Image of a volcanic eruption courtesy Pixabay.com/Life-Of-Pix

There are lots of good reasons to have a strong supporting cast in an RPG.

They permit interactions which reveal or highlight aspects of a PC that otherwise might get an infrequent airing, for one thing.

Trusted NPCs can serve as proxies for the PCs, or can supplement their skill-base.

Or, fourth, they can facilitate plotlines and subplots that would otherwise be impossible to deliver to a PC.

It’s the latter function that I want to focus on in this article.

The Splash Vector

There are any number of situations in which there’s a reasonably obvious right-or-wrong decision to make.

Sometimes, there may be more difficult choices on offer, in which a player character can be presumed to exercise due caution and do their homework, because of that degree of difficulty.

“Disasters never ‘just happen’,” as the tagline from the Canadian documentary series Seconds From Disaster reads; “They are a critical chain of events…” Break the chain, and the disaster doesn’t happen, or doesn’t have as severe an impact (depending on how late in the piece that opportunity is taken).

In essence, it takes a minimum of one error or circumstance to create the potential for disaster, and at least one more to transform that potential into a manifest reality. Even then, there may be multiple opportunities to discover the imminent catastrophe and mitigate, minimize, or even avoid it entirely; these have to be avoided if the catastrophe is to become both inevitable and to exact it’s maximum toll.

Which is what the GM usually wants to flirt with – the deeper into this chain that a twist can be inserted to turn catastrophe into disaster averted, the more dramatic the situation is, and the deeper the architect of that salvation has to become immersed in the plotline.

Last-minute escapes are more thrilling than situations which are easily averted.

The decisions involved in these chains of events also deserve some scrutiny. These are either overt actions or failures to act, and – if overt – can either be what would normally be the correct choice (but isn’t, for reasons unknown to the protagonist of the story), or can be an incorrect choice of action; either will deepen the crisis to it’s next stage. Even the right action taken too late, or not taken strenuously enough, can transform the danger of a potential disaster into an actual disaster.

But here’s the thing: if you’ve been able to engineer a situation in which the usually-correct choice of action will only make things worse, and a player can usually be relied upon to make the correct choice of action – for reasons I’ll get into in a moment – why wouldn’t you make them the central figure of the resulting drama?

It’s only when you can’t rely upon the Player to let his PC get into the situation all the way up to his neck, or when your plot objectives require a less immersive situation – another point I’ll get to in a moment – would you need to target a Supporting Character and draw the PC into the situation by virtue of the relationship between that Supporting NPC and the PC, i.e. to utilize the Splash Vector.

The diagram to the left illustrates all this in the simplest way possible. This catastrophe chain consists of 9 events, though the last three are variant outcomes of the same singular event, If I’m honest.

Event 1 is the initial error or condition. It defines the blue-to-green zone, which contains mistakes or circumstances. The green-to-red zone deals with discovery and response. Event 2 is discovery after the fact, and is the reason maintenance work on aircraft has to be inspected before that vehicle returns to the sky. Both of these, on discovery, lead to inconvenience or difficulty, nothing more.

Event 3 is where things start to get interesting. Something unexpected starts to happen; it has to be correctly assessed and diagnosed and the correct remedial action, if any, taken. At each point from event 3 onward, the correct choice of action leads to the previous outcome – so, getting the assessment and diagnosis right, and taking the right action, at Event 3 leads back to mere inconvenience or difficulty.

Getting to Event 4 means that one of the trilogy of actions at Event 3 was not handled correctly for whatever reason. It represents one final chance to resolve the situation before it becomes dangerous, or to discover and correct that Event 3 failure. Success leads to Worsening Difficulties, preventing the situation from escalating leads to Potential Danger. Making the wrong choice leads to Event 5.

Event 5 means that the situation represents a potential danger. Pilots and the masters of other vessels have a phrase, “Pan-Pan-Pan” (sometimes just “Pan-Pan”), which is used to alert others that the vehicle has an urgent situation that does not yet threaten the lives of those aboard or the operation of the vehicle itself. Handle this correctly and the best possible outcome is that the danger remains only potential; the slightest shortcoming in response (including simply taking too long to reach a decision) leads to the danger manifesting, and leads to the declaration of an emergency, and a mayday to alert others that assistance may be needed.

Event 6 arises from a failure to prevent the situation from worsening – if there was ever an opportunity to do so, of course. The best outcome from this point is an emergency satisfactorily resolved, with passengers, crew, and/or vehicle having been placed in danger but not harmed to the point of disaster. This event represents the final opportunity to prevent such a disaster, and margins at this point are usually razor-thin.

