On failure, its outcomes and its implications: derailing

Abstract

There are basically three topics in this miniseries. Thread one is about (mechanical, die-rolling) failure per se and how it feels to me as a player. Thread two is about failures, interestingness and the Fail Forward/Roll Dice or Say Yes sort of concept. Thread three, today's topic, is about the interaction between failure, narrative and the coherence (in the non-Forge-jargon sense) of the play experience.

Failure, Narrative and Coherence

So, what am I wittering about today?

There's two interrelated ideas here. I'd like to begin by talking about how various types of failure affect the coherence of a game, and the table atmosphere. Then I'll discuss the relation between spontaneity/improvisation and narrative coherence. Finally, I'd like to tie these back into the main topic of the miniseries, by talking specifically about the way mechanically-mandated spontaneity affects coherence.

Sleeping Policemen

You are driving comfortably down a road (or in my case, being driven) and encounter a sleeping policeman. If you are lucky and alert, your progress slows rapidly and there is an uncomfortable bump before you resume; the smooth progress of your journey has been interrupted.

If you are unlucky or unwary, there is a sudden violent jolt causing passengers in the back to cry out, or you slam on the brakes at the last minute causing a disconcerting lurch and annoying the car behind you.

Failure in games can be a lot like this.

For international readers, a "sleeping policeman" is also known as a speed bump - a raised band across the road that is very uncomfortable to take at speed, used to enforce low speeds in dangerous or residential areas. If you encounter a real sleeping policeman in the road, it is considered very impolite to drive across them.

Now, there are different types of failure and these do not all have a jarring effect. Rolling a miss on an attack roll is a failure, but the game usually continues smoothly. Where failure becomes a sleeping policeman, or even a brick wall, is when the failure affects the narrative.

Types of failure

Broadly speaking I'd say there are three types of jarring failure. One is where a character fails in a way that feels inappropriate: Zorro repeatedly fails to stab the generic guardsman, Sultry Hotlips is unable to find a single seducable suspect, and Gorlak the Wise can recall not a single detail of the history of his own country. These failures are usually results of mechanical quirks mixed with sheer luck, and often a GM can spin them to avoid the jar.

A second, related type is where the failure breaches the tropes that are in play. Valorous Darksword nobly volunteers to hold back the rebel tide while the rest of the heroes dash to prevent the sinister ritual, and is immediately felled by a single lucky arrow at extreme range. There is an elaborate heist, and as the first of the infiltrators strolls through reception for the vital first stage of the plan, she discovers her stolen security pass doesn't let her through the secondary doors. These tend to be a mixture of player mistakes and bad luck, and are harder to work around.

The third type is where the failure prevents progress, and here investigative games are the posterboys. Detectives running out of leads, either through inattention, rolling badly or ruining them with unwise decisions, are a constant complaint on Call of Cthulhu forums. In more combat-based games, the party may feel they've run into a challenge they can't beat. A White Wolf game might feature situations where the players simply don't know what they could meaningfully do. Mostly these can be boiled down to saying: "Well, I flubbed that roll, and now I don't know what to do."

Failure and Coherence

Because these failures interrupt the flow of the narrative, they tend to disrupt the coherence of the story unless the GM is extremely skilled at improvising. It can often feel as though the next step is simply being told what to do by the GM. Players stop, discuss and may well spend a long (and sometimes frustrating) time working through options before they find a way to keep the story moving.

A very common situation is that players keep looking for ways to progress with the plan or route they are already taking, because that seems to be pyschologically how we're wired. Especially where narratives are concerned, we tend to think in terms of making the thing we're already trying work, rather than stepping back and looking for another option.

In a Call of Cthulhu game, the intended way to progress may well be a clue the players forgot or mishandled. In these cases, the GM may have to find a way to remind them of it or convincing them to try again. It's difficult to do this without it feeling very metagamey and heavy-handed.

Broadly speaking, I think the issues with coherence are firstly, that seeing one plan fail and having to switch to another naturally tends to break the flow and disperse momentum; and secondly, that a narrative of "after doing X, the heroic party started doing Y, but it didn't work, and after trying several times they talked for ages and did Z instead" simply feels less natural and satisfying than "after doing X, the heroic party did Y". A linear progression tends to feel more comfortable, and build up momentum more efficiently, than one with dead-ends and retreats.

