Beyond Unreasonable: Challenge Failure Modes
All players expect to be thrown into the deep end from time to time by the GM, with no idea of how to solve the conundrum that confronts them. Most commonly, this results from characters not having the skills that would solve the problem via the most straightforward course. Instead, the character has to take what he does know and add some creative juice to extend what he does know into the area of what he does not.
More embarrassingly, sometimes these situations come about because the character misunderstands the scope or scale of a challenge that they know lies before them. Overconfidence is a sure way to find yourself in a truly sticky situation.
Either way, the player is forced to rely on the GM’s vested interest in not aborting the game prematurely. That’s usually a safe bet; while GMs love to cultivate reputations as absolute fiends who make players scramble for every gold piece or experience point, most of that reputation is – to say the least – inflated.
- It is one of the GM’s responsibilities to ensure that there is always a way out – though I make no promises that it will be a palatable choice.
- I won’t kill a PC on a die roll under normal circumstances.
- One exception is when a Player is making a Heroic Sacrifice of his character. That always needs to be respected.
- That rule goes completely out the window when the player makes a mistake and ignores all hints and warnings regarding the decision. Stupidity is always fatal – eventually – and lifts all responsibility for turning into a Killer GM from my shoulders.
- With one exception: I won’t end any other players’ fun because one player does something idiotic. No matter how illogical on its face, I will find a way to bring the wrath of god down upon the character who’s player has earned it, while limiting ‘splash’. Even if that means foregoing the aptly-deserved punishment until a more opportune moment presents itself.
- And another exception: even if the players were all in on the decision, and equally culpable, I won’t let the campaign be killed by a foolish decision. I consider a lingering torture to be a more appropriate response. I know, I’m a softie.
- A consequence is that I have an oft-unspoken deal with players: I won’t use save-or-die spells or technologies if they won’t, unless I can be reasonably confident that for a particular reason, the spell won’t have any effect, in which case it just speaks to NPC villain personality.
My personal philosophy is more complex than that of most GMs simply because they haven’t put any thought into this area of metagame rules.
For example, let’s take your typical D&D Fighter, give him a reasonable level of intelligence, and drop him into the control room of a nuclear reactor about to go into meltdown. Obviously, he has no idea of what the myriad of flashing lights and dials and meters and push-buttons and knobs means. However, it’s reasonable for him to interpret alarm bells as an indication of trouble, and klaxons as an attempt to bring the attention of someone to a problem that needs immediate intervention of some sort. But that’s where the freebies end.
Were I the player controlling a character placed in this situation, I would be frantically looking for ways to make some interpretation of my metagame knowledge regarding reactor controls available to my character, and relying on the leniency of the GM to find any half-reasonable suggestion acceptable as a way of moving the plot forward.
I would start by asking the GM if there are any chairs. I would then ask, based on my knowledge of basic proportions and the inhabitants of the world with which the character was familiar, what race are the chairs designed to best-suit? Let’s assume that the response is “human”. I would then take a seat and ask about any writing – I should be able to determine what language the characters appear to derive from, even if what they are saying is beyond me, or so I would argue to the GM. Perhaps an Intelligence roll, I will suggest.
This is actually a trap for the GM, establishing the principle that an INT roll and a reasonable justification permits that translation of metagame knowledge that I mentioned. But it’s an entirely reasonable suggestion, and there are no obvious grounds for denying it. One INT roll later (presumably successful) and the answer comes back that most of the characters seem to resemble human script.
Now, it’s time to use that principle to give the character a fighting chance at saving both himself and the situation. I inform the GM that my character is going to assume, based on these two findings, that standard human conventions will apply – left or anticlockwise for ‘less’ or ‘off’, green for ‘safe’ or ‘go’ and red for ‘not safe’ or ‘stop’, unless he can give me good reason why my character would not.
