How much character do you need? Or, to phrase it another way, how much character construction do you have to undertake in advance when creating an NPC?

You could simply have a random character generator throw up something to avoid any character construction, but the results never seem to mesh properly with the situation in which the NPCs are encountered. Actually constructing a bespoke NPC is infinitely preferable – but it’s also a lot of work, especially if you have to do a lot of them,

Today’s article was going to be about a shortcut that I figured out last week for NPC generation – a companion piece to The three-minute-or-less NPC (creating personalities quickly) and last week’s 3 feet in someone else’s shoes (getting into character quickly) – but a lot of what I had to offer was very similar to the technique described in an earlier article on the subject, The Ubercharacter Wimp – and furthermore, assumed knowledge of some techniques that I haven’t yet shared, which would have been awkward to any readers who aren’t telepathic.

So the intent of this article is to plug that gap, and talk about Partial NPCs – and the most important decision of all, how incomplete can they be while still being fit for their intended purpose.

The Continuum Of Construction

First, let’s establish a frame of reference. NPC construction can vary in depth from complete characters, as well documented as any of the PCs or perhaps better, at one extreme, to nothing more than a couple of descriptive notes – an idea for an NPC – at the other. These two extremes form the end-points of a continuum, a straight line with multiple degrees of completeness in between.

Character Concept
Character concepts come in three flavors. You have characters that are there simply because someone has to be doing that job, and the PCs are going to interact with that someone – the Guard at the Palace Gate, for example. You have characters that the GM thinks are a cool idea, or that use a cool ability, or have a cool magic weapon, or whatever, that are essentially a “cool gimmick delivery system” – “A Beholder with Ninja training sounds like fun”. And you have NPCs that are window dressing, on tap and ready to be used the next time the NPCs interact with someone – “Barfly Number Nine” – who are nothing more than a personality or circumstance, but who can be used to flesh out the game environment, interact with the PCs, and perhaps supply one crucial piece of information on occasion.

These three types, when that concept stands alone with no further development, is a character concept, Concepts can be one partial sentence in length or a short paragraph, or just a link to a block of text elsewhere in the case of someone whose sole function is to have a particular magic item.

The Complete Character
A complete character can be even more strongly defined and complete than a PC simply because the GM knows more about the game world (or can create more as necessary) and hence a complete background history can be included, as can notes on characterization, how to get into character, and so on.

The points in between
There are all sorts of points in between these two extremes in which you have a character who has been partially created or defined, but which does not yet reach the full standard, and this is important because creating more character than you need is a waste of prep time.

Actually, that’s not quite true. Creating more character than you need right now is an investment in prep time; if the character is going to turn up time and time again, then time spent now – when you have the character clearly in mind – can be more efficient than doing it later – if you have the time to spare during your current game prep.


So the completeness that is desirable and the completeness that is required for immediate purposes are two different standards, with the first being as high or higher than the second. In other words, the target should fall somewhere in between immediate need and complete character, and the differentiating factor is the likelihood of reappearance, because that likelihood is loosely related to future “immediate needs”, but everything that is more bang than you need right now is only “nice to have”, it isn’t essential. It’s something to be done if you can spare the time.


When I first started generating NPCs for my campaigns, I started with the character concept and then created the character to fit that concept in exactly the same way that I would generate a PC – I’ll go into more detail on that in a moment, because the details will become important. But almost by definition, this will usually create more NPC construction than you need right now, so as I grew in experience I came to realize that it was far better just to do what was immediately needed and then reverse-engineer my way to earlier parts of the character construction process if I needed them for something else later.

This is almost always possible (there are some exceptions, such as original Traveller, but I regard that as being a flaw in the game system design because it forces the GM to do more work).


This technique relies on the skill of prioritization – of knowing what exactly you are going to need right now, and what might be nice to have but isn’t essential. Decide what you need right now and later you can use the game mechanics to derive how the character gets those immediately-useful scores.