It's worth noting that fiction narratives do typically involve changes of plan - in particular, fantasy epics, detective stories and crime stories are fond of things going wrong. The route through the pass is blocked by the minions of the dark lord, and now you must plunge into the Hideous Netherworld that was repeatedly mentioned earlier. The obvious suspect turned out to have another explanation for the deeply suspicious things she was doing, she just refused to admit it to anyone for reasons that, when you consider it's a multiple murder case, look astonishingly spurious and ill-judged (hello, Wasting Police Time). The heist went wrong because of something unexpected and probably ironic, so unless it's All Part of the Plan, time to improvise! After all, The Plan going perfectly isn't usually very satisfying.

I think one of the differences is that in these cases, the change of plan is usually smooth. The merry band are on their way to the pass, find the minions, and enter the hideous netherworld instead. The infiltrator discovers a new security door has been installed and turns aside. The investigators realise that the suspect is incredibly bland and probably something else is going on. This is partly because of the way a single author can summarise and elide information to a smooth conclusion.

With a group, things are a bit different. People can keep trying slightly different tacks on the same idea - "what if we search the settee? the shower? under the lino in the kitchen?" - because they've committed to it and the instinct is to exhaust all possibilities. Groupthink can also come into play, with everyone reinforcing the same ideas.

Because of the fact that you're not actually your character, and the way game mechanics tend to focus attention on relatively immediate goals, it's also quite easy to get sucked into the means and kind of forget about the ends. I've run a game of Call of Cthulhu where the hook was "find out what happened to this guy", where the players first fixated on working out what was suspicious about him (nothing), and then on the very suspicious workshop (fair enough). They got really into the clues in the workshop, but there was a point where they were chasing leads I hadn't actually got answers for - things intended to be vague hints and long-term hooks, or bits of background I devised so that the mystery didn't exist in a void. When I couldn't answer any more questions, I had to remind them that their paid assignment was in fact "find out what happened to this guy" not "investigate a workshop". Which is not to criticise, but just to spotlight how easy that is to do.

Similarly, it's very easy in non-mystery games to get sucked into, for example, concocting increasingly elaborate plans to break into somewhere or overcome some enemy, rather than reconsidering. Which is to an extent fair enough, because I really like concocting elaborate plans. I think a thing here is that if you concoct an elaborate plan and it works, that's fine; but if you devise a series of plans over the course of two hours and have to abandon each of them, eventually realising that maybe you shouldn't do this after all, it's unsatisfying.

I think tension has quite a bit to do with this. If you encounter an obstacle in your plan and overcome it, it ratchets up tension and then you have a sense of great achievement. If you encounter an obstacle but foresee it and change the plan immediately, you feel clever. If you try and fail to overcome it (and perhaps particularly when you fail a die roll, because it's purely random and mechanical) then the tension just oozes flatly. One way to recover is to try again and succeed, because emotionally that feels like the obstacle just grew but you overcame it anyway. But this can suck you into a cycle of frustration that still ends in a flat and unsatisfying change of plan.

To come back to my line above, fictional changes of plan tend to involve planning to go along one track from X to Y, but seeing a looming problem and switching track at the points in the direction of Z. Gaming failure-induced changes of plan normally involve heading merrily off to Y, finding the line is blocked, getting out to try and clear the debris, seeing whether you can lift the train over it, spending an hour not only discussing the use of a series of cranes but also starting to build them out of sheer enthusiasm, realising it won't work, giving up with a sense of unsatisfyingly dispersed tension, reversing back to the points and driving off to Z.

Failure and Frustration

All games (that I care about) feature failure, and in many ways it's what makes the games fun. As my friends have observed, simply choosing to succeed at things is only fun for a limited time and for a very specific type of experience. The threat of failure is crucial to the play experience. In many cases, failure is only a cause for passing regret, or even contributes to building the tension and excitement that will make the final success more satisfying. Sometimes it has longer-term consequences, which may be positive or negative, but in either case can build towards a fulfilling narrative. For example, the failure of a crucial roll may lead to a character death, and the failure of a plan can lead to the loss of a stronghold to intruding rivals.

Things become more frustrating when jarring failures come into play.