“I know what a meter is, because there’s one to measure pressure on a pressure cooker and I have a high cooking skill,” I will next propose to the GM. This is a relatively flimsy line of argument that could easily be blocked by the GM simply by asserting that while that may be true, Pressure Cookers have not yet been invented, so my Cooking Skill doesn’t let me know anything about them. But I’ve already established some momentum of agreement with the GM, and I’m clearly trying to work with the situation he’s presented, with some solution in mind (which brings curiosity into the equation as another factor on my side), so I would expect a cooperative GM to let it slide.
“Okay, so I’m going to extend the logic of human conventions to all the meters that I see. I’m looking for red on one side, or green on one side, and using that to estimate which side represents safe operation of whatever this huge machine is,” I tell the GM. “If there’s no such indication, I’ll assume that left is good and right is bad. I have no idea what buttons are, so I’ll ignore them unless these are the kind that stay depressed when you press them.”
This is another invitation to the GM to take a step on the slippery slope of being co-operative. If he accepts the invitation, he’ll tell me that these buttons are indeed that kind, or at least some of them are, which justifies another INT roll to deduce the basic operating principles of buttons from those examples, then generalize that to include the other buttons. If he’s not ready to let me off that easily, he’ll tell me ‘no, none of the buttons are of that type’.
This approach works because when the GM puts a character into a situation, he expects that character to do something about it, with the ‘something’ being defined by the education, skills, personality, and capabilities of the character. It usually doesn’t matter too much whether or not the ‘something’ is what the GM expected, so long as he can see that it’s a reasonable choice under those conditions and given those parameters. Since that’s clearly what I’m doing in this case, the GM is more likely to play along than not.
So, now it’s time to apply some simple logic that will give the GM a chance to bail me out of the situation that he’s dropped me into. “I’m looking for the panel with the greatest number of red lights and meters indicating danger, as it’s the most likely to control whatever function has the problem. I’m locating the nearest control to each meter or red light. If it’s near a meter, I’ll push that button for a second and then push it again to return the control to where it was, and see if the light goes out or the meter dips toward safety or gets worse. If it gets worse, or doesn’t seem to do anything, I’ll move on to the next one. If it’s a knob, I’ll turn it a tiny bit to the left and see what the meter does. If it seems to help, I’ll turn it slowly until the meter is in the green, or at least out of the red, or it’s all the way to the left. If it seems to make things worse, I’ll turn it slowly to the right in the same way. When I’ve done everything I can with this panel, I’m repeat the process at the next worst, and so on. If none of that solves the problem, I’ll look for writing and try to decipher it.”
This outlines an entirely reasonable course of action, supported by logic, and has already inclined the GM to consider it satisfactory. But it gives the GM every excuse in the world to kill the character out of hand – revealing themselves to be a bad GM, because this is a really unreasonable situation to drop such a character into – or to decide that somehow, the character finds the right combination of control manipulations to solve the problem.
As this example shows, whenever the GM puts a character into a situation, the GM has to have answered one vital question about the situation: What does the GM expect the character to be able to reasonably do about the situation?
Some General Principles
Whenever there is only one valid course of action, it needs to be pretty obvious what it is – and that, too, is a sign of a bad GM, railroading the characters, though sometimes that’s permissible, especially when it gets the character into an adventure that would otherwise not take place.
In most cases, there should be at least two identifiable alternative choices for the character to select between. And that, by definition, means that there has to be something that the GM expects the character to at least attempt to do about the situation, and a path to salvaging events if a required die roll would result in catastrophic failure.
GMs should spend at least as much time and effort making sure that any given challenge or encounter has a way for the PCs to cope with the situation that they can reasonably find as they do making that challenge or encounter as difficult and interesting and, well, challenging, as it can possibly be.
That’s not always an easy thing to do, because it pulls the GM’s thinking in two mutually-contradictory directions.
A Practical Approach
My technique for solving this problem is to employ iteration and step-wise refinement, one of the ‘magic bullets’ that are always the first tools that I reach for.
- Consider the challenge that you are proposing. Is it difficult enough to be challenging? If it is, move on to step 5.
- If not, think of a way to increase the difficulty level. Before doing so, however, ensure that there is still a reasonably obvious course of action open to the PCs.
- If the condition is met, incorporate the refinement and then return to step 1. If the condition is not met, consider whether or not there is a way of making the course of action more obvious.
