Last Sunday night, I was the guest on Geekerati Radio. We talked about games and narrative, topics lodged firmly within the Gameplaywright wheelhouse if not also sown broadly about its idea farm. You can listen to the roughly hour-long discussion at blogtalkradio.com.
One of the interesting things that the show’s host, frequent GPW commenter Christian Lindke, asked me was to what extent the dice mechanics of Feng Shui contribute to its ability to express a satisfying story. Feng Shui‘s central mechanic, in case you’re not familiar with it, is to roll two six-sided dice, one of which generates a positive number between 1 and 6, the other a negative number between –1 and –6. The two results are summed, and the sum added to a skill total that varies in a range from 5 to 15 or so (for player characters, anyway). Obviously, much of the time, the dice effectively cancel each other out and the skill number stands as the overall test result.
Christian’s question, as we explored the idea, was whether I thought predictability enhances narrative.
It made me think back to an experience using GURPS as the engine for a pulp genre campaign game. I remember, at the time, being impressed a how well GURPS‘s 3d6 bell curve rendered the dispatching of low-stat grunts by heroes very predictable, while at the same time making an array of truly death-defying feats with high target numbers particularly hard to achieve. This struck me then—as now—as firmly in keeping with the pulp sensibility.
A game with a broad range of unpredictability in test results, on the other hand, can make it very difficult to tell a story that makes any sense as a story. For a low-level D&D character, for example, the d20’s overwhelmingly broad range of equally probable results wholly trumps skill values ranging from +1 to +5 or so. Even with skills at which these characters are most proficient, the random factor is overwhelming.
This isn’t quite fair to D&D, given that target numbers can be adjusted with relative ease to manipulate the probability of easy tests being successfully carried out. But it’s still the case that the random factor is a much greater part of a D&D skill check than a Feng Shui skill check, and this fundamental difference seems to color their suitability for narrative play quite a lot. And it’s hard to imagine a good story qua story—the kind of narrative that someone without an ownership investment in one of the characters would be interested in hearing—having massive unpredictability in whether its protagonists overcome relatively easy obstacles or fail to do so.
High predictability in RPG tests makes me think of the story-game community’s interest with the “Yes, but…” GMing technique. According to this idea (which I’m aware I’m presenting here as a much more inflexible dogma than it is actually taken by anyone), the GM avoids contradicting player statements about what their characters do, instead offering complications that are dramatically commensurate when the declarations get dodgy.
“I step up and slay the dragon with my sword.”
“Yes, but the dragon’s acidic blood melts your blade, and his death throes maim you.”
This player-driven capability to determine what happens in those narrative-centric games strikes me as similar (to a point) to traditional dice-driven mechanics where there is a high degree of predictability in test outcomes.
Predictable dice systems do seem to have a certain synergy with highly narrative games. But on the other hand, to suggest the truth of such a thing, one must account for the fact that the predictable narrative is also the deathly boring narrative. But-but, even though some of the best developments in narrative are surprising, even shocking, at the same time, it’s critically important that they make sense given what has gone before. “Surprising but inevitable” is the ideal to which storytellers aspire.
The bell curves of Feng Shui‘s system, and GURPS‘s system, emphasize predicability while also allowing for highly unpredictable outcomes in a small percentage of cases. In Feng Shui, results of 6 on either die “explode,” for example. The hero always has a chance to miss even the greenest mook. These possibilities, when wedded to a skilled GM’s ability to extemporize an explanation for the unexpected event, can make for truly excellent game-stories that accommodate the need for predicability in most cases but unexpected results in small doses. But when the GM is forced to justify senseless outcomes on a round-by-round basis, it quickly becomes clear to everyone that the game is generating something other than a satisfying narrative.
What do you think? Does predictability enhance a game’s narrative quantities? Fight against them? Operate on an unrelated axis entirely? Do tell!
The way I would put it is, people are unpredictable enough to more than make up for any predictability of the dice. Likewise, the unpredictability of the numbers produced by the dice mechanics is less important than what the game does with those numbers. In low-level D&D the whiff factor can be overwhelming to the point where luck seems to trump everything, but even there the cleverness of the players can climb on top of the heap. On the other hand in a game like In A Wicked Age, the dice leave a lot of the results of the mechanic up to luck, but you know that whatever the outcome, it can be something that changes the course the story takes. Likewise, Maid RPGs over the top randomness gives it a frenetic quality, and success and failure alike feed back into its general zaniness.
I think a player in a given situation has a range of outcomes that they will find acceptably dramatic and realistic- bearing in mind the tone of the game. It’s a very subjective thing. Someone’s favourite system is probably a system where the resolution mechanic closely matches their conception of what that range should be. If not all of these outcomes are possible, the system is ‘too predictable’; if it allows outcomes they don’t approve of it is ‘too random’. If they’re really lucky, they get to play with other people with similar ideas of range.
For me one of the primary purposes for randomisation is to shake the players (GM included) out of any complacency. If there’s a possibility that even a nominally routine encounter will suddenly go very bad, it encourages the players to pay attention and bring their A-game all the time. If there is ‘too little’ randomness, you have to elide the routine more often as it’s not interesting enough to warrant play time.