Method 1: Natural Progression

“Natural Progression” is the same sequence of construction steps that you would expect to be carried out in the construction of a PC, which is the same as the way the rulebooks describe the process. As a general rule of thumb, it can be broken into seven steps:

  1. Concept/Personality
  2. Stats
  3. Skills or Abilities
  4. Abilities or Skills
  5. Equipment
  6. Flaws, weaknesses, disadvantages
  7. Personality revisited – history, background, etc

Complete all seven and you have yourself a complete character.

Clearly, if you only need the character to have one specific ability, there’s a whole bunch of work – steps 2, 3, 4, and most of 5 – that is completely wasted. What’s more, some of these steps have subdivisions, and sometimes you don’t need the whole step to satisfy your immediate needs.

In terms of any intelligently-designed Partial Characters technique, this is about as inefficient as you can get. Let me show you a better way…

Method 2: Functional Progression

“Functional Progression” performs character construction by deciding the desired outcome from any earlier steps of the character generation process and worrying about how you get there later. In other words, the steps of character construction are performed in the order needed to deliver the character’s practical function within the plot and nothing more. What this means will become clearer as we proceed.

combat vs roleplay

It didn’t take me very long to realize that two different forms of Functional Progression were needed – one for characters designed to be roleplayed in an interaction with the PCs, but that weren’t likely to be involved in combat, and one for characters designed to fight the NPCs but not to do much meaningful talking.

The Combat Model

The Combat Model of a Functional Progression divides character construction into fourteen steps:

  1. Concept/Personality - the starting point that identifies this as the model to use
  2. Attack & Defense - in D&D / Pathfinder, these are To Hit and AC. In the Hero system, they are OCV and DCV. They define the character’s chance to hit with an attack and his chance to be hit.
  3. Damage Capacity - How much damage can the NPC take?
  4. Equipment/Damage - What does the character use to inflict damage and how much damage does it inflict? In the Hero System, this is also where you decide PD and ED. Note that you don’t care WHY the character has the scores that he does, just the end results.
  5. Key Abilities - Does the character have any combat-related tricks or abilities or superpowers? What can they do?
  6. Key Flaws, weaknesses, disadvantages - Most game systems don’t go in for defining these. The Hero System and GURPS do. I like to enunciate them even in those games that don’t need them. Key Flaws are personality traits that can lead to the character making a mistake in battle, Key Weaknesses are holes in the character’s defenses (which he may or may not recognize) and Key Disadvantages are anything else that might hinder the character in combat.
  7. Key Skills - Any skills the character might have that are likely to make a difference in combat under the circumstances in which the character is to appear? This is likely to be a very short list.
  8. Other Key Stats - Many systems have other stats that you might need in a fight, like Initiative Bonus. Sometimes you need to know a DEX or STR check. Some abilities and game systems might dictate knowing the characters WILL or INT in order to deal with some forms of attack – but you only worry about those if the PCs have those attacks.
  9. Other Stats - Work backwards from the numbers you’ve assigned to derive the stats on which they are based, then set the levels of any remaining stats accordingly.
  10. Other Abilities - fill out any non-combat abilities.
  11. Other Skills - fill out any non-combat skills.
  12. Other Flaws, weaknesses, disadvantages - identify and detail any other manifestations of personality within the game mechanics that apply to the NPC that you don’t expect to make a difference in combat.
  13. Other Equipment - this is where you list any other equipment the character might be using.
  14. Personality Revisited - history, background, etc.

This list strips out just the things that you need to know before the character can fight and does them in a sensible order of priority, one that permits the GM to jump off the list at any point, or jump down to a later item if it is likely to become relevant. By definition, it excludes everything you don’t need to know, and focuses not on the mechanics of getting what you do need to know, but on allocating end results to those mechanics.

The Roleplay Model

You need to know different things when the NPC’s role is to talk to the PCs – possibly to relay information, possibly to react to information.

