You are here

The Shadow of Yesterday: NPC rolls

So, we had ended last session of our Shadow of Yesterday campaign on a cliffhanger: the villain, tied and captured, spells a Zu sentence (a powerful form of magic), but when we played next time and I detailed the scene (the Zu sentence itself), I rolled on the villain's behalf, only to find that he failed miserably, in a sense, ruining the high note I kind-of-hoped to kickstart the session.

So, what happened, I think? As a GM, NPCs are the tools I use to push players or to demand action from them, so having to filter my "move" through a roll kind of blew up the situation players would react to. I think this is a strong point of games where the NPCs don't roll dice and your input manifests directly as a change to the situation.

Additionally, this roll prompted a question by one of the players: how does failure manifest for Zu, which odd effect does this magic school have, when you fail? This created another misconception, in the sense that failure in that game, for me, is a tool you use to transform the player's input into something different, not necessarily something you apply to your own input.

This got me thinking about replacing npc rolls, next session, for a static result.

Department: 
Actual Play
Tags: 
fantasy

Comments

Ron Edwards's picture

Hi Joaquin, let's see what I make of this. As the title says, 1-2-3.

One: Games in which players roll for certain things that their characters do and the GM does not roll for those same things when NPCs do them. This topic includes two subsets regarding whether they can still fail.

Subset A: When success and failure for the NPCs are handled strictly as a direct interpretation of the moment by the person playing them (“the GM” for simplicity of discussion). This is, in fact, the main technique one finds from the earliest games in the hobby, especially early D&D and Tunnels & Trolls, and it often applies to player-characters for lots of things. Outside of combat and some highly specific actions, whatever they do that’s congruent with who they are, they do it, and often things inside combat too. The big step-away from this concept is found in RuneQuest and The Fantasy Trip: In the Labyrinth (someone will have to tell me about other early seminal games, e.g., Traveller), but probably surprisingly to some, it’s maintained in first-generation Champions.

So that’s something to consider right away: what you’re describing, in its most literal form, isn’t some odd technique or innovation. Nor is it intrinsically problematic. If there’s some messed-up mis-use of the authorities’ distribution at the table, it can give rise to “yes I can no you can’t but wait wasn’t I over there” murky breakdown, but so can lots of systems, and we don’t have to consider that to happen necessarily.

I’ll develop this further in Topic 3, so stay tuned for that.

Subset B: When success and failure for the NPCs are still possible outcomes, but the mechanics are handled player-side and include opposition and potential NPC failure. I wouldn't be surprised if it shows up earlier, but in my library, the first games which formally pose "foe attacks, its attack is X difficult, resolve by rolling your Defense against X," are Legendary Lives (first publ. 1991) and The Whispering Vault (1994). The same thing applies to anything the NPCs do: you roll to oppose, and that’s the only kind of rolling there is.

As you can probably see right away, this method really doesn't differ from the GM rolling for the NPCs; after all, dice don't care which hands actually touch them, or whether the GM is rolling for the NPC’s low (easy) attack target number or you’re rolling for your defense against the NPC’s high (hard) attack target number. The main difference is that the resulting probability curves tend to include less mitigation, i.e., two rolls' effects can't "cancel out." Whatever other nuances may be included are highly specific to a given title. This particular arrangement of mechanics doesn't appear relevant to your concern ... except that it does matter in reference to subset A above, which as I said existed long before B came along, also to be discussed in Topic 3.

My point in comparison during this campaign is Apocalypse World, because I'm running another campaign using that game, so I'm mentally comparing both games all the time. And my train of thought here was: in AW, in a similar situation I could've narrated a soft move (announce future badness, for example), and then the situation would be on players' hands.

Ironically, I see the relevance of your subset B paragraph, partially because yes, only one roll means the distribution curve works as expected, but also because there's an (arguably minor) shift in the narration, where the situation (a Zu sentence spoken) is already there, and the weight of the scene is on the Resist roll done by the players, instead of being distributed between my NPC roll and theirs. Also, due to how resisted skill rolls work in this game, if the NPC fails (a 0 result), there's no need at all for the players to roll on behalf of their characters, which is what happened. Of course, it wasn't a FAIL situation, it was just a curious mismatch of expectations that turned a potentially risky roll for the players into an NPC failure and a Zu farming scene.