Events 7-9 represent disaster, with various degrees of mitigation. Loss of life is now inevitable, only its degree is now in the command of those ‘controlling’ the situation. If we’re talking about an aircraft, event 7 might be a successful crash landing with successful evacuation and no casualties but a wrecked multi-million dollar investment, or it might be less than a third of the souls aboard being lost. The maritime equivalent is the ship sinking but all the passengers and crew rescued. A “mayday” call means that the commander of the vehicle is anticipating an Event 7, even if that’s only a worst-case outcome. Event 8 is a moderate disaster – some casualties are expected, but there is also an expectation of some survivors. Event 9 means that there is no hope of survivors (though sometimes there are a few, anyway).

Some chains of events may have more opportunities in given stages, or none at all. The worst situations go directly from Event 1 or 2 to Event 7, with the crew never being aware of the danger they are in until it is too late to do anything about it. These were once commonplace, but have mostly been engineered out. These days, the worst case you can reasonably be expected to confront is a chain that runs Event 1 to Event 6 to Event 7-9.

Olympian Heights vs Confirmation Bias, Logic Errors, and Hubris

Putting a PC on the path of such a disaster chain – depending on how softly you define “disaster” – is a regular occurrence in RPGs, because the implication is that there is something that can be done about the situation – of the PCs are sufficiently quick-witted and aware of the situation.

They are aided in this respect by the GM, who usually issues multiple warnings of imminent cataclysm, and by the separation between player and character. Because they are not personally in danger, this presents the players with the perspective from Olympian Heights, and that alone can permit a more rational decision-making process than those that would be experienced if they really were in such a situation.

There are several basic responses to emergency situations:

1. Counterproductive Fight-or-flight
2. Denial
3. Confusion
4. Freezing
5. Panic
6. ‘We have to do something, now – this is something’
7. Incorrect action taken through misdiagnosis of the situation – at best, this only consumes time, at worst it makes the situation worse
8. Correct action in insufficient measure, too late, or both
9. The correct action to prevent or mitigate escalation

Now, how many times does a player have a PC choose anything but the last three – and with the full expectation that they are choosing the 9th and last option? It’s rare. Very rare.

That means that any of the other outcomes have to be applied to the character by the GM. Options 7 and 8 are usually tolerable, because that still leaves the player in command of the character, though there can sometimes be disputes about the character realizing in time that their action wasn’t having the desired effect.

Options 1 to 6 are less tolerable, even if they are reasonable responses for this particular character in this particular situation.

There are three major reasons for characters making the wrong choice of response – Hubris, in the form of overconfidence; Logic Errors which lead to incorrect or insufficient responses being made; and Confirmation Bias, in which we become so convinced of what we are doing that we can become literally blind to anything that contradicts out interpretation of the situation, and which I discuss in the context of players in this article.

Players are susceptible to each of these, but are less likely than their PCs to succumb because of the Olympian Perspective. (That perspective can also leave them more vulnerable to Confirmation Bias because it reduces the price-tag of speculation; most player Confirmation Bias results from the player assuming that his theory of events is correct, leading him to ignore as red herrings the evidence that the GM intended the players to use to disprove incorrect theories).

That means that unless you can be sure that there’s no opportunity to mitigate the situation prior to Event 4 at the earliest, PCs make unsatisfying targets for these situations.

Which brings me back to the Splash Vector, but first I want to talk a bit more about the choice of whether or not to expose PCs to a situation. In other words, I want to discuss a few more reasons why there’s a problem, and look at the scale of that problem, before I demonstrate the ways in which the Splash Vector can provide a solution to the GM.

Immersion Depth and Plot Significance

The more easily the correct course of action can be determined, the earlier in the catastrophe chain the sequence can be broken by an intelligent character behaving rationally.

It follows that the earlier a PC becomes involved in the plotline of such a situation in an active way, the lower the significant impact that plotline will have on the character, and the lower the player’s immersion within the plotline.

The logic given at the end of the previous section has already stated that only a few situations relative to the total pool of possibilities are definitely suitable for PCs. The reduction in impact means that some of the remainder can still be of use, with the assumption that the threat will be defeated at some intermediate point – Events 3-6, in other words. However, the immersion factor means that the earlier within this sequence that the intermediate point is reached, the less valuable the situation is to the GM.

As a practical measure, anything that can obviously be dealt with at Event 3 can be ruled out. Easily-solved problems at the Event 4 stage are trivial and also not, therefore desirable. Problems that can be easily resolved at Event 5 are therefore the minimum that are of value even as a subplot.

