better bell curves, variant die sizes for different things (similar to Savage Worlds), un-unified mechanics (quelle horreur), and once or twice even cards have been proposed.  But it wasn’t until recently that I was able to articulate why they felt inappropriate.

Using a d20 offers “large margin of failure.”  This is a thing that many players dislike.  Players, if they have certain powers or abilities, or if they have decided upon a certain course of action, generally want to succeed at that course of action.  They don’t want to feel like their shining moment in the sun has been robbed from them.

That is why many game systems are making the dice less important.  I say less important because gameplay and results are becoming less dependent upon a good roll of the dice.  Failure, of course, is also less damning when the dice are less important – it’s increasingly harder to get “one-shotted” if the results of the dice are more mathematically stable.

This has been a trend for a very long time now, and there are a variety of ways to do it.  Rolling summed pools of dice is one very good way to generate a nice bell curve, where most results are “in the middle” and you suffer dramatic failure or dramatic success much less frequently.  Another way to make the dice less important is by rolling a metric crapton of them such as in World of Darkness or Tenra Bansho Zero, which both operate under the principle of “throw enough at it and you’re bound to win a little.”

Another popular way to do it is “yes and yes-but” (also called “yes and yes-and” and fail-forward design), wherein your results are either a win, a partial success, or a failure.  This is common in more story-focused games, where partial successes allow the introduction of new complications to a situation, but don’t result in preventing players from keeping on keeping on.  “Yes and yes-but” has been very popular of late (both Fate and Dungeon World, two of the biggest recent games in gaming, use variations of this mechanic).

The traditional method of making the dice less important, used by Pathfinder and D&D, is to add bonuses until it doesn’t matter what you roll.  Once your bonus is bigger than the actual number of the sides on the die, the die’s not quite so relevant anymore.  This method suffers, I think, from an “arms race” mentality; players and GMs alike keep pushing the numbers higher and higher because each sees the other pushing the numbers higher and higher.  It’s a vicious cycle.

And one of the latest new games, 13th Agemerges fail-forward and bonuses together with its use of the Escalation Die, which provides an ever-increasing bonus to be added to rolls as combat continues.  This still suffers the same issues as bonus-based gaming (and all the mathematics involved therein), while adding an additional complication and some extra lookups in combat.

One thing that I absolutely didn’t want to do is let the dice fade into irrelevance; this is something I’ve talked about here before.  We’ve dramatically reduced the number of overall bonuses to the dice, so that the results of the dice always matter.  And we’ve stuck with the d20 precisely because it’s so damn swingy.


Because we like supporting emergent roleplaying, roleplaying wherein the story is not what anybody was expecting.  And we’re trying to support this style of storytelling and roleplay from a completely different angle that other approaches have taken.

Fail-forward design, for instance, ultimately makes it very hard for the characters in the game to actually legitimately fail.  And without that risk, without the risk of disaster, the stakes aren’t as high for the players to “get it right.”  Caution gets thrown to the wind, and gaming becomes a story not of heroes overcoming obstacles, but of reality-TV show dramatists overcomplicating everything in their lives.

This is a bit of hyperbole, which I am prone to, but that is definitely the feeling I have gotten whenever I have played a fail-forward game.  It doesn’t feel as immersive, or as emergent, so much as it feels a bit scattered in the way it races from plot point to newly-created plot point in a whirl of sound and color.

Is it an interesting game design, and does it lead to interesting games?  Sure!  But they’re not the kind of game that interests us to build.  So.  How to handle the bad-dice situation of “the plot can’t go forward without this die roll succeeding,” while also providing meaningful and genuine failure consequences?

The first is actually pretty easy.  Kill “detection / observation / spot / search” or whatever your intended “find a clue” mechanic is.  Is there a clue?  The players find the clue.  The story moves on.  Nothing that is required to happen for the story to continue should be based on a roll of the dice.  Do the players need to get the clue that tells them where to go next?  They get it.  Do the players need to forge a key out of sky-iron to get into the fortress?  Let them find some sky iron and let them find a craftsman who can make keys.  That sounds like a story in and of itself, no?

That’s easy.  It’s the easiest part of not blocking players and stories.  The hard part is guaranteeing the dice aren’t cooperative; that’s where the d20 comes in.

There’s lots of chatter about the relationship of bell curves in dice mechanics.  Many people prefer summed multi-dice rolls because they are more likely to provide a result closer to the statistical mean.  This provides them with consistency, and consistency is comfortable.  But what’s the point of being comfortable when playing a game?  The entire point of making something a game is stakes!  There are lots of ways that failure itself can be uninteresting, but the presence of failure is what makes winning important.

Consistency leads to a thought process wherein you already know you can win–your modifiers plus your probably-consistent dice roll will lead to success.  But there’s no point in rolling if you already know you’re going to win.  This reminds me of the SNES game Earthbound (Mother 2 for purists), wherein if you encountered an enemy that you were way more powerful than, it simply ran away.  Consistency is the enemy of emergent roleplay, because if you know the results you can’t not know the results.  And emergence is dependent upon not knowing what’s coming next.

But because it’s so swingy, the results of a d20 are inconsistent and can’t be depended upon.  There is no bell curve to the results of a d20.  But in the design of our numbers, we’re building as if 10.5 is the typical result (it is in this case the average but, since there is no bell-curve distribution of values, it is not the typical result).  This number design, combined with the reduction of bonuses, makes the binary success or failure of any given roll up to the results on the die.  Failure is still an option but, because nothing that is story-required is based on die rolls, it cannot stop the game.  It can only stop a particular action.

We’re calling this “Fail? Next!” design.  If you fail (say, for instance, you are confronted with a lock your group can’t pick), that course of action is blocked to you, forcing you to try a different course of action.  Do you go back to town and strike up a bargain with the thieves’ guild to let you borrow one of their expert lockpickers?  Do you try and climb a wall and break into the building you’re trying to enter a different way?  Do you strike up a loose friendship with someone who works in that building, and then pressure him or her give you a tour of the place?  Do you get angry and just kick the door in, risking announcing your approach prematurely?

Fail-Next design encourages oblique thinking, wherein the players (and the guide, since rolls are made in the open) must look for alternative potential solutions when the direct solution or the solution they were considering is met with failure.  And “finding another way” is very much within the realm of emergent roleplay, wherein the story can potentially be something that nobody was considering.

The world is open to possibilities.  How will you meet it?