You are here

Monday Lab: Whoops

A big lab this time on a big topic: failed resolution, of any kind, for any given set of rules. This is no small thing and may rate as the single most undeveloped topic in the entire activity, to date. Before talking about bad, good, constructive, unconstructive, fun, not fun, or any such thing, we had to back way up and discuss what it even is.

I can't promise that we actually got anywhere or concluded anything with this discussion, but I do think that some air was cleared and some issues were brought forward. "Reflection necessary for reflection," perhaps.

Some notions came to me during the session which I didn't get to, or didn't want to divert to myself.

For actions which are ordinarily understood as "I can do this," or at the most extreme, "Anyone can do this," then the probability of success really isn't about competence in a positive sense. I think we should recognize it as the diminished chance of failure. Although mathematically it's merely an inversion, (e.g., saying 25% chance of failure instead of 75% chance of success), I don't think it's trivial semantics. I think it's a much better way to understand what the dice or relevant procedure is even there for.

To be more specific, think of the possible reasons for "failing less."

  • One may be literally more skilled, experienced, or competent in some way, in terms of a level playing field, i.e., basis for comparison. This works very well for narration when one doesn't fail.
  • Similarly, one may interpret "not failing" in terms of pure plot protection, especially for moments in which the action/effects are a bit strained to describe in terms of skill - think in terms of older D&D saving throws, for which being "higher level" can only be interpreted as "you're more important."
  • Or, should failure should happen to occur, adverse circumstances are already stated and known, or available for incorporation, and are therefore "said again" to explain it.
  • Or adverse details are stated de novo, as part of narrating a failure, and even though we had no idea about them before, they make sense in the moment in some way, or are inherent to the action and therefore can be invoked.

I am pretty sure that some of the other participants have similar summaries or conclusions, so I hope they'll provide them too.

Department: 
Seminar

Comments

Today, I recalled a game that I remember had a sophisticated take on success and failure. This was Theatrix (1993). It was diceless -- without a fortune mechanic just as the Amber RPG. It has two steps, each with a flowchart that the GM uses to decide how to describe the outcome of a particular action.

The Basic Resolution Flowchart guides the GM through questions such as "Are they capable of the action" (decided primarily by whether their skill score is equal or greater than the difficulty) and "Does the plotline require a particular outcome" (which is left to the GM). There are other question and following the decision tree leads to six possible outcomes I won't go into here. 

What is interesting is the second level of flowchart, each specific to a different kind of action -- in this case I'll use the Combat Resolution Flowchart. Each provides suggestions for how frame different outcomes.

For example, we have the row for "Try to Limit the Damage to Suit the Lesson" which you reach if the plot makes no demands and your skill isn't up to it and the GM has decided not to bring "real damage." This roll them splits into Not Capable, Capable, and Very Capable.

Not Capable suggests describing the opponents deft attack and stinging damage.

Very Capable suggests an unexpected element added to the situation.

Capable is broken down further into Skill, Physical, and Environmental explanations and suggests various ways these elements give essentially a warning consequence.

Flowcharts like these might be useful in further considering the various issues we discussed in the lab: eg What does failure look like for a very competent character? what does failure look like for an uncompetent character? what does failure look like if you've got a high skill level but the skill rarely gets used, etc. 

The Theatrix charts themselves don't address the exact issues were discussed, so we'd have to bash up our own set.

 

Ron Edwards's picture

This game sure brings up a lot - possibly "everything." You probably won't be surprised that I consider it to be very far along a sincere but ultimately failed trajectory of RPG play and design. However, it's full of useful and important things. (I think they are painfully embedded in this whole thing about "the lesson" and "what's good for the story" under nanny control, but that's not relevant here.)

Maybe it's odd, but it's not the mechanic's structure as a flowchart which interests me, but strictly its content. To parse failure into usable fictional categories is really important, pride-of-place worthy. A lot of games do this in terms of how extremely one fails (e.g. Legendary Lives, discussed below), but here I'm focusing exactly on fictional description, i.e., cause.

This is what's so absent from most RPG texts: to acknowledge that failure is a momentary and contextual event - what character is this, what situation is this, what ations or conditions are under way. To narrate outcomes in this way is one of the tacit "good GM" skills which people rely upon, or perhaps seek, but it's not taught via the texts at all.