That’s at least half the potential plotlines that have been ruled out.

But it gets worse. Not every problem metastasizes into the next stage of the disaster chain. A huge number of potential inconveniences or difficulties never escalate beyond the irritation stage. I have no data on which to assess the relative proportions, but each Stage comprises more situations than the next higher Stage.

Let’s see what that means using 4 different values: 75% metastasize, 50% metastasize, 10% metastasize, and 0.5825% metastasize (I’ll show you where that last number derives from, along the way):
 

• 75% advance to the next stage, 25% are easily resolved at the current stage:
  • 75% of 6 become 7-9, so 6 is 4/3 of 7-9.
  • 75% of 5 become 6, so 5 is 4/3 of 6, which is 16/9 of 7-9.
  • 75% of 4 become 5, so 4 is 4/3 of 5, which is 64/27 of 7-9.
  • 75% of 3 become 4, so 3 is 4/3 of 4, which is 256/81 of 7-9.
  • 75% of 2 become 3, so 2 is 4/3 of 3, which is 1024/243 of 7-9.
  • 75% of 1 become 2, so 1 is 4/3 of 2, which is 4096/729 of 7-9.
  • So, for every stage 7-9 situations, there are 5.6 stage 1 events. Minimum. And about one in six mistakes leads to a disaster.
  • Another way of stating this is that for every 800,000 Stage 1 events, 200,000 are easily resolved and 600,000 will escalate to stage 2.
  • Of those 600K Events, 150,000 will be easily resolved and 450,000 will escalate to stage 3.
  • Of those 450K Stage 3 events, 112,500 will be easily resolved and 337,500 will escalate to stage 4.
  • Of those 337,500 stage 4 events, 84,375 will be easily resolved and 253,125 will escalate to stage 5, where they will be useful as a minor subplot or better.
  • Which means that useful plots comprise an estimated 31.64% of the total. And that’s with an unrealistically high proportion of escalation.
     
• 50% advance to the next stage, 50% are easily resolved at the current stage:
  • 50% of 6 become 7-9, so 6 is 2x 7-9.
  • 50% of 5 become 6, so 5 is 2x 6, which is 4x 7-9.
  • 50% of 4 become 5, so 4 is 2x 5, which is 8x 7-9.
  • 50% of 3 become 4, so 3 is 2x 4, which is 16x 7-9.
  • 50% of 2 become 3, so 2 is 2x, which is 32x 7-9.
  • 50% of 1 become 2, so 1 is 2x, which is 64x 7-9.
  • So, for every stage 7-9 situation, there are 64 stage 1 events – and 1 in 64 mistakes leads to a disaster. I don’t know what the error rate is during aircraft servicing, but there are an estimated 39,000 commercial and military aircraft in operation currently, so that would be about 2 newsworthy air disasters a day at a 1/64 rate. That seems to be about 1/365th or so of the true rate – call it 1/400th for ease of calculation, or 0.5825% metastasizing – which is where the fourth of the values comes from, but is getting ahead of ourselves a bit.
  • Another way of stating this is that for every 64,000 Stage 1 events, 32,000 are easily resolved and 32,000 will escalate to stage 2.
  • Of those 32K Events, 16,000 will be easily resolved and 16,000 will escalate to stage 3.
  • Of those 16K Stage 3 events, 8,000 will be easily resolved and 8,000 will escalate to stage 4.
  • Of those 8,000 stage 4 events, 4,000 will be easily resolved and 4,000 will escalate to stage 5, where they will be useful as a minor subplot or better.
  • Which means that useful plots comprise an estimated 6.25% of the total. And that’s with a still-unrealistic proportion of escalation.
     
• 10% advance to the next stage, 90% are “easily” resolved at the current stage.:
  • 10% of 6 become 7-9, so 6 is 10x 7-9.
  • 10% of 5 become 6, so 5 is 10x 6, which is 100x 7-9.
  • 10% of 4 become 5, so 4 is 10x 5, which is 1,000x 7-9.
  • 10% of 3 become 4, so 3 is 10x 4, which is 10,000x 7-9.
  • 10% of 2 become 3, so 2 is 10x, which is 100,000x 7-9.
  • 10% of 1 become 2, so 1 is 10x, which is 1,000,000x 7-9.
  • So, for every stage 7-9 situation, there are 1 million stage 1 events, with this progression rate, which is the one I instinctively selected as being “realistic” until I did the calculations in the previous results group, and hurriedly inserted a reasonably accurate value earlier into the article!
  • For every 1,000,000 Stage 1 events, 900,000 are easily resolved and 100,000 will escalate to stage 2.
  • Of those 100K Events, 90K will be easily resolved and 10,000 will escalate to stage 3.
  • Of those 10K Stage 3 events, 9,000 will be easily resolved and 1,000 will escalate to stage 4.
  • Of those 1,000 stage 4 events, 900 will be easily resolved and 100 will escalate to stage 5, where they will be useful as a minor subplot or better.
  • Which means that useful plots comprise an estimated 0.01% of the total. And that’s with a still-unrealistic proportion of escalation.
     