If your character repeatedly fails at supposed areas of competence, this breaks the contract you have with the game. Playing an action-adventure game where Curt Manly PI continually misses clues, falls over during brawls and makes a fool of himself in front of attractive suspects is not in keeping with either the genre or the character you have made. Playing an ace sniper who can't hit the broad side of a barn ruins the experience because, since only what happens in game narrative is real, you are not in fact playing an ace sniper. The player can easily become frustrated with their play experience.

The failure of a genre-appropriate plan is jarring because, to a large extent, games depend on tropes. In a serious narrative, the lone hero who nobly holds back the tide may sacrifice herself, but she holds back the tide. When the Big Heist goes down, the plan has a 100% chance of going awry in some way, but the plan still gets executed. If you're lucky, there's a way to deflect the failure into another trope that still works for the genre, style and characters you're playing. If not, it can be frustrating that you did the character-appropriate, genre-appropriate thing and it didn't work. It's even more frustrating if you did the character-appropriate, genre-appropriate thing instead of the mechanically-optimal thing.

There is an issue with many games in that the simple, direct approach is often mechanically optimal. This is partly because PCs in many games are very good at strongarm tactics and bad at other things, and partly because the direct approach often averages out in the PCs' favour. Failing one roll in combat is rarely disastrous. Involved tactics often require using skills you are less good at, and/or taking one-shot rolls where the entire plan comes crashing down if one person fails their Disguise check.

For me at least, this is unsatisfying because I'd often rather try a clever plan than kick down the front door and shoot everyone.

Running into a wall where you have no idea how to progress the narrative is unsatisfying because your momentum disperses, and because I think on some level it feels like a failure as a player. Zorro's inability to stab a single guard can be put down to appalling luck tonight, but you having followed up the wrong leads or tried a tactic that won't work is all on you. You picked the thing that you think should work - and it doesn't. That creates a dissonance which frustrates you, unless that failure itself offers understanding of why you failed and leads naturally to an idea of how you can move on to succeed, in which case it can feel like a satisfying and dramatically-necessary failure.

Overall, I would say that when frustration arises, that itself disrupts the coherence of the game, because it breaks immersion and the sense of enjoyable focus that maintains momentum.

Spontaneity and Coherence

So, what of spontaneity? Why would it matter whether content and events are created beforehand or improvised on the spot?

I think there are two things gone on here. One is that spontaneous play of any kind can sometimes result in coherence issues. However, I think mechanically-mandated spontaneity is particularly prone to cause problems.

Setting out a stall

At this stage, I should say that I don't have a problem with largely improvised play. I'm not advocating for All Prewritten All The Time. I've played in fantastic games that were largely made up on the spot by skilled GMs, and fantastic prewritten adventures, and I've played (and run, and decided not to run) underwhelming prewritten games and ad-hoc games that didn't quite work.

In some ways, a fully-spontaneous game run by a GM skilled at that playstyle is even stronger than a prewritten game in terms of producing a coherent narrative, because they can avoid the need to fit in crucial bits of plot or clue or cause-and-effect that the designer intended. The GM can ensure that everything that happens forms part of a continuing narrative that makes sense in the world they're depicting, and tweak things to subtly direct players towards the content they want to focus on, without ever hitting a wall. It may well be that chunks of improvisation in a strongly pre-planned game are more disruptive than a fully-improvised experience.

That said, let me outline a couple of drawbacks.

Dilution

In its most simple form, improvisation adds something to a game that is likely to be more weakly connected than most existing elements. Let me put forward some arguments for this.

Recently I ran two games of Lovecraftian investigation on two successive days.

The first was a scenario that I spent two (2) years planning and hand-crafting specifically for three players and their existing characters. I considered the players' tastes and playstyles (as best I could), and the skillsets and personalities of the characters. I wove the NPCs into a social web with interconnections and attitudes. I carefully layered clues upon clues, ensuring that each part of what happened in GM-reality was accompanied by clues that allowed the characters to discern it and to eventually understand the quite implausible truth.

I found people willing to read it and comment, and filled in the gaps and flaws they identified, and added additional detail and explanations. I mulled the scenario over in my mind countless times, and pictured each of the NPCs and their possible interactions with players - not deliberately, just because it was a long process. There were several potential ways I could think of to approach it, each with their own arc.