Other games seek to do this with modifiers based on contexts of "difficulty," which is a whole branch of discussion which probably has to be divided into chapters. They are often far more picky and weird and laborious in the text than people actually use in play. As one major example, I point to D&D 3rd edition, which includes two (2) entirely different systems for modifying the target number, which the DM chooses between at will. Prior discussions have shown me that many people learn one of them and think that's the system.

Here's a related topic: to assume a levelling effect from character to character, i.e., skill X is handled in a particular way regardless of who is doing it. That's the key difference in what Theatrix is presenting, that when guy A doesn't kill the goblin with an axe, it's different from how guy B would (also) not kill the goblin with an axe.

I'm reminded, in this, of the meltdown among Gloranthaphiles when Hero Wars was published, which resulted in hatreds and fan-schisms which persist and constantly flare up today, twenty years later. It's essential to RuneQuest that any single skill is exactly the same thing from character to character; it's essential to Hero Wars (later HeroQuest) that no such content/mechanic exists.

LorenzoC's picture

LorenzoC's picture

Edit: sorry for the above comment, it's a by-product of my daughter playing around with a barcode reader while I'm browsing AP.

It's probably not the most interesting or profound thing, but I think one subject that can't be overlooked when discussing this issue is instrumentation and (more importantly) how it's implemented.

When we roll to see if we succeed or we fail, we tend to use instrumentation that isn't particularly sensitive; even then, the split between success and failure is generally implemented (dare I say, idealized) to achieve the most unpredictable (and thus exciting?) results possible.

So having a 75% chance of success is excellent, and possibly considered even "boring". But let's look at what this is meant to represent. In my activity, I do not operate heavy machinery. However, I did learn how to drive a forklift, because my co-workers aren't necessarily always there, and I have to use our forklift relatively frequently. "Learning" the activity took about 15 minutes. I operate a forklift at least 3 times a week, and in 17 years, I've never, ever had any accident with it, nor I ever failed at doing something that it's meant to be done with it. Would I ever even try to maneuver a potentially dangerous machine if I felt I had a 25% (or even 5%) chance of running into trouble? Never.

So what I feel we're looking at is a combination of elements - the nature of instrumentation, both in terms of sensitivity and implementation, leading to the desire of making a test an exciting, meaningful process and not a bland confirmation of our competence in performing that task. The most natural answer becoming that we shouldn't be rolling at all, if we're good at it, which leads to a new problem - being good at something means getting less "game" out of it, robbing us of vertigo and dramatic potential, and making doing that thing potentially less and less interesting the better we become. 
If engaging in an activity we have elected as important to our character means succeeding on 2+ on a d20, we lose the excitement. Then we fail, and we get angry.

Looking back at instrumentation, we also observe that making resolution mechanics "simple" is a top priority for most - and it makes sense. But this can also make the data provided by the instrumentation itself to be necessarily simple, and more often than not this translates into binary results. We roll to see if we get a yes or no, with "yes" being the only usable result, so if you get a "no" either you get to repeat the roll or the activity is over. What is lost here is that when we're engaged in challenging activities, "failure" can be declined in many other ways that a simple "no". Time, for example, is a generally ignored factor in roleplaying games (unless we assume repeated attempts count as a timing mechanism, but still that means "yes" is the only usable outcome).
In practical terms, the most common and observable form of "failure" is "taking more time". You're good at parallel parking, you do it in 10 seconds. I'm bad at it, it takes me 1 minute. You're a good climber, you're over the fence in seconds. I'm not really good, it's gonna take me time (and I'll probably ruin my pants).

Let's say I'm good at picking locks, but I roll poorly - if the end result is that I still open the lock, but it's a long process, it may be easier to rationalize and narrate. I broke my lockpick, I dropped it, I didn't really know the type of lock (but I still made it!). 
I tried to jump over a cliff, but rolled poorly. I fall sort, slam against the rocks, catch a root at the last moment. I'm not dead, but I'll waste time (and probably lost some of my resources in the process). 
Anything that helps us turning the process of failing from "Nothing happened" to "Oh snap, what now?" is useful.
I made the example of rerolls in Trollbabe as something that goes in this direction, in my opinion. I fail, and may recoup from that failure, but that failure is something that happens, and moves things, and produces a new situation. Very few games care about making failure usable and not just a "your turn ends, over to the next" moment.

As a conclusion, I have an example from last night's Pathfinder 2 game. Right at the end of the game (it was 1:00 AM and we were rather exausted, having started at 21.30), the players have entered some sort of burial ground in the middle of the woods and as some of them explore an ancient shrine full of petrified corpses, the rest who stood outside are assaulted by a pair of dracolisks (yes, in Pathfinder dragons and basilisk are sexually compatible). 

