Today’s Save Vs. Rant is about the design philosophy of NPCs in RPGs. While every game system has its own guidelines and rules for the design of characters (some unifying the systems for PCs and NPCs, some treating the two completely different), the basic philosophy behind designing NPCs remains the same across the entire spectrum of RPGs, and in today’s episode, we’ll discuss different approaches, exploring their strengths and weaknesses.
Erwin Schrödinger was a physicist who, in an attempt to disprove the apparent absurdities of quantum mechanics, designed a famous thought experiment that describes the behavior of fundamental particles. The simple version is, you put a cat in a box with a flask of acid, with the flask attached to a detector that would check if a certain radioactive particle had decayed, then either break the flask (killing the cat) or not depending on the results. Until the box is opened, the closed system is unobserved, and can be described as being in both states at once – the cat is alive and dead (although, if you don’t open the box soon, you’re just going to have a cat that’s dead two different ways).
Physics is, apparently, more fun when you’re abusing imaginary animals, but the experiment proved to be so successful that it’s recognized as a pretty solid illustration of how these systems actually work, thus allowing Schrödinger to join the ranks of Albert Einstein in the coveted category of “scientists who tried to disprove something and accidentally proved it,” which is pretty crowded in the quantum physics world. I digress…
When we talk about an character being Schrödinger’s NPC, what we mean is that any element of the NPC that is hidden from the players (or simply not written yet) can be ANYTHING up to and until the moment that the players discover that information. Any given beggar is a disguised fairy, former criminal, down-on-his-luck former spice merchant, ruthless assassin and slumming royalty until he does something that serves to conclusively prove (or disprove) one of these theories.
Now, virtually any NPC is going to have some elements that are not yet defined and, if need be, the DM will reasonably fill these in as appropriate. You probably aren’t going to decide if each and every NPC you create prefers sweet or savory snacks, is a cat person or a dog person, believes they have ever seen a ghost, has credit card debt or knows what “dabbing” is, and while many of these might be inferred from other information about them (any random middle class character probably has credit card debt, an ultraloyal Russian gangster probably prefers dogs, and so on) for most characters, these factors can exist independent of any other information about then. What we are suggesting, is that, until such time as you have to answer these questions, you get in the habit of thinking of them not as being “neither” or “undecided,” but as being “both.”
The practical element of this suggestion is that instead of simply deciding – on the fly – which way the answer tilts, consider – even as a kind of thought experiment or exercise – exploring multiple options, and only locking in one once it becomes clear that it is necessary to do so.
So the Russian gangster lets slip that he has a bet named “Sasha.” There’s no reason, unless it becomes clear that you have to answer the question, for you to decided that Sasha is a cat, a dog, or even a trained black bear or python until you have to. There might even be scenarios where it is desirable to decide the answer based on what is best or worst for the story, and base the answer on outside information.
Suppose you were thinking that Vladimir (a naming decision I just made) was a dog person because, y’know, loyalty. What would it imply, though, if Vladimir owned a cat instead? Suppose that a cat doesn’t seem to suit Vladimir’s personality. You could easily rationalize that Vlad just has different tastes than some others in his position, and there’s nothing wrong with that. But, perhaps, Vlad has the cat for a reason that does tie into his loyalty. Maybe the cat belonged to a former partner (romantic or otherwise) who has died, been imprisoned or simply gone away, and Vlad is caring for the animal out of a sense of duty. The player characters are looking for a way to ruin Vlad’s reputation with his boss who, it is revealed, is currently in hiding. Perhaps the cat would not take well to the boss’s method of concealment. Perhaps he’s constantly on the move or in an airplane or another environment that would be uncomfortable or impractical for his feline companion. When the player characters begin their research on how to ruin Vlad, one of their contacts drops a hint about the cat’s special relationship. Now, Vlad is a cat person – or at least his pet is a cat – and we’re locked in, but up until then, we might abruptly decide that it’s better for the story for Vlad’s pet to be a dog, or the python, or a very talkative parrot.
Now, some may recoil at what we’re suggesting – after all, isn’t it cheating to change your mind about things to suit the player’s decisions? And the honest answer is, it certainly can be, and to that end, it’s important to discuss why “cheating” exists in RPGs.
The well known Rule 0 of RPGs is “The DM Is Always Right,” but this rule is definitely to be interpreted in the spirit that with great power comes great responsibility. Different groups will, of course, have their own comfort levels with things like rules calls, narrative flexibility and the oft-maligned “cut scene” privilege, but there are two aspects that serve as a very good test:
- Will this make the players doubt that they are playing a game?
- Will this make the player characters doubt that their actions have consequences?