Bonus Tip: Your combat sequences will almost always feel less artificial if you can include at least one non-combat NPC as a bystander who will nevertheless get involved in the sequence in a non-combat way. Think of the classic Barroom Brawl in The Trouble With Tribbles with Cyranno Jones helping himself to the drinks and meandering through the fight. Think of a dust-up in a restaurant while the Maitre De squawks. Think of the tourist interrupting Superman in the middle of a brawl to ask for an autograph. Think of a seneschal offering snide comments about the combatants during a duel between a disliked courtier and a PC. It doesn’t matter what the fight is, or what it’s about, or where it happens – a non-combat element adds a touch of levity that adds immensely to the entertainment value of the fight, and makes it less about the game mechanics and more about the interaction between characters.

The list of elements in the roleplay model will look very familiar at first glance. They are just organized into a different order. But then you look at the definitions, and little subtleties begin to manifest themselves:

  1. Concept/Personality - Always the starting point, identifies this as the model to use.
  2. Key Flaws, weaknesses, disadvantages - These are all about the character’s personality and not his combat vulnerabilities.
  3. Key Abilities - Similarly, this is about what the character knows and what he can do in an interpersonal / information-gathering situation, and not what he can do in a fight.
  4. Key Skills – By now, the difference between the meaning of the term “key skills” in this model and the meaning of the same term in the combat model should be clear. Knowledge, crafts, and interpersonal skills.
  5. Key Stats - Int, Wis, Will, Cha, etc, depending on the game system.
  6. Personality Revisited - history, background, etc.
  7. Non-Combat Equipment - Anything that might help the character know or learn something or perform some non-combat task, that might help him seem credible in his role, etc. A Hat of Disguise probably won’t help you in a fight, but it’s vital to know about it in a roleplayed encounter.
  8. Everything Else:
        8.1 Other Stats
        8.2 Other Skills
        8.3 Other Flaws, weaknesses, disadvantages
        8.4 Attack & Defense
        8.5 Damage Capacity
        8.6 Equipment/Damage
        8.7 Combat Abilities

You don’t need to be especially observant to notice that the bulk of character construction has been lumped into the “everything else” category. That’s because, unless the character is going to enter combat on a subsequent appearance, and you know it, this stuff need never be done.

Oh, there may need to be the occasional highlight – “He’s carrying a mace painted like a clown’s face and wearing enchanted chain mail of some sort” – but you don’t need to worry about what these magic items actually are. Maybe the Clown’s Face would need some additional explanation, but you get the point…

Partial NPCs

A partial NPC is one in which the entire character construction process has deliberately not been carried out because most of it is unnecessary for this character being encountered in these specific circumstances.

Sounds simple, doesn’t it?

The Combat Model: Seven grades of NPC completeness

When it comes to combat, there are seven grades of completeness that I employ to decide how much of the character construction process needs to be completed.

  1. Flunkies - Next to no detail needed, and one common concept fits all. Combat Model elements 1-4. Optional but unlikely: element 5 if any; element 6 if unusual; elements 7 & 8 if they are likely to be needed. Same scores for all.
  2. Combatants - Described as an individual, otherwise the same as Flunkies.
  3. Major One-off Foes - Combat Model elements 1 through 7. Optional but unlikely: 8 if they are likely to be needed. Anticipating the likely need for this character to interact with the PCs in a roleplaying sense, I will also complete items 1-3 from the Roleplay Model.
  4. Lieutenants - What the character knows is more likely to be important, as are any limitations, so I will complete Combat Model elements 1 through 8 and Roleplay elements 1 through 6, but without going into great depth on roleplay elements 4-6.
  5. Arch-Enemies - Same as Lieutenants, but with more attention to the roleplay elements. An Arch-Enemy is a recurring enemy of one or more of the PCs, so there will be emphasis placed on the relationship with that character/those characters.
  6. Recurring Enemies - A recurring enemy is one that is expected to make multiple appearances in the campaign, but which is not tied to any individual PC. I treat them the same as an Arch-Enemy but without the relationship focus, and anticipate completing the process of making them Complete over time.
  7. Complete Characters - The only time I create a full character before that character has even entered play is when he is going to be an NPC ally of the PCs and part of their team. I consider it important that these characters be built to the same standards as would be expected of any other PC, and I am usually happy to consult with the best character-creating player on the course of their future development – while keeping some cards close to my chest, of course.