The fight goes on for a while and I'll focus on one character, a barbarian whose specialization is becoming a man-bear hybrid while raging. Among other things, he gets a climb speed. So we get a few cool scenes with him using it to chase around the dracolisks who expected their flying speed to give them a bigger advantage. The fight is brutal, one character dies, at some point a dracolisk dies too and the other, left with a small bit of life, attempts to flee, flying away.

I'll upload the map for better visualization:

The dracolisk is fleeing from the right side opening, and ends its turn right at the border of the map, 30 ft into the air. A few characters shoot at him but don't manage to kill him. It's the barbarian's turn - either he hits, or the next turn the beast will fly away.
The barbarian climbs the column and jumps at the creature. Here the game has us counting inches and goes over an incredibly complex process to determine the DC, but it ultimately comes at 30. The Barbarian's talents drop that to 20. The scene is fantastic, the beast flies by the crumbling arc, the barbarian jumps on the column, clawing it as he frantically climbs, and jumps at the flying lizard. He gets 19. 

There's wooing, people suggesting I give him a bonus (which the rules suggest as a very generic "the DM may modify this DC based on the situation") but it's immediately clear nobody would find that satisfying. So the barbarian lunges and misses, and drops to the ground.
The player ends the night on "I don't want to play this game ever again". He wasn't serious, but the frustration was palpable. 
It's the quintessential case of a game telling you you can attempt something, giving you very precise indications on how to do so and making it sound like a relevant instance of play, offering you a range of build options geared toward doing precisely that thing (which the player had picked)... and then the instrumentation returns you a solid 35% chance of failure (probably because this needs to be a potential challenge when you're level 15 or so - the characters are level 7).

LorenzoC's picture

I realized I used a lot of "we", which result in a somewhat pedantic tone. My bad, it wasn't intentional.

noah_t's picture

So having a 75% chance of success is excellent, and possibly considered even "boring". But let's look at what this is meant to represent. In my activity, I do not operate heavy machinery. However, I did learn how to drive a forklift, because my co-workers aren't necessarily always there, and I have to use our forklift relatively frequently. "Learning" the activity took about 15 minutes. I operate a forklift at least 3 times a week, and in 17 years, I've never, ever had any accident with it, nor I ever failed at doing something that it's meant to be done with it. Would I ever even try to maneuver a potentially dangerous machine if I felt I had a 25% (or even 5%) chance of running into trouble? Never.

Lorenzo, as someone not very well-versed in statistics, I actually found this observation really interesting. It made me consider how an 80% skill in Runequest can be rephrased as a "20% chance of failure skill." Having not seen how these ideas go into actual play, do you find this influencing your prep at all? If a player possesses an important skill at 40%, say, you can bet that you'll be seeing more failure. Likewise, that 80% skill will see much more successes. Does this influence your thought process at all as you think about how to make success/failure interesting in play? I'd be really curious, too, to see how players interact differently with two 'identical' skill lists, one expressed as % success, the other expressed as % failure.

LorenzoC's picture

Honestly, when I prepare I generally imagine a situation, pick up a map or picture a room in my mind and don't interface it with the player's skills. Going through the steps of thinking "will they possess the necessary skills to face/solve this?" is something I don't generally do. I don't know if it's good or bad, but my reasoning is that I need to walk into the scene with as little expectations as possible, letting the details become usable in play.

So when I picked this map, I didn't even know it would be dracolisks (I had a few monsters set out in the area, including a pack of aarumvoraks on the hunt for a succulent pair of rust monsters). This was their lair, but it wasn't granted they would be there.

I described the map, naturalistically, and it was the players who brought the verticality of its elements into play (but putting some verticality in scenery is something I've discovered being useful, in this type of tactical wargame-ish scenario).

As for the main point, looking at the chance to fail and how skill mitigates failure (which is something Ron mentioned) is a good shift in perspective.
An 80% chance of success sounds great. A 20% chance of failing horribly and facing dire ripercussions is absolutely terrifying and possibly paralyzing. It's also rather incompatible with the presumption of being good at something - imagine an activity where you fail 1/5 of the time. You're not good at it. Every attempt is a rollercoaster.

And this goes for complex activities too. Driving a forklift is easy. Rope-walking is not, but circus artists perform it daily with a tiny percentage of failures.