- If there is, incorporate both refinements and then return to step 1. If not, return to step 2 and think of a different complication.
- Are there at least two different solutions to the problems being presented for the characters to choose between?
- If not, is this one of the reasonable exceptions to that requirement?
- If it isn’t, then you need to introduce circumstances that will enable a second fairly obvious approach to be taken, and which will permit the time for the players who will be involved to identify both solutions and choose between them. When you have done so, return to step 1, as the whole challenge needs to be re-evaluated in light of the change.
- The challenge is now both difficult enough to be challenging and yet provides a choice of approach by which the PCs can express their individuality. But before moving on to the next challenge, you should make some notes on the possible solutions and how the encounter will unfold if the anticipated choices are made.
- It’s also vitally important to be aware of the motivations of those characters who are causing the challenge to take place and who are participating in it, because those are the guidelines that will enable you to improvise if and when the players choose a third path. Which they will, at least some of the time (if not most of the time).
Using this procedure, you start with a simple situation and keep complicating it until it achieves a difficulty standard that you deem appropriate. I want to talk about that for a moment.
An appropriate difficulty standard
I apply vastly different standards in the beginning, middle, and end of an adventure. In fact, one of the reasons for breaking a campaign up into discrete adventures is to facilitate this differentiation.
Early encounters & situations
In the beginning, I am more willing to lead characters into a situation by the nose, or by the railroad track, because if I don’t, there might well be no adventure. It’s one thing to bait one or more hooks and throw them out for the players to nibble on, but you need them to swallow something or the day’s play will be a big, fat, nothing. The best solution is to have a default defined, an adventure hook that will come looking for the characters if they are still hanging around nibbling and won’t give them a chance to say no. But this isn’t always possible, and it produces a rich but very complicated campaign structure.
You get a far simpler structure when you can simply parachute PCs into an adventure regardless of any plot trains they might have to catch – so long as their final stop is a destination that forces them to make significant choices and get involved. The other advantage to this approach is that you can integrate briefing materials into the plot train, and even view the plot train as doing nothing more than leading the PCs, breadcrumb by breadcrumb, to these necessary briefing materials. This line of argument contends that the players are unable to make properly-informed judgments until they have received the briefing from the GM, and that’s a completely valid line of argument.
Most adventures are more like a game of snakes and ladders in some respects – parts of them are locations where you roll dice (or make choices) to progress, while other parts represent a long way around (a backup that comes into effect if you don’t find and choose to use the most efficient ‘ladder’), and still others are as inevitable a complication or setback as a snake. Take one of those paths and – until it plays out by delivering you to somewhere that again permits forward progress – and events can be as inevitable as the tracks of a railroad, or a pebble rolling downhill.
The artistry in structuring an adventure lies in preparing only those parts that you need in order to convey the action and choices available to the players while containing any restriction on their choices and preparing for the choices that they are most likely to make, and avoiding wasting time preparing for parts that you will not need because they aren’t part of that pathway.

The left part of the figure shows the way we like to pretend our adventures are, a sort of idealized image. “Ladder” elements are in blue and – if successfully navigated and all the right choices are made – shortcut or provide a viable alternative path through the narrative. “Snake” elements are red and yellow, and represent a setback – potentially all the way back to the beginning – if all the wrong decisions are made, especially due to ignorance because the PCs took a ladder to somewhere near the resolution. The figure on the right gives a truer picture, with extremely messy and largely improvised passages of play, some benefiting the characters and some not. Again, if every possible decision is correct, even though made blindly to at least some extent, the adventure can be short-cutted dramatically – but the closer to the finish line you get, if you haven’t followed all the steps and understood their significance despite the GM’s best efforts to surprise players with plot twists and originality, the more likely you are to end up going backwards. I’ve also indicated in green the boundaries between early, middle, and late in the adventure.