The second scenario was randomly generated from a table in the space of three minutes, for characters randomly generated in the same period. I nobly appropriated real places to lend a touch of verisimilitude, since I could just visualise stuff in my head instead of inventing it. The plot was rudimentary and fleshed out as I went along, sometimes in response to player questions or actions. The main NPC had a motivation of "works for aliens because reasons" and no clear defining traits.

My group enjoyed both scenarios. They made it (very politely) clear that they enjoyed the first scenario immeasurably more, because it was a painstaking piece of richly-detailed investigation that was coherent and allowed them to progress logically and naturally through the weird events that drew their characters in, whereas in the second one they ran around to wherever the next clue seemed to point and looked for somebody to punch.

The second scenario was highly spontaneous, but I was making a strong effort to keep things coherent and not throwing in random events simply to spice things up. It was still far less a coherent experience than the first, and it suffered from it. I can't say there was any particular pre-intended arc in terms of exactly what kind of story it was trying to be; it ended up sputtering about formlessly somewhere between an investigation, a revelation and a pulp adventure. I had a very loose idea of what was going on, but because each scene and piece of the game was fairly spontaneous, they did not feel well-connected, so what we ended up with was a bunch of stuff that happened to some people.

Broadly speaking, my argument is that mechanically-mandated spontaneity can detract from the play experience by marring the narrative coherence of the game. I think it does this in several broad ways: by generating new and typically urgent elements that are more weakly (if at all) related to the existing experience of play; by demanding emotional and intellectual focus that could otherwise be devoted to the group's intended game; and by creating tangents that can leave the campaign feeling quite removed from the original concept.

None of these is necessarily a bad thing, but in the wrong circumstances this injection of "interestingness" can get in the way of the game you actually wanted to play.

For ease of use I'm going to use the term Happening to indicate one of these situations: it might be a GM Intervention, or a consequence of Failing Forward, or a result of the players accumulating a certain number of Awesome Chips in their pool. Regardless, the game mandates that the GM should devise something interesting that happens right now.

I really do not want to get into big terminology arguments because that's not the point. I am using Narrative to mean, broadly, the sweep of events in the game together with things like tension and dramatic arcs. Coherence means how well the events of the game seem to fit together into something that makes sense.

The reason I'm talking about this here is that, as per the rest of the miniseries, "consequences for failure" is a sizeable source of Happenings.

Because they are

A Happening wasn't part of the original idea, so is inherently likely to feel bolted-on, because it is bolted-on. You had to devise an event somehow related to a failed die roll and incorporate it into a pre-existing game.

The argument from time

Here's a second argument: Happenings are likely to be less coherent because the GM simply has less time to devise them.

GMs generally spend a reasonable amount of time planning games, whether that's between sessions or simply when devising the original campaign premise and setting. They tend to try to ensure the game makes sense, and may specifically aim to create a particular type of narrative or a dramatic arc. However, they usually have mere seconds to think up a Happening, so it's natural that they are less rigorously thought-out than pre-planned aspects of the game.

The "No True Scotsman" position

Finally, I think you can make an argument that the actual mechanic of Happenings inherently argues against them. To whit: if X was a natural and interesting outcome of the ebb and flow of the game, something which felt strongly tied into the current situation and events, the GM probably wouldn't have waited for a random game event to implement it.

Obviously this is a generalisation. Some things seem like interesting possibilities, but perhaps something to only implement if the PCs act in particular ways or get unlucky. Sometimes the GM doesn't want to interrupt when the group has momentum, simply to throw in an idea they think is nifty. In some cases the GM knows a particular 'scene' or location will be a big focus, and plans some ideas for interesting events that could occur there depending on the actions of the players.

On the whole, though, I don't think GMs typically have a set of really good ideas suited to a specific point in the game just sitting around waiting to be used if Jessica rolls a 1. I believe most of the time, if they come up with an idea which would naturally follow the current course of events and fits nicely into the genre and tone and narrative arc of the story, the GM implements it without necessarily even thinking about it as "an event". This is just the everyday business of at-the-table GMing.