What this means in game terms is that either the instrumentation acknowledges this, or we get to a point where we rationalize why we don't always succeed. This in my experience boils down to circumstances - I can drive a forklift, but would I be able to perform that exact maneuver I need, under stress, with just one chance to succeed, while people are hurling stuff at me, etc?

How the rules acknowledge this concept of urgency or danger or stress is important, to me. It's become an important part of my recent design, but I think it's present in some form in many games, even if not explicitely acknowledged. What I'm finding not particularly enjoyable is when games (and here Pathfinder 2 was "guilty" of this) demand that the DM figures out bonuses and penalities that ultimately make the odds "make sense" in the specific situation. It's a process that (aside from being not particularly intuitive and possibly stressfull) places a little too much importance on what I (the DM) feel would make sense here. My preference is for rules to take care of that so that I can be surprised, rather than disappointed.

noah_t's picture

I wonder how much more dramatic and satisfying the player's roll of 19 would have been in the actual play you describe if the instrumentation had allowed you to arrive at the DC much more quickly. Do you have a sense of how the player might have been less disappointed in the outcome?

It's one thing to do something desparate in-character without much hope of success. In fact, doing something desperate and coming thiiiiss close to actually succeeding makes up some of my favorite narrative memories. It's another beast entirely to have to go through a complex, time-consuming (possibly tedious and discouraging) rules adjudication without much hope of success.

I asked about skills and prep in relation to this because I've been reading and playing games that do a lot of their world-conjuring in the skill lists: Burning Wheel and Runequest: Roleplaying in Glorantha.

With both, I feel I really need to understand the skills and how they relate to the world to engage with their success and failure in compelling ways. Success and failure in the "Plant Lore" skill, say, isn't just information about the situation at hand, but also about the fauna and ecosystems of Glorantha.

I'm still thinking about how the potential for success/failure could improve this process of studying a skill list. But, like you, I avoid thinking "Oooohh, the characters are really going to have to have this skill to succeed," as that usually ends with planning the session in advance. 

One last thought which your interesting and thoughtful reply shook loose for me: I wonder how reframing success and failure in the way we're discussing would shift how players built their skill lists. It allows a player to look at failure as something substantial and interesting to engage with and play toward, instead of just a void.

I might actually try framing it this way in my next session zero. Which skill are you interested in seeing your character fail, slip and struggle at 70% of the time? Are you as interested in playing this particular gap in their abilities honestly, dramatically and to the hilt as you are the areas your character shines in? Which failure are you as excited to player towards as you are success?

LorenzoC's picture

I think that kind of vertigo you describe is generally welcome and exciting when you perceive you're doing something exceptional or when the circumstances justify it.

In this situation we had that - the dwarf artificier tried a difficult crossbow shot - a weapon he's not exceptional with - and he needed a 15 or plus to hit. He barely missed but the table still reacted with excitement and laughter.
The problem with the bear-barbarian was that the check was perceived as rather pedestrian in difficulty (he got a 5 or 6 on the die and would have succeeded on an 8 or so, I think) and on top of that, the player had invested a huge amount of resources in becoming capable of doing exactly that kind of thing - his best attribute, a class feat, a general feat, 2 skill advancements. He was supposed to be great at this.

As for the rest of the comment, I think I understand better now what you mean, and yes, I think I most often then not take the kind of skills the game has to offer into account. But this kind of overlaps with what has been said across various videos and discussions about skills and skill lists. For Pathfinder 2, in particular, everyone can attempt every check, and at low levels expecially, a good roll matters more than being good at the activity. This has the predictable consequences.

noah_t's picture

Whoa! Not being familiar with the Pathfinder system, I hadn't realized the test was so pedestrian, Lorenzo. I though the bear-barian had to roll a 20 on a D20 and rolled a 19. That they missed an easy, in-concept roll is even more of a bummer.

No worries about connecting to your current play, Jon This always happens to me. When I was playing Imp of the Perverse, I was breaking every protagonist I encountered in fiction down into a Perversity and a Greatest Strength.

Ron, it's interesting to see these different paradigms of success and failure. I'm realizing I've only encountered a very small slice of this spectrum (mostly the same area occupied by Darkurthe Legends, where the situational 'input' into the roll makes a huge difference to their 'output,' particularly in the case of 'middling' successes).

I have about a billion questions about how these different paradigms give players different 'instruments' for playing their characters, but also not a wide enough experience to frame them in specific ways.