In The Middle
Middle-section challenges are far more ‘pure’ in terms of the absence of railroading. Players have to get involved in challenges as a natural outgrowth of decisions made by them or natural consequences of the evolution of the situation in which they have found themselves, both interacting with the GM’s need to pose challenges to the PCs that will force them into choices. My goal as GM in these situations is therefore simply to advance the plot, and as a result I am even more forgiving and willing to be cooperative if a player puts up a reasonable line of argument than at any other point in the adventure.
It must be noted that there is a difference between thwarting a chosen course of action and failing to advance the plot. If the PCs go down a dead-end, there should be some way to navigate back to where they were when they made that choice – perhaps with the handicap of the enemy’s plans being further advanced, or other options being curtailed as a consequence. These aren’t failing to advance the plot, they are simply adding setbacks and resolutions to the plot.
Heading For A Climax
When adventures are approaching a climax, however, the safety nets start to disappear. The die is now cast, and the choices made, and the players have to live with the consequences with far less protection from me as GM. In particular, stupid choices in this phase of an adventure can be extremely hazardous. By this point, I have made sure that the players have access to everything they need in the way of briefings to make properly-informed decisions.
Reasonable and Unreasonable Challenges
This article is rapidly approaching its conclusion. The best way to get there is to return to the example offered earlier, of the D&D character confronted with an imminent meltdown, and consider how that challenge would be handled differently at the beginning, middle, and end of an adventure, and whether or not that makes the challenge one that’s beyond reasonable.
As a way of getting a character into an adventure, it would require the GM taking an extremely generous position for it to be reasonable. The character should have every right to expect to survive the situation. So long as the GM is as willing to engineer that survival just as much as he has clearly engineered the encounter, it’s fine.
In the middle part of an adventure, some of the safety nets are gone and unless the situation is clearly the result of a player choice, the encounter is an unreasonable one, even if the GM adopts a reasonably cooperative attitude. That’s because it’s stretching credibility to the extreme to put a character into that situation and expect them to be able to find a reasonable course of action. There is no way of solving the problem without a cooperative attitude on the part of the GM and such an attitude, carried to the extremes necessary, erodes verisimilitude. As an encounter, it goes too far. Having the PCs discover a reactor that’s working perfectly well and not having its controls protected against inadvertent and ignorant control manipulation is also unreasonable, because such a character would have no reasonable justification for knowing to leave things alone.
As the climax of an adventure, however, the reactor heading for meltdown is entirely workable, provided that the GM is cooperative enough to enable any reasonable actions on the part of the characters – and, more importantly, not to punish any reasonable actions undertaken from a position of ignorance.
Failure Modes
That’s as far as most expert GMs go. But there’s one final consideration that I want to bring to your attention, one thing more that only the most elite do when preparing an adventure, and that is to consider the Failure Modes for each challenge.
When you think of an encounter before it takes place, think for a moment about all the different ways that it could go wrong. That includes being boring, being too easy or too difficult, having a flaw in it’s logic (which will inevitably be discovered only after the players are already committed to it), and failing to achieve any overall plot objectives that you have for it.
Each such ‘way of going wrong’ is one of two things: non-fatal to the encounter, or a Failure Mode.
Non-fatal Errors
If something goes wrong, but it doesn’t terminate the encounter, kill a PC undeservedly, kill believability, or make it impossible for the encounter to achieve its plot objectives (if any), it is a non-fatal error. It’s a complication – perhaps not one that you expected or intended, but one that’s taking place nevertheless. This type of complication is called ‘non-fatal’ not because it doesn’t kill anyone (though it may not), but because there is a way to recover from the mistake.
Finding and implementing that recovery path immediately becomes the GM’s top priority. If necessary, you should admit to that being an unexpected move or an unexpected complication, and take a five minute break to think about it – and by “think about it” I mean consider your recovery options.
Failure Mode
The term ‘Failure Mode’ comes from aviation, and it means all the ways a part or procedure can fail. The design objective and desire is always for these components and processes to ‘fail safe’ no matter what the cause of the failure might be. Not all of that terminology is applicable to RPG adventures, but enough of it survives translation to make the term a reasonable one – especially if ‘fail’ is restricted in definition to the things that can go wrong that aren’t a non-fatal error – i.e., that can terminate the encounter, kill a PC who doesn’t deserve to get killed off, that eviscerates your campaign’s believability, or that makes the encounter worthless because it is no longer able to achieve its plot objectives.