I have a lot of experience GMing Burning Wheel, which is similar in this regard (NPCs have to roll to act against PCs), but I've never perceived this same issue, which now leads me to believe there might be something else undergoing in my expectations for that game. A circumstance I didn't disclose about the session is that we are streaming the campaign. We're not professional streamers in any sense, and there's hardly 1 or 2 in the audience during each session, but this circumstance might have had an effect in my expectations regarding the game or my role as GM (I'll probably write more about this in another comment).

Ron Edwards's picture

Two: Failed resolutions, conducted in a contingent way - why do they happen in play, ever, at all? For a first pass at this topic, which I maintain is hard-core central to any question of play procedure, see Monday Lab: Whoops. It's been developed through a wide range of posts here ever since, as different people found the concept to be powerful and helpful throughout a bunch of different games. I probably should starting thinking about doing another Lab, "Whoops One Year Later" or something like that.

I think this is a big topic for you. Let's take it to the points of play in which you evidently are OK with someone failing an action-oriented resolution: the players, regarding their characters. There you are, in play, using The Shadow of Yesterday, which I know well, and someone has a rating of 2 in the Freeload ability. No modifiers or bonus dice apply, so here, I’ll roll with three real dice: 0, 0, +1, which with the 2 ends with a Great result. Yay! Presumably this has some kind of “carry events along into what happens next” effect.

But ... here, not rolling this time, but saying I get -1, -1, -1, which leads me to a result of 0, the minimum possible: Failure. What is that, at your table, and to you? The game text tells you nothing concrete. I’m not demanding an answer from you actually here at this discussion, but asking you to consider the following issues.

Is it a full stop? Do you conceive of play in which someone rated very highly in some ability (which 2 is) can just ... fail? If so, then do you consider it a failure of ability or of circumstances? (basically, the contrast between “you suck” vs. “life sucks even when you don’t”) Do you straightforwardly get to choose between these as you see fit in the moment? And how about consequences? Is there now guaranteed to be suffering upon a failed resolution? Are the characters shut out from eating and resting tonight?

There’s lots more to this. Instead of a full stop or “big no,” do you consider it a cue to bring some other thing into the situation? Is it now necessary for everyone to get up from a prone-in-play position, brush themselves off, and continue with “try something else,” possibly to wait for your presentation of whatever that might be? If these or similar inquiries can be answered with yes, then is it not really a failure but just a way to add some detours or color into how they eventually or amusingly “succeed by accident” to get some food and lodging this evening after all?

Again, I really don’t want you to write answers here to any of these questions, even if you’d like to. You don’t have to defend anything or prove anything. Instead, I ask you to consider privately why, for that NPC, for that moment in play, did the game design and your ideas about play require a roll? As you see it, what is necessary to be the case, for any instance of play, for anybody’s dice roll, including the chance to fail, to be a good thing? When, during play, is that not the case?

Well, if I had to meditate on this particular situation, regarding the meaning of failure, and the question the player asked ("how did he fail?"), I think I dismissed the possibility of adding a twist (other than "he fails"), because it'd detract me from focusing on the players and shift that focus to that particular NPC, how failure interfered with his intention, etc. In my mind, conflict resolution is a tool for players characters and not exactly for NPCs. Although, in light of what you've written, I think if I could use NPC failures to twist a situation and make it more interesting.

It didn't help that this particular iteration of a roll was using a magic system, where traditional rules for conflict resolution are, if not suspended, kind of forgotten sometimes (in tSoY at least).

As you see it, what is necessary to be the case, for any instance of play, for anybody’s dice roll, including the chance to fail, to be a good thing? When, during play, is that not the case?

I'd say a failure chance is a good thing when there's some opposition between what the player wants and what the setting or NPC want, in such a way that failure might allow me (GM) to mess their intention or transform it a bit into something else. What I'm hesitant about is when NPCs have to roll in such a way that, if they fail, players don't need to react to it at all (roll to Resist). In such a case, it looks similar to something the rules discourage, which is having NPCs roll to affect other NPCs, or having them roll unopposed.

Ron Edwards's picture

I'm still of the position that any meditations can be kept private - so you know, one of my reasons for saying this is that even here, there's a chance for internet-style dogpiling in response which no matter how wel-intended would be destructive.

Now, you did bring up something which I think is OK to chat up a little, mainly because I'm talking about myself/my ideas and not you. It's this "twist" business, specifically, the idea is that a missed roll is a signal for a GM to introduce new information or events.