I’m an advocate of post-hoc GM reasoning, imbuing the dice rolls with meaning after the fact (and even incorporating the explanation as a modifier in future rolls), so I may well have a larger tolerance for randomness than many. Too little randomness limits the scope for this kind of improvisation, too much makes is necessary so frequently it loses its charm.
The outcome of an encounter should be primarily driven by what the characters do. ‘Too little’ and ‘too much’ randomness both minimise the impact of player decisions and actions. The subjective sweet spot gives the players the ability to balls things up in a way only good luck can rescue, minimises the possibility of dismal failure given excellent play, and offers satisfying gradation between the two extremes.
To agree with Ewen, I’d say predictability is most important when (a) you’re attached to a particular outcome, and (b) failure isn’t fun.
When I heard this come up on the show, I just about called in. The fact that you have a reliable zone of AV in Feng Shui is why the stunt system works, and why it’s not exactly penalizing players the way that post-Exalted players sometimes say it is. In theory, in Feng Shui, you should be able to design your character’s action for the turn and have a reasonable expectation that it’ll go off, once the AVs of enemies is figured out. That expectation isn’t there, even with a big dice pool, in Exalted. I’ve seen a lot of big dice pools come up zero. 🙂
I really want to contrast this post with Robin Laws’ post about Win Space today. But the time’s not here for me to do it, quite.
There’s a difference between the atmosphere and the practice, though. In practice, individual rolls matter, but instinctively the fact that my 4E Warlord is as likely to roll a 2 as a 20 is somehow irksome. I tend to come back to ideas that allow for players to spend a resource to improve rolls at the right moments.
But what’s a right moment? Is an RPG really just a gamble to see if you can get a game to match up with the story you meant to tell? Or is it a game of telling a great story out of the random pieces you’re given during play by die rolls and conflicting actions?
Expectations are certainly key to the impact that dice mechanics have on the narrative, though. I like the 2d6 resolution because of the way that it intuitively adds meaning to the less likely outcomes — where 9-11 on d20 is middling, a 7 on 2d6 is a sturdy, reliable value. It’s normalcy. It’s competence.
I’m reading too much into it, I know.
Eh, I don’t think you’re reading too much into it, Will. Randomness in a game can be really problematic, but it can also be a boon.
To more directly answer Jeff’s question, I would say that randomness does undermine a games narrative flow. This is of course only true if we agree that most games are intended to follow a general arc of plot, while allowing for some individual action and yes, randomness in terms of the exact path to the ultimate resolution.
The problem is when a player devotes creative effort to a task, be it in combat or otherwise. One of my favorite question to ask is “How do you do it?” hoping to elicit some cinematic description from a character. The problem is that this creative endeavor needs a reward to be meaningful when repeated. If a player declares their character cartwheels into the room while firing her gun at all the gangster goons, only to roll poorly and fail, then she is less likely to invest into the narrative creatively in the future.
At the same time, there needs to be some risk of massive failure or glorious success, otherwise it ceases to be a game with any meaning to it. (Not that collaborative storytelling isn’t a critical part of the experience…)
So, less randomness would be better. As it is, though, I’ve always stolen from Will and used some sort of resource mechanic to balance out poor rolls when playing in a D20 environment. To be honest, though, after thinking about it, I might move to take advantage of a Feng Shui style mechanic.
Well remember with Feng Shui there is also a built-in player expectation of HK “cinematic” reality with all the resultant acceptable over-the-topness (for good or ill) that goes with that.
And Jeff? The ability to Take 10 or Take 20 certainly adds a level of predictability to post 3.0 D&D…
I’ve been thinking about the bell curve for some time now, having once been a devote of GURPS and finding the d20 mechanic overall disappointing because of it’s lack of bell curve.
It took playing more 4/e to finally understand what a linear system offered that a bell curve system doesn’t. And that’s what I’d like to touch on for a moment: Uniformity.
A +2 on a d20 roll always means a 10% bump. For a 3d6 roll, that bump swings all over the place based on where along the curve the original is. Because of that, I am growing more and more of the opinion that a bell curve system works if you treat the individual characters as islands — that they will on their own improve their chances. The players will know when a +1 isn’t worth it, and spend resources (character currency & in-game resources) elsewhere.
But 4/e is so much about helping someone else be awesome, granting each other bonuses. If my +2 to, say, Jeff was much less useful than giving that bonus to Will, then I’m in a sense penalizing the group by wasting my resource on this lesser help, even if situationally it would still be *somewhat* useful. To tie into narrative, whatever it is that my character is doing in 4/e — say, flanking a foe — is universally helpful. The story of these characters is one of teamwork.
(You can also apply this idea to penalties — the fact that I get a foe a -2 in a linear system is equal, regardless of the foe. In bell curve system, I am being relatively more or less competent depending on the foe.)
I recognize this is a half-formed thought, but a comment on a blog isn’t sufficient enough to think this one all the way to wherever it’s going. Still, thank you for the post — you have dislodged something I now need to chew on.
– Ryan
Ryan- the linearity is only true up to a point. The relative benefit is still highly non-linear. Someone who needs an 11 to hit gets a 20% increase in how often they hit with a +2 bonus. Someone who needs a 16 to hit gets a 40% increase from the same bonus. So, sooner or later, you lose the linearity.