Identifying Failure Modes
I always take the time to at least think about what could go wrong in an encounter. While this is possible to do in advance to at least some extent, you can generally do a better job at the last possible minute; that’s because advance planning requires you to make assumptions that may not be accurate forecasts, while last-minute analysis replaces those assumptions with campaign or adventure history, and the relative certainty that it represents.
That being said, having as much leisure time to think about a problem as you need can make advance planning a more attractive option – if you make general plans and avoid getting too wrapped up in specifics. This enables you to restrict your just-in-time analysis to a little a fine-tuning and to decisions concerning the best way of dealing with the failure modes that you might be presented with.
Preparing for Failure Modes
Preparing for a failure mode essentially involves setting up a means by which the ‘fatal error’ can be rendered non-fatal, permitting a recovery to take place. That sounds simple, but it often involves weakening the encounter by making it easier to resolve in a manner the PCs will consider satisfactory. To combat this, you either need to once again toughen the encounter or problem to compensate, or you need to find a way to conceal the potential for recovering from the ‘fatal error’. Sometimes it can be as simple as inserting a line of NPC dialogue into the middle of the encounter, to be uttered only if you need to prompt a PC to activate a solution that you had delivered to them earlier in a concealed form – “Whatever you are trying will not work. Only the Book Of Anthanalum and you do not even bear its’ hilt” – which tells the PC that this encounter is equating a book with a knife, and that the empty hilt that’s in his backpack somewhere (if used correctly) might provide an answer to the challenge.
The odds of needing more than one Failure Mode recovery ‘prop’ in an adventure are relatively low, so you might even be able to designate multiple roles for the one ‘tool’ to play, depending on what you need at the time.
Responding to Failure Modes
I tend to think of such approaches as incorporating a plot ‘toolkit’ into the adventure that I can manipulate as needs present themselves. In general, I favor broad principles rather than specific solutions, enabling me to customize them into a specific solution when I need one. Salvaging an encounter with a Fatal Error, i.e. a Failure Mode, almost always amounts to a plot twist within the encounter in some shape or form, and those are better utilized only when you need to, and when you have all the specifics of the circumstances at your mental fingertips.
The only certainty about a Failure Mode is that doing nothing is not an option. That’s because there’s no such thing as a “Fail Safe” when it comes to plots – failures are either non-fatal or they have to be dealt with – and right now.
A Lesson From The Past
I learned these lessons early, when I ran a published TSR adventure in my D&D campaign revolving around a crashed space-ship of flying saucer configuration. The players damaged the ship’s systems badly enough that the central reactor was placed into a dangerous state, figured out what the alarms and klaxons and automated warnings were supposed to tell them, got the computer to tell them where the engineering section (not called that, of course) would be found, and attempted to bring the reactor under control using the methodology and lines of reasoning that I have described in my example. This occurred in the middle of the adventure, and so I found myself facing a conundrum – permit a premature climax to the adventure that left half of it unplayed, or risk the rest of the adventure being a total anticlimax.
I chose the first choice, but never forgot about it. Because it was up to me when sufficient damage had been done to cause a problem; I chose a standard that seemed appropriate at the time, but which had unintended consequences for the adventure. In retrospect, I should have allowed the auto-repair functions of the ship to at least delay the consequences being a problem until I got to a point where the adventure needed a climax, then had the situation worsen as some further damage became ‘the straw that broke the camel’s back’.
So the lesson is now there for everyone. When you pose a challenge, or set an encounter, always ask yourself what you expect the PCs to do about whatever situation they are going to find themselves in. Make sure that you have an answer to that question, and that what you are about to inflict is always a Reasonable Challenge, and your campaign is more likely to survive its encounter with a bunch of untamed and unruly players.
Discover more from Campaign Mastery
Subscribe to get the latest posts sent to your email.
Comments Off on Beyond Unreasonable: Challenge Failure Modes