My brain is fried, so that I'm having trouble thinking of anything except the games I'm playing at the moment. With Legendary Lives, and its 10 graded outcomes from Catastrophic to Awesome, each meant to be a specific degree of success and/or failure, it can sometimes feel, when you're in the middle of the results (Inferior, Poor, Passable, Good) like the 7-9 problem, but multiplied. Sometimes it is very clear what a failure should be based on the situation, sometimes it is very clear what a success plus complication should be; but other times -- it seems like simple success/failure would be more straightforward. We've had a few occasions where we've had to stop and try to figure out just what a "Poor" result means in a given situation.

It's more clear at the ends of the spectrum: Catastrophic results tend to be obvious, as do Awesome results (and if they aren't obvious it means we probably shouldn't have been rolling in the first place). They've also had the biggest effect on the plot: we've managed to stay honest with our successes and failures with the story taking shape in the way it has due to the specific skill rolls that the players have succeeded and failed at. There have been any number of situations where things would have swung very differently with a different result. But the "middle" results do seem to have less impact overall on the story, although they may serve as a kind of pacing mechanic or a way to weave in more material that can THEN be exploited and called upon when an Awesome or Catastrophic comes around.

noah_t's picture

The way you're observing the 'middle' results interact with pacing is really interesting! I'm definitely going to watch out for those dynamics the next time I play a game with the possibility for 'middle' rolls and crits. In actual play, I'm still getting the hang of dealing with catastrophic failure or critical success. There's a certain amount of detachment required to be willing to let the dice totally reverse the momentum of a scene, and I sometimes find myself stymied when a player rolls the equivalent of a 'critical.'

LorenzoC's picture

This is a really good point because when listening to the discussion I was several times tempted by the idea of "degrees of success" or "degrees of failure" being an easy way out. It's really not that simple and I think they don't matter that much unless the game is very explicit about what to do with them and supports all of the results in the range as usable.

Ron Edwards's picture

... not because they're unique, but because each exemplifies a completely different concept for failure despite superficially appearing similar, i.e., skill-based.

Legendary Lives (following up on the above comments too) - here, the point is that nothing, not even column shifts, protects a character from the full range of performance (10 distinct & specified narrations), every action, every time. You can reduce the "bad side" to small percentages, but play has shown that getting the range of Catastrophe to Poor really far down requires ability scores almost beyond imagining. My reading is that the dice, in this game, are very much defined as "dumb luck" and "malevolent fate" and "blind justice" (these useful terms are from Tunnels & Trolls), with personal/intrinsic skill being merely there as minor frequency modifiers for these effects. A criucial point is that the lower and higher ends, and not just the most extreme ones, pretty much anything but the middle, for a single roll tend to bring conflicts to a close in highly decisive ways.

Darkurthe Legends - here, the point is that skill elevates or protects you from the harshest possible result, but not from straightforward failure, which is always a possibility no matter how good you are. So being good at something is very good because at a skill value of 5, you only Fumble on a 1 out of 20, and more importantly, are nicely buffered against most of the incidental penalties that everyone suffers throughout an ordinary combat and which can bring a less-skilled person into far more likely Fumbles. 5 is also the typical threshold for higher social placement in one's chosen "box," for the required skills, so it makes a lot of sense conceptually/procedurally to be this good at the important things. As I mentioned, though, "ordinary" failure is inescapable in this game, no one is too good for things not to work out. Narration is therefore critical and the text is very good at explaining or implying how - that cause is important, not just "you miss," and to use what I call the "kibble" of the situation in the moment for explaining and applying that failure. So here, the dice and specifically failure are a means for confrontational situations to change-up at a very fine grain.

Haven: City of Violence (prompted here in part by Jon's mention of the game at Discord) - this game exemplifies a trend that I think bears some harsh scrutiny. For context, you add attribute+skill, roll to get that number or less on d20, no big deal there. My point is the rather severe "difficulty modifiers" based the task of the moment, which kick in immediately, so that doing anything except utterly automatic things like breathing (this is a quote) will incur penalties from -2 to -25. Played as written, the system hits you with 50-75% chance of failure even for something you're skilled at, in any circumstances that resemble a typical RPG conflict that interests anyone. So there's a double whammy: first, your target number, and second, the difficulty modifier. Now, I am not merely saying "high whiff factor, I hate failing, waaaah," I'm saying that the logic of rolling is strange. What is a skill-based target number without the difficulty factor, what does that mean?

Bonus feature: the Forge thread from when I played the game back in 2003, Haven: City of Violence. Unfortunately the RPG.net link in that discussion is dead; I talked about the roll modifiers in detail in that.

Add new comment