• 0.5825% advance to the next stage, 99.4175% are “easily” resolved at the current stage – calculated “realistic” values:
  • 0.5825% of 6 become 7-9, so 6 is 1717x 7-9.
  • 0.5825% of 5 become 6, so 5 is 171.7x 6, which is 29,472x 7-9.
  • 0.5825% of 4 become 5, so 4 is 171.7x 5, which is 5,059,554x 7-9.
  • 0.5825% of 3 become 4, so 3 is 171.7x 4, which is 8,685,292,912x 7-9.
  • 0.5825% of 2 become 3, so 2 is 171.7x, which is 149,114,663,056x 7-9.
  • 0.5825% of 1 become 2, so 1 is 171.7x, which is 25,599,083,786,461x 7-9.
  • So, for every stage 7-9 situation, there are roughly 25.6 million million stage 1 events!
  • For every 25.6 million million Stage 1 events, 99+% are easily resolved and 15 thousand million will escalate to stage 2.
  • Of those 15 thousand million Events, 99+% will be easily resolved and 8,737,500 will escalate to stage 3.
  • Of those 8,737,500 Stage 3 events, 99+% will be easily resolved and 5,090 will escalate to stage 4.
  • Of those 5,090 stage 4 events, most are easily resolved and 3 will escalate to stage 5, where they will be useful as a minor subplot or better.
  • Which means that useful plots comprise an estimated 0.00000000001171875% of the total. Or 99.99999999998828125% of possible plotlines are useless for the purposes of directing at a PC.
     

Fortunately, we don’t have to think of all those possible mistakes as GMs – we get to cherry-pick one that we can make interesting or relevant. The full list, potentially, includes everything from not storing enough soft drinks to forgetting to attach the engines when the vehicle was last maintained, and all points in between, plus every possible weather configuration of note, and the risk of collisions, and, well, anything and everything you can think of, quite literally, and more that you don’t.

Modern design and engineering has redundancy and safety measure heaped upon redundancy and safety measure. That’s why so few of those problems escalate beyond the inconvenience stage. But, if we broaden the concept of “disaster” to include adverse personal developments, there are far fewer protections. Nevertheless, the principle remains clear – more plots are almost certainly unsuitable for PCs than plots that are suitable. That was the case with every possible metastasizing rate that we considered. In fact, you need a rate of just over 89% escalation before it stops being true.

And some people think it’s easy coming up with Adventures that are interesting, internally logical, and engaging – just a side-note observation :)

The Splash Vector (cont)

The Splash Vector simply means that instead of targeting a PC directly, you target an NPC with whom they have an established relationship, who then asks the PC to step in and get them out of trouble at some point deep in the disaster chain. In other words, we target the NPC and hit the PC with the ‘splash’ so that the PC can’t dodge the problem by taking the relatively obvious escape routes early on in that chain.

Players can have no objection to NPCs exhibiting the full gamut of possible reactions to a crisis. NPCs are just as susceptible to Confirmation Bias, Logic Errors, and Hubris, as anyone else – and the GM doesn’t even have to require a roll to test for it. In fact, they can be “forced” to (mistakenly) make the worst possible choices until they have escalated a situation to the point where it will be difficult (and interesting) to solve without a disaster taking place.

A lot of those rejected plotlines – and they outnumber the directly-useful ones, remember – are suddenly back on the table.

Let’s take an example:

NPC has an accident of some sort and are subsequently sued. They hire a lawyer they saw on a TV advert, and don’t tell anyone about the accident or lawsuit. They lose the case and now have 30 days to come up with a substantial sum of money. They do the worst possible thing and borrow the money from a loan shark with ties to organized crime for what they think are ‘easy terms’.

That alone might be enough trouble for the PC to have to deal with, especially if the goal is merely to seed the campaign with the presence of a gang boss who is known to the PC – this is a great way to bring them to the PC’s attention. Or you might need to escalate matters a little:

The loan shark manipulates circumstances so that they default and then blackmail the NPC into doing illegal acts for him. Which leads to the NPC being arrested and charged. The NPC’s blind faith in people leads him to retain that same TV lawyer again. At which point the PC learns of the situation, and has to somehow extract his friend / partner / relative from it.