Most adventures will have one representative of groups D through F, maybe two in extreme cases. There may or may not be a couple of representatives from group B. Most of the NPCs who are expected to engage in battle will be group A.

The Roleplay Model: Seven grades of NPC completeness

I use a similar scale for roleplaying encounters.

  1. Casual Encounters - Next to no detail needed, usually encountered singly rather than as a group, though there are exceptions, which I will discuss below. Roleplay Model Elements 1-4, in as little detail as possible.
  2. One-offs of substance - A little more detail needed, so Roleplay Elements 1-4 in full.
  3. Recurring Encounters - More detail again. Roleplay Elements 1-4 in full, plus some quick notes on Elements 5 and 6.
  4. Intimate Recurring Encounters - These are the same as Recurring Encounters but with an important relationship to one or more of the PCs. Roleplay Elements 1-4 in full, a little more detail in Elements 5 and 6, and some notes on the relationship(s).
  5. Supporting Cast - Supporting Cast are NPCs who don’t fight alongside the PCs but who are frequent associates of the PCs. Roleplaying Elements 1-6 in full, and some notes on Element 7. I will also anticipate the possibility that the character will encounter combat at some point and do Combat Model elements 1-4.
  6. Central Interactions - Supporting cast members who are central to the campaign or the plotline, at least for a while, deserve a little more attention. Roleplaying Elements 1-7 in full, and just in case, Combat Elements 1-4, with notes on Combat Elements 5-8.
  7. NPC Team Members & Complete Characters - It’s very rare that full characters are needed unless they are NPC team members who live and fight alongside the PCs. In most cases, I will use the Roleplay Model “complete character” process rather than the Combat Model equivalent because I want these to be more than Flunkies to the PCs, I want them to be people with whom the PCs will interact. But there are occasional exceptions.
The Mob and other groups

As much as possible, I try to think of groups of individuals as a collective NPC. I do this largely without needing to analyze the logic behind the decisions I make, relying on my experience, a little education on group psychology, and a healthy dash of cynicism. What is the strength of a mob? Average plus one for each member able to contribute to one goal at the time. What is the INT or WIS of a mob? If they have no leader, it’s that of the lowest member – but any attempt to influence the mob has to reach all members of it, or they will resist attempts to persuade or dissuade them from whatever they think they need to do. Ditto the moral restraint of the mob. On the other hand, if they DO have a leader, these scores will be those of the leader, plus an X-factor for having the support of the mob pushing them on. And so on.

It’s so much easier to generate one set of stats for the group as a collective.

The Impact Of Game Balance

There are two major game architectures out there – the classic “character class” model, and the “construction points” model. Both the Partial Encounters systems blatantly ignore any attempt to balance characters in favor of fitting the character to the role that he or she is to play within the game, and I make no apologies for that. Ways can always be found to balance the books if you want to obsess over trivia, but there are better things to do with your time as a GM.

But that’s not to say that I don’t bear some reasonable standard of Game Balance in mind. The standard is always one that is relative to the PCs capabilities, and this is a key component of the initial concept. “Worse than the PCs” is a valid standard for grades 1 and 2 of the combat rankings, and for anyone short of full team members in the roleplay rankings. “Worse than the PCs except in their specialty” is another – though it then requires a further statement of the NPCs ability in their specialized field. “Better than the NPCs but slower to improve / advance” is another valid choice. “Better than any one PC” is a valid choice. “The equal of the entire group of PCs put together” is fair enough – for an enemy who is high up in the gradings, or for an NPC whose function is to serve as an advisor / mentor without doing the PCs job for them. The role in the plot is what dictates what I want an NPC to be able to do.