The logical flipside of this would seem to be that when the GM is called on to generate an interesting event, the ideas they have available will on the whole be less natural to the narrative. They are things whose time had not come, or that seemed like vaguely cool possibilities that would throw off the perfectly good contrasting narrative the players were building up. In some cases they are simply things that could happen rather than things which seem either dramatically or logically necessary. Yes, a man could burst through the door holding a gun, but it isn't always fitting.

Dilution again

So, I think Happenings tend to be more weakly connected to the narrative than content the GM devised otherwise. This weakness combines with a second feature of most Happenings: they are urgent and demand attention.

It's entirely expected that the coherence of a setting and a scenario vary. It's not a real world, and some parts are going to be, at least initially, mostly backdrop or genre dressing or filler. In many cases players and GMs alike can politely ignore or handwave parts of the setting that seem a little fragile, underplanned or distracting. This makes them less salient and keeps the campaign moving in the group's intended direction. However, Happenings are generally things that must be addressed, and soon.

The fact that players and GM are obliged to confront them tends to highlight any weaknesses of a Happening, whether those are logical, narrative, mechanical or even social.

Obviously these weaknesses can apply to any part of the game, spontaneous or not. It might not make sense, on closer examination, that the enemies were able to lay an ambush somewhere the PCs only recently decided to go without telling anyone. A pulpy arc where the PCs repeatedly clash with mooks (occasionally getting captured and daringly escaping) and learn more about a plot, slowly building momentum to confront the big boss, may become dischordant and boggy if a mook encounter turns into a mini-investigation in more of a detective style. A challenge may turn out to rely on skills the PCs don't really have, or misjudge how much of their capability they lose without access to magic, or conversely become a walkover that feels hollow. And the revelations or decisions that happen in play may run counter to the initial pitch, the tacitly-agreed tone of a game, or specific player or GM preferences.

For example, here is something that may superficially seem reasonable.

It is a game of epic fantasy, and the party are trekking across countries in the course of a mighty quest. Currently they find themselves caught up in the darkness afflicting the Heathlands, which has been growing rapidly and threatens to overwhelm crucial strategic locations, potentially sending the whole population into flight or even enslaving them. After several exploratory forays and skirmishes, they worked out the root of the problem and are now winding up for a major milestone in the game: confronting the Grey Titan, Zchaan, with an ultimatum to abandon forging artefacts for the Howling Tyrant, or to perish before their blades. They are challenged by a party of gnomish scouts hunting for raiders, led by the dutiful princess Yark, and a social encounter begins. It was just supposed to be a basic exchange of information, but the players roll badly, and a Happening is mandated.

The GM reviews the situation, and decides that the scouts compel the party to travel to Gnomvark, the great citadel, and discuss their business with the Archmage. The Archmage, however, is distracted by local matters and by a grand ball to be held shortly, which the party may be compelled to attend before they can get an audience and thus permission to leave.

This is a stressful diversion at a time-critical point in their mission, so it will add tension and provide a mild inconvenience for their bad roll; it also avoids having an unwanted fight against what are supposed to be valued allies, which would be inappropriate for the campaign. There's nothing specific planned for the encounter, as the GM wasn't going to bring the gnomes in just yet, but they can explain what they're doing, maybe get some information or resources, make some allies, and soak in the flavour of gnomish culture. This will prove useful when they return to Gnomvark and perhaps make the players more interested for that visit. Visiting kings, exploring other countries and attending balls are classic parts of epic fantasy. They can get back to the main quest soon.

Unfortunately, this also takes the momentum that was building up and drives it into a swamp.

The party had been progressing smoothly through a series of narrative points that built up naturally to a climactic encounter with a powerful entity, which also fits into their long-term goals. Having learned the truth, their instinct is to immediately act on that knowledge and shut down the local keystone of the spread of evil powers before matters get worse. The Grey Titans were not thought to be inclined to favour evil, so there is suspicion and doubt over the truth of matters and Zchaan's motivations; will this end up as a fight, a negotiation, or even a rescue mission? The supposed in-game urgency of their quest enhances that motivation: not only do they want to halt the spread of evil in its tracks, but the faster they can do so, the sooner they can continue their journey in the hopes of putting an end to the Tyrant forever.

Unfortunately, although the Archmage is a notable NPC in the background of the campaign, this diversion doesn't tie in to the existing flow of the game. It doesn't form part of their discovery of events in the Heathlands, it doesn't seem to build on their desire to stop Zchaan, but nor is it an obstacle they can actively overcome and gain further momentum from. It is a tangent.