Or another:

NPC receives an email from a Nigerian Prince in exile who needs help in recovering 1 billion dollars in gold. The PC would know better, but the NPC is taken in – and has their money and their identity stolen.

That’s enough for some entertaining role-play between the two (provided that the PC somehow learns of what’s going on, probably through the boasting of the NPC), but the actual process of closing an old credit card and replacing it with a new one, etc, is likely to be more tedious than entertaining, so anything more than a roleplay can probably be hand-waved away. This is absolutely fine if having the NPCs identity be stolen is just an establishing condition for a bigger problem, in which case it’s better from the perspective of the GM trying to engineer “interesting times” for PCs that this PC remains ignorant of the situation for a little while longer.

…That identity is then used to acquire a credit card, which is used to purchase high-end electronics, which are resold – and, of course, no attempt is made to repay this money… Meanwhile, the NPC is so confident that he’s about to become wealthy that he goes into debt splurging and buying gifts for everyone. At which point the PC realizes that something is wrong, extracts a confession from his friend (face-palm when he learns of the Nigerian Prince) – just as the police show up to arrest the NPC for credit fraud…

It doesn’t matter what the trouble is, an NPC can either get into it more plausibly than a PC can, or can make a poor choice that makes a bad situation worse.

Splash Vector Requirements

Of course, for this to work, you need an NPC with the right qualities and personality, and you need to have established a relationship between the NPC and the PC in game time. The more remote that relationship, the greater the risk that the PC will say “No”, or decide that it’s too much work or too difficult a problem.

But it’s not for the GM to state that there’s such a relationship, or to dictate it’s depth and whether or not there is sufficient strength in it that the PC will put himself out so much for the NPC. That’s all up to the player.

That means that the GM can’t take the relationship for granted. Instead, he should ensure that the relationship develops by involving both PC and NPC in mutual events in-game prior to drawing on that investment.

There are times when you can drop an NPC into the plotline “cold” with an alleged relationship to a PC and have it accepted as the plot hook, but it feels forced, because it is.

As a general rule of thumb, if the relationship is a personal one, even if it hasn’t been established in-game, a personal crisis is justifiable – a relative having gone missing, or being held hostage for an impossible ransom, or whatever. If the relationship is not, a personal crisis probably won’t play, but if the promise of “adventure” is high, you have your next best chance of pulling it off.

Once an NPC is established in-game – and you’ll never know for certain whether or not they’ve been established enough until the time comes to pull the trigger – anything becomes possible. If the relationship is not yet enough, the previous rules of thumb stand.

Confidence

One technique is to use these facts to boost confidence in the relationship, via a simple four-point plan.

1. Embed the NPC as a regular figure within the campaign. Have them interact with the target PC on a number of occasions.
2. Use the rules of thumb given in the previous section to engage the PC in a high-adventure personal crisis, i.e. one that could be safely used with a drop-in NPC.
3. Have the NPC provide significant assistance to the PC at some personal cost, repaying the debt incurred in (2). The (2)-(3) combination greatly deepens the relationship.
4. If the response to the events of (3) give you confidence that the relationship is ready-to-use, go ahead with the real test of the relationship. If you are still lacking in confidence, return to step one (interactions) to more deeply embed the character.

Of course, you will have to be subtle about it – this plan falls apart if there’s any whiff of orchestration involved.

A broader field of opportunity

For that reason, I will often embed several NPCs into a campaign with no intent to use any of them for anything in particular – then wait to see which ones “take”. This gives a broader field of opportunity, because it means that I can pick and choose which relationship is most “ripe for the picking” at the time that I need one.

There are a couple of indicators that can be usefully employed to measure that readiness. If both you and the player can name the NPC without looking it up, that’s one sign that they are entrenched within the campaign. If the NPC is fun to play, and it has become easy to involve him or her in some way, that’s another. The combination makes it even more certain.

Of course, you will need to customize these indicators to suit your group, and the way that they play. A player who is naturally good at remembering NPC names (or who maintains a list of them) will obviously discount the efficacy of the first of those signals that I mentioned.

A little cold-blooded prep can also go a long way – deliberately seeding the campaigns with NPCs whose relationships you expect to need at some future point, i.e. deliberately emplacing the NPCs that the future plotlines will require. If nothing else, this gives you the chance to see whether or not those necessary NPCs “take” or if you need to tweak them or even replace them.

Think of this as giving some method to your madness when it comes to casting choices. Having a direction is always a good thing!