The Role of GM experience

The more inexperienced you are as a GM, the more you should treat NPCs as being one or two grades better than they need to be. It takes a lot of expertise to be able to judge correctly how far you need to go.

I know of one would-be GM who thought that to create an NPC that was the equal of the whole team put together, the right approach was to add all the PCs stats together and use the total. This completely ignored synergistic mechanics within a character’s construction, and non-linear progressions in ability, and a whole slew of other such factors. A far more balanced approach is to take the best score of the PCs in any given stat and add 1 for each PC after the first that he is going to oppose. And maybe +1 or +2 to some things because he is going to be facing multiple opponents simultaneously.

Greater experience also makes you more adept at winging it if circumstances propel an NPC into realms you didn’t expect them to go, so there is less penalty for underestimating the role that an NPC will play in the game.

The Initial Standard & Unpredictability

Even the most experienced GMs will get it wrong occasionally. The players will ‘take’ to an NPC like a duck to water, recruiting what was expected to be a one-shot NPC. A villain will be so much fun for everyone that he has to make a return appearance at some point. A character who was never expected to see combat ends up on the front lines. A character who should have been mincemeat gets a spot of luck and benefits from some clever thinking (on his part) and/or sloppy thinking on the part of the players and gets away, or successfully pulls off a Wizard Of Oz routine that makes him look far more effective than he really is. An NPC turns out to be tougher or more effective than he should have been.

There are so many ways to get it wrong.

The Build-as-you-go solution

As you become more experienced in GMing, if you aren’t already, you will discover the counterintuitive solution: Do Less and Build As You Go.

Even if you think a character is going to be recurring, build to a lower grade – the first time they appear. Then, if your prognostication turns out to be false, you’ve wasted less effort. If you are correct in your expectations, you can always bump them up a grade or two before their second appearance – and then again, before their third, and so on. Always keeping an eye on what you expect that NPC to be providing to the plot in those appearances, of course!

Don’t be afraid to skip ahead on the Model Hierarchy of elements as necessary. Build as much NPC as you need right now and extend on that when it’s warranted.

Flunkies should take one minute to create, two at the outside. Combatants and Casual Encounters should be two, maybe three minutes, at most. Work smart, work efficiently, and work hard – and you will work quickly.

I like to estimate how long generating a complete character to the required standard will take (in real hours and minutes). After the first couple of levels, I divide that into blocks of roughly 10-20% of the total – and that’s how much additional time I will invest in building on the basis of what I’ve done already.

Creating a first-level D&D character should take less than an hour. Maybe 30 minutes to do a complete one – with a full personality, history, etc. Creating a fifth-level character might take 2 hours. Creating a tenth level character, maybe 3. And so on. But those are how long it would take me – you might well be different. So use your numbers to allocate time. Do whatever you need to do in order to have the NPC ready for his function in the next adventure, and then – if the anticipated grade is high enough – spend whatever time remains in working on the next item in the appropriate hierarchy step of the Model. Yes, there will be exceptions to these rules – if you expect a character to be defeated, you might need to do a complete equipment list because the PCs will loot it, for example.

Apply the principle of doing what you need, just in time for you to need it, show a little love regularly to those NPCs who warrant it, and use a little common sense. At the VERY least you should be able to cut your character prep time in half. Or, more likely, to about 10% of what it would be if you created complete characters every time.


These are taken from the Pulp campaign that I co-referee. While you might not be familiar with the Hero System, it will be clear to anyone who knows that game that this is just about everything you need to know in a basic fight scene. What should be more obvious is the brevity, showing just how little you really need to create in order to manage a complex scene.

The scene is a nightclub with a number of patrons. Two different groups of NPCs, one led by “The Sikh”, are going to have a fight, with the PCs in the middle, not knowing whose side they should be on. The PCs have just rescued one of their number who was kidnapped by the owner of the club. That owner has left the club, leaving the execution of the PC in the charge of her Lieutenant, “The Sikh”.