By dealing with the Archmage situation, the players have to switch their focus and engagement from "dealing with the Zchaan situation". They lay aside emotional investment in that plotline, excitement about the coming encounter, and the half-formed plans and in-character ideas they are (consciously or otherwise) forming relating to that plot thread. Now they have to focus on a location they had no previous interest in, an NPC (well, several) they have no particular reason to deal with, and some events that haven't yet offered them any hooks.

Players in this situation may well expect that the Archmage sidetrek will prove to be relevant to the main quest. A careful GM will try to ensure this happens, but as this Happening is spontaneous, they will need to do a lot of work quickly to make it so. They haven't prepared any clues, omens or hooks at this stage; worse, players might bite too early on hooks intended for later use, perhaps assuming they must relate to their current activities (or else, Chekhov's Gun-style, why would they exist? Foreshadowing is hard in RPGs). The players may spend quite some time looking around for relevance that does not exist, and potentially leaping on every possible hook. This can build frustration; and when they realise the gnomes aren't relevant at this stage, disappointment. None of this contributes to the momentum of the Zchaan arc. At the same time, it doesn't really work as a relaxing and informative interlude, because the players haven't just finished a dramatic section - they were anxious to do so.

Broadly speaking, the problem is that this is narratively inappropriate. It's very much in-genre for the party to attend an important event, and even to be dragged to one unwillingly while on an important quest. However, dramatically speaking this ought to happen when there isn't a huge amount of momentum in the first place, such as during a travelogue or in the lull after a storm.

Wait, hang on

Could this Happening be deployed better with planning, though?

Ironically I don't want to devote the time that would be needed to turn it into an elegant piece of GMing, but let's see.

It is a game of epic fantasy, and the party are trekking across countries in the course of a mighty quest. Currently they find themselves caught up in the darkness afflicting the Heathlands, which has been growing rapidly and threatens to overwhelm crucial strategic locations, potentially sending the whole population into flight or even enslaving them. After several exploratory forays and skirmishes, they worked out the root of the problem and are now winding up for a major milestone in the game: confronting the Grey Titan, Zchaan, with an ultimatum to abandon forging artefacts for the Howling Tyrant, or to perish before their blades. They are challenged by a party of gnomish scouts hunting for raiders, led by the dutiful princess Yark, and a social encounter begins.

The party are asked to visit Gnomvark, the great citadel, to discuss their business with the Archmage. When they arrive, they find the Archmage distracted by local matters and by a grand ball to be held shortly, which the party may be compelled to attend before they can get an audience and thus permission to leave. During their attempts to gain an audience, the party realise that there is in fact a connection between Gnomvark and the Zchaan situation. One of the expected attendees may be connected to the smuggling of the evil artefacts; moreover, there is a traditional connection between the Titans and the gnomes, and so the Archmage may have some insight into the situation with Zchaan.

In this particular case, I don't think it matters much whether the party are compelled (by poor rolling or even GM fiat) to visit Gnomvark, or given the option. Either is one valid approach to these matters. I think this more planned approach is a slight improvement on the previous.

Although the party will be forced to wait around to talk to the Archmage, they have a sense that this frustrating sidetrek may be a natural stepping-stone in their quest. Rather than the irritation of feeling "precious" time tick away pointlessly (or the immersion-damaging sense that time doesn't really matter) and waiting for the ball to be over, they might instead be impatient for the ball to happen because it's now part of their arc.

This also opens up the possibility that Gnomvark itself is important now. If an emissary of the Tyrant is attending, are the gnomes in danger? Can the Archmage be involved? Whereas the Archmage was previously just a potential source of exposition, she might now be a threat, a mole in the "good" alliance, a dupe, a reluctant pawn, a dangerously neutral force, or an ally in need of help. They may need to persuade her of the emissary's treachery to help them combat the spread of evil, or even to foil a plan that threatens Gnomvark itself. While not directly relevant to the Zchaan situation, this is a chance to crush a flowering evil or to cement an alliance with the gnomes, which contributes strongly to their main quest and potentially to the Zchaan quest as well (especially if the emissary's plan relates somehow to the artefacts).