One common issue that assails beginner DMs is having an entire story planned out in their mind, and knowing what, exactly, will happen in that story. While a lot of DMs have specific stories they want to tell and, furthermore, specific events with predetermined outcomes. Anyone who has played in a game like this knows that, regardless of how well written it is or how engaging the narrative, it is an underwhelming experience. One common way in which DMs railroad these sorts of games is through the exact kind of Schrödinger’s NPC we’re discussing. The wizard who always has the right counterspell. The expert who just happens to be skilled at every field. The evil warlord that flawlessly predicts every one of the PC’s methods of infiltration so he is never off guard (though let’s circle back to this last one in a minute). These situations, which undermine the dice rolls, words on the character sheets, and rules in the rulebooks, make the game itself meaningless in pursuit of desired outcomes.
But, even if you don’t do it for this reason, you can separately run the risk of ruining the consequences of actions. You can, for example, make up an NPC’s stats, possessions and abilities as you go, but if you do this specifically in response to, for example, an unexpected attack (“this important NPC is currently wearing a shield ring!”), success on an unlikely pickpocketing check (“You rolled a natural 20 to pick the captain’s pocket? Um… there’s nothing in there.”) or merely an undesirable choice (“You can’t capture the werewolf, because he possesses the power to shed his outer coat and escape your grasp!”). When these things happen to frequently, players start to notice the pattern and realize that certain things are going to succeed – or fail – to get them the desired outcome regardless of the results. If you use Schrödinger’s NPC too frequently to persecute or assist (regardless of whether you’re doing this for one or to PCs or the group as a whole), the players begin to realize that things are always going to turn out a certain way and may lose interest. Again, we might consider our evil warlord where he might have prepared for the specific attack types the PCs possess – having immunity to whichever element the PCs use in a fight, suddenly having a battalion protecting their rear when the PCs attack from there, or simply having the exact counter-spells prepared for each of the PCs spells. The players might rightly realize that preparing or approaching the battle with the warlord creatively is futile and stop considering other options.
But even the exact same scenario can work given that the explanation is believable and the outcome still feels like a game. Consider, once more, our warlord. If you were trying to portray a warlord of astounding intellect who is known for being a tactical genius, the PCs might very well find any possible plan they consider being foiled by a foe that flawlessly anticipates their plan. To finish this illusion, perhaps the players could still circumvent (or at least, believe they could still circumvent) this characteristic through in game wheeling a dealing. Perhaps they could deceive him, or apply a tactic that he is known to have a weakness for, or simply catch him off guard through careful observation. And, as a final sneaky note, this doesn’t even have to be the case, the players just have to really, truly believe that this is the case. Consider these two narrative scenarios:
The player characters, in infiltrating the warlord’s castle, determine that they can enter the warlord’s castle clandestinely through the catacombs – an option the DM suggests among a group of other more obvious sounding routes of attack. The catacombs are, however, full of undead raised by a necromancer the warlord hired for the purpose of attacking the PCs as they approach the castle. The necromancer and his 5 skeletons and 2 zombies are slain by the PCs, who emerge from the catacombs into the waiting arms of a half dozen guards. They sound the alarm, and the whole castle is on high alert. Throughout the rest of the adventure, the PCs are attacked by roving groups of troops, drawn the to PCs. In all, the PCs have to evade and defeat 4 bloodthirsty patrols. Finally, the get to the warlord’s chambers, where he, his three bodyguards and his general are ready to do battle.
This course of events is likely to frustrate your players and make them believe that there was no possible way to attack the warlord at even the slightest advantage. Now, consider this second scenario:
The player characters, in infiltrating the warlord’s castle, discover references in the records of the warlord’s noble family to a family catacomb that had a long-forgotten entrance for the now-defunct priests of the local death god to enter to bless the dead. They search for and discover, the entrance and begin to infiltrate through the catacombs. While in the dungeon, they come upon an unfriendly necromancer who has taken up residence in the dungeon. They defeat him and his 5 skeletons and 2 zombies, eventually discovering the entrance to the castle. When they emerge, two guards – the customary, but largely unnecessary defenders of the catacombs – confront them but are, likewise, defeated. As they continue through the castle, they have to contend with 4 sections of the castle watch. Finally, they arrive in the warlord’s chamber where he, his general and his three bodyguards are scheming.
Exactly the same encounters, flavored (and potentially positioned) slightly different, but the stories are completely different. The first feels like each encounter is just a sequence of uphill battles to the goal, while the second suggests that this is an improvement over what would be expected in a frontal assault. While discussing exactly what you can do to help maintain that illusion of choice and sense of accomplishment is a rant for another day (and we will definitely rant), the basic guidelines of making it feel like a game where choices have consequences is the gold standard for making that determination.
This blog post is, of course, really only scratching the surface of Schödinger’s NPC (which could be discussed in regards to narrative, rules, worldbuilding or even just the incredible I know a guy rule, courtesy of LA Based Twitter user David Nett), to say nothing of the other aspects of character building, and, our schedule permitting, we hope to touch on this topic more.
Next up on Save Vs. Rant – economics in games! We promise it won’t be anywhere near as dry as it sounds!