Combat Examples

THE SIKH: Strongman, Huge 2-handed sword, think Raiders Of The Lost Ark: OCV 7 DCV 7 PD 6 ED 3 SPD 3 BDY 20, each attack can hit up to 4 tgts (one sweep) 2D6 normal + 2d6 HKA, easily distracted.

20 GENERIC TONG FIGHTER UNDER SIKH COMMAND: OCV 7 DCV 7 PD 4 ED 2 SPD 4 BDY 12 1d6 Normal + 1.5d6 HKA, will try to protect customers. Will attack PCs if they come within reach.

24 MARTIAL ARTISTS: Kung Fu style, OCV 5 DCV 6 PD 3 ED 3 SPD 6 BDY 10, 1d6 normal + 1.5d6 HKA, 3 throwing stars 1d6 RKA each, will try not to harm customers but will go through them to reach Tong. Will attack PCs if attacked.

10 POLICE: all armed with batons except Lt. who has a .38 Webley Revolver: OCV 3 DCV 3 PD 3 ED 1 SPD 3 BDY 8, 1d6 normal + 0.5d6 HKA.

So we have one group who behave a little more like good guys, but who will attack the PCs on sight, and another group who behave a little more like bad guys but who will not attack the PCs unless provoked. The Sikh tips the balance in the fight against the Martial Artists, but the PCs can tip it right back the other way – if they help the Martial Artists.

54 Flunkies, 1 Combatant, total time elapsed: <6 minutes.

Roleplay Examples:

Also at the “Jade Palace” tonight are:

  • 22 member Wedding Party: Bride, Groom, 2 sets of Parents (hostile to each other), 2 Bridesmaids, 2 Groomsmen, 12 guests, will panic, bride expects Groom to defend her; he will try but is hopeless;
  • Diplomat & Guests: 1 person from German Embassy, 2 businessmen, 3 female escorts, will overturn table and hide, will use escorts as distractions if necessary;
  • Diplomat & Guests: 2 people from Spanish Embassy, 1 Spanish Guest, 2 wives; most will attempt to flee, wife of diplomat wants to get involved in the fracas because they are ruining her night out;
  • Chinese Military Officer and Texan Arms Salesman: doing an arms deal for a “Pederson Device” which turns a bolt-action rifle into a semi-auto rifle. Chinese Officer will attempt to sneak out, Texan Salesman will pull a pistol and shoot wildly at anyone who gets in his way as he tries to escape; kill with a throwing star.
  • Shady Businessman: Immaculate white suit, 2 Bodyguards (.38 webley revolvers), one Floozy, is convinced that the Martial Artists have been sent by a rival, will attempt to escape under protection of bodyguards.
  • 28 General Customers with wives/girlfriends: Talking, drinking, watching show, etc – 1/3 will duck for cover, 1/3 will try to flee, 1/3 will just panic
  • 18 Patrons at the bar: Will run towards the stage and climb up, looking for a stage entrance to use for escape
  • 4 stage hands: will flee out performer’s entrance and lock it behind them, preventing performers from escaping.
  • 21 Entertainers: Will stay on stage, performing, will fight with patrons from bar – stage magician, 2 comics, an MC, 2 female singers, 8 chorus girls, band leader, 5 piece band, Chinese strongman in leopard skin with weights, big gong on stage with hammer, porcelain statues. Strongman will help the PCs by throwing objects from stage while using Gong as a shield.
  • 12 reserves: from whatever faction the PCs oppose (split evenly if PCs go after everyone)
  • 20 Staff (Generic Tong Fighters): waiters, kitchen staff, etc, all of whom are also Tong providing Club Security;
  • 10 Police: on stakeout outside, 3 rounds to make entrance after fracas starts.

152 NPCs, Total time elapsed: 30 minutes.

And a good time was had by all.

Related Posts with Thumbnails
Print Friendly