The party also have a potentially interesting decision. Do they try to insist they are released early, progressing their Zchaan quest but deliberately neglecting the opportunity to find out what is going on here? Do they set aside the urgency of challenging Zchaan for the sake of learning more about events in Gnomvark, and the gnome-Titan history, and perhaps gaining some advantages? Do they want to confront the emissary, or leave them free for now, and how would that affect their encounter with Zchaan?

I'm not saying it's the most coherent sidequest ever, but I think it works a lot better in terms of not breaking the momentum of events and maintaining focus on one primary quest, while still obliging the party to delay and deal with unwanted social activities that get in the way of crushing evil. However, it requires planning to build up the connections between different elements that make for a more natural progression between parts of the campaign.

Why is this a problem, again?

Yes, sorry, that did get rather long. You know I like my examples.

The problem is that urgency and weak connections combine to dilute the coherence of a campaign. If you are moving through a series of events devised by the GM (at least broadly), which combine into a genre-appropriate and dramatically-satisfying narrative, this gives you a coherent and probably enjoyable experience. If you're then suddenly obliged to transfer your attention to something else which is happening, but which isn't particularly relevant to the things you have been caring about so far, your emotional connection is interrupted.

Not only are you interrupted in your experience of what was a reasonably satisfying narrative by something else; you have to start more-or-less from scratch with the new focus, work out what it is and deal with it, and then try to regain your mental immersion (or momentum, or engagement) with your original narrative.

Each time this happens, you are spending a smaller proportion of your time participating in a continuous narrative arc which builds up emotional energy for a satisfying payoff. This is true even if the interruptions are really good. Your overall enjoyment of the arc is diminished because you keep losing that focus and momentum by experiencing it in smaller chunks interspersed with other things. If the interruptions instead detract from or even deprive you of things you hoped to do, which mechanical Happenings can easily do by changing the situation, you've also lost an opportunity.

Look, it's basically multitasking. If you're writing a long report and regularly have to stop to answer the phone, you'll lose your focus, your train of thought, your holistic grasp of the report. Your satisfaction in putting together a high-quality piece of work will be reduced, and it's likely (in my experience at least) that the quality will also be impaired, because you struggled to regain focus and enthusiasm after the interruptions.

The Creativity Tax

When a system mandates Interesting Consequences for X, each time X occurs, the GM must devise an Interesting Consequence.

Remember how I said player creativity isn't a finite resource? Well, that's not quite true. Players are, especially as a group, highly creative and able to come up with ideas (usually more than enough to keep a GM on their toes and occasionally throw a whole plot wildly out of kilter) but their creativity is not actually unlimited in the short term. Nor is GM creativity. Nor are attention span, concentration, or investment in a game.

One of the disadvantages of mandating Interesting Consequences for failed rolls, even when they are of a challenge type whose immediate consequences for failure are not inherently interesting, is that it costs time, energy and yes, creativity.

In theory creativity is not limited, but in practice, it kind of is. On a good night, I can improvise wildly and come up with content fluently enough that at least some of my players can't tell I hadn't planned it all beforehand. I can see the world in my mind's eye, and intuitively sense the natural consequences of some failure or oversight. I bubble with imagination and have only to pick the juiciest, most tempting morsel to suit the palates of my players.

On a bad night it's all I can do to have an orc slip in a pool of blood, a policeman happen to be strolling by, and a man burst through a door holding a gun.

Most GMs are (rightly, in most cases) proud of their creative streak, because that's partly what draws us to games and to GMing. But that doesn't mean our creativity is unlimited in the short term, or adaptable to every situation. And each bit of creativity is a tiny emotional investment in the world and its details, and another piece of information to grasp and remember, and that adds up.

Whenever I am obliged to improvise, I draw on the same pool. Some of this is passing improvisation: holding a dialogue with a player, or fleshing out the details of a location they chose to visit. There's also reactive improvisation, where I work out the natural consequences of player interaction with the setting, whether that's experimenting with a weird device, socially manipulating NPCs, setting a Klingon workshop on fire, building an orphanage or whatever. Sometimes they focus heavily on a colour element and I have to spontaneously make important decisions that will shape the setting: what exactly is the ideology of this murderous kobold cult, and how does it tie into general kobold society and other races' interactions with them?