I suggest that this is absolutely a desperation device which evolved into hobby practice specifically in the context of GM as den-daddy and controlling, guiding force. Designs conceived and published in this context result in players' resolution rolls (or any equivalent) that don't actually resolve much, diminished into atomic moments which are only as important as the governing eye-and-hand of the GM permits, case by case. Once play has been homogenized and diluted to the point where the players can't "cause trouble" with their actions, however, then somehow events (as preferred and conceived by this GM  person) have to be introduced somehow. Since the failure roll doesn't do much as such, it's treated as a signal to do this.

There isn't any intrinsic logic to it; it's merely something that happens intermittently and therefore the GM receives a useful intermittent reminder to "make something happen," because otherwise nothing is happening or will. It might as well be a cuckoo clock set at random intervals, except that the illusion of system-influenced causality can be imposed - "see, you failed your Freeload roll, so here is Bob the Informative Fellow, showing up to tell you something." Calling it a "twist" is merely putting pressure on this GM - frankly, an unenviable position - to come up with something entertaining out of the blue.

Ron Edwards's picture

Three: The issue of control, manipulation, and deception. Here, the root concern is when anyone’s thoughts or feelings about what might be about to occur, in the thick of resolution methods in action, shifts from a provisional hope into a plan.

In my little vocabulary at this site, we’re talking about shifting the current methods of situational authority into outcome authority, perhaps canceling whatever the latter’s methods may be. There is no grey area here; doing this is instantly apparent to everyone, and the question is whether it’s acceptable and, critically, known to be acceptable, or not. (It may be helpful or interesting to see my Phenomenology videos about what play is and how it works.)

Topic 1 can help to understand this, I think. Consider when the GM says “he grabs you and shouts, ordering to submit,” with no roll, and that’s cool. Consider then when you throw the guy over the balcony, perhaps with a successful roll I guess, and the GM says, “he grabs the railing and doesn’t fall,” and then, even, “he climbs down to the window below and escapes.” Both are saying what the NPC does without a roll, but the time when he grabbed you and the time when he broke his fall (and maybe escaped) are definitely not the same thing. The GM stopped playing and started writing.

People tend not to like sudden shifts in authorities’ current construction, i.e., when it was not known or accepted that someone at the table can do that, and then they just do. Shifting from I casually hope this happens to This is what will happen right there in the moment faces some unhappy effects among the people playing, which historically has led to the allegedly masterful or subtle version in which the NPCs’ actions are nominally or apparently rolled for, but they’re actually resolved preeisely as stated above. In other words, shifting authorities specifically to the outcome category and lying about it. I assume you are not talking about doing this in your game, but I think a lot of us can see that it’s an elephant that is, if not actually in the room, definitely right outside the window looking in.

Topic 2 is relevant as well, because the player-side mechanic is absolutely designed to directly defy any such possibility of deception. It is, perhaps, the most outright commitment to “yes, your NPC guys can too fail” in the hobby prior to the 2000s. I think you would do well to consider why such a defiant mechanic was deemed desirable by those role-players, including the authors.

In fact, let’s go right to the cestus-armed fist boxing punch: it is flatly toxic to desire to have X happen to the extent that you have pre-writtten it, mentally and emotionally, to happen, and yet also to use devices of play which admit to the possibility that it will not. One is inevitably trapped between being unhappy with outcomes in play vs. sneaking and lying in order to be happy with them.

Fortunately, now that it’s happened, you are seeking solutions. It would appear the hobby has provided a big “yes, do that” from its earliest days: just say whether NPCs succeed or fail, we all know that’s how it works, and no lying about it. That works, so, hooray, awesome, problem solved ... except that the trouble was not the failure at all, it was the pre-writing for outcomes.

And that is really a problem, and the solution you’ve proposed isn’t going to make it go away. Even if you shift to “NPCs do not roll,” if you maintain this notion that “the GM guides the story” and therefore use this NPC-action device to determine outcomes, then the day will come when some player’s rolled result, or some player’s understanding of what is happening, or even some statement of yours made a few moments previously, is going to pull the rug right out from under your writer’s plan again. And you’ll be right back where you started before this post, and if five decades’ history is any guide, sooner or later you are going to resort to manipulative and eventually deceptive means of making play cease to be play and start doing, instead, what you want.

Here’s really good reference for related matters, from recent play of the (almost) same game: Troubles in Solar Town.

Add new comment