Thinking up something that happens when Bob rolls a botch, a fumble, a critical failure, a GM Intrusion or whatever draws on that same pool.

I can't speak for others, but personally, I would rather be devoting my energies to inventing content that somebody wanted in the game, rather than content demanded by a piece of mechanics. If players display unexpected interest in something I threw out, that's great: I want to reward that, to roll with and feed whatever momentum is building, because keeping everyone (myself included) engaged is crucial to running a fun session.

Sometimes that momentum is in an unhelpful direction, but even then, giving players a tasty morsel that says "yes, you're allowed to interact with that" before gently directing their attention elsewhere feels better for everyone.

If players are doing something unexpected in our gameworld, then I want to have the focus to interpret and gauge the appropriate consequences of actions; to amicably and reasonably judge whether there is an X, or whether tactic Y should yield interesting results. If I can do that, then players will feel freer to play with the gameworld, to try things out, to predict the results of their actions and to expect their choices to seem meaningful.

I don't want to ever be in a situation where I've expended enough of my brainpower coming up with GM Interventions that I'm too tired to play along with what my players want to try.

Fractal Distract...als?

In extreme cases, the group can end up following nested loops of distractions from whatever they were originally doing.

Of course, this is by no means unique to random Happenings. Following up a series of consequences is fairly core to sandboxy play, as opposed to prewritten adventures. It often reflects what the players find most interesting. But when these consequences are improvised on the spot in response to die rolls, rather than purely narrative considerations, I believe they're more likely to distract from the premise of the game.

If consequences result from the decisions made by players interacting with the world, there's an extent to which the players are interested. To take up an existing example: if I decide to sever a supporting beam and everyone ends up covered in orc dung, then I at least will generally feel a sense of involvement in the whole Orc Dung Debacle that ensues. If I decide to reject the Belligerent Duke's plans and tell him he's a waste of oxygen, causing the whole party to be exiled, not only do I have an emotional involvement in the consequences, but I probably considered them at least a little and decided they were interesting before making that decision.

If what happened was that I rolled a 1 on my d20, causing dung and/or exile, that doesn't feel emotionally like something I was involved in. It feels like a random event, because it was. Even if the GM does a good job of narrating it and helping it feel natural, it still doesn't involve my choices.

Because mechanical Happenings tend to be relatively common, if they have quite far-reaching consequences it's not especially difficult for these to stack.

So you are engaged on a noble epic quest to something or other. You attempt to negotiate with the Duke, and roll a 1, and get yourselves exiled for a pilgrimage. While travelling to the Cathedral to do your penance so you can continue with the quest, you roll a 1 and get lost in the mountains. While seeking shelter from a storm before trying to return to the road the next day, you roll a 1 and are captured by goblins and taken miles away to the goblin labyrinth. While breaking free from the goblins so you can escape the labyrinth and get back on the road to the Cathedral, you roll a 1 and break a supporting beam and now everyone's covered in orc dung. While trying to battle the goblins covered in orc dung, you roll a 1 and lose your heirloom axe. While trying to recover your axe from the orc dung, you roll a 1 and realise it has been carried away by a goblin thief. While following the goblin thief across the countryside, you roll a 1 and are drugged and shanghaied by a merchant...

You end up going off on side-tracks to follow up consequences and their consequences, and not actually engaging very much with what you intended to do or the GM's original content. And not using the GM's precious content is fine (to a degree, I mean, I am always sad if I don't get to use something I thought up), but then there's an extent to which writing a setting and campaign premise is superfluous if the actual game will be substantially driven by spontaneous decisions based on random events. In our example here, the noble epic quest has rather degenerated into a Sinbad-style picaresque.

On Numenera, Interesting Failure can be very uninteresting Interesting Failure spotlights, can derail and hands the power back to the DM that those types of games so often claim they want to give to the players Imagination is a limited resource - where do you want to spend it? Arbitrary "interesting" failures, or Intrusions (see also Intrusions for Numenera critters), or on improvising content that reflects what the players are actually choosing to do? On making arbitrary events "interesting", or on supporting non-arbitrary events? Interesting Failures dwell on those failures and absorb game time that can be used for more gameplay. Bifurcating Failures tend to derail. Dealing with consequences of failures can lead to highly tangential games where it's difficult to follow through on anything.