Monday, 28 April 2014

Pathfinder NPC levelling

Abel Grimmer 002

For quite a few years now I’ve been dissatisfied with the NPC rules shared by D&D 3.5 and now Pathfinder. There are a few problems with these rules that grate on me and I was really hoping for an overhaul in Path; sadly not. Since I’m running Path at the moment I’m going to do some playing about with the rules in the hopes of getting something more satisfactory.

The main problems as I see them are as follows:

  • NPCs gain hit dice per level as the same way as PCs. These hit points model ability to survive accidents, illness, and attacks. PCs gain this survivability through repeated life-and-death struggles that hone their ability to survive, generally involving vicious battles. Many NPC classes involve few or no such incidents, and therefore there is no reason for them to be particularly tougher with increased level. An expert tailor is no better at surviving a wolf’s attack than an apprentice tailor.
  • NPCs gain increased attack bonus as they level. This has the same problem as the previous point; outside “adventuring”, very few professions require combat skill, nor are they likely to increase it. Those civilians that have militia training, bar brawls, pirate attacks or bandits to account for can easily do so by multiclassing into a warrior class, or allowing an NPC feat or ability that models this.
  • NPC levels make PC levels narratively wobbly. As usually modelled, most NPCs would be 1st level peasants or warriors, which gives them a maximum skill rank of 4 (including class skills) in two or three skills, and a single feat. This excludes them from most of the profession-appropriate skill feats, such as Sea Legs for a sailor (5 ranks in Profession [Sailor]). It also means they’re liable to fail the skill rolls used daily in their job, which often have DCs from 15-20 and disallow taking 10. To reliably climb and balance at sea without falling to his death, a sailor would need to be about level 4, even with Skill Focus and class skill bonuses.
  • NPCs gain XP in exactly the same way as PCs – occasionally by fulfilling specific missions, but the vast majority comes through combat. This means that the very best way for tailors to improve their tatting and hemming skills is by bludgeoning goblins over the head with a mace. It also makes it virtually impossible for NPCs to gain any levels at all, keeping their skills at very low levels, even if their entire life is devoted to using those specific skills to stay alive.
  • A merchant requires Appraise 20 to determine the value of a common item. Let’s ignore high-pressure sales techniques, auctions, busy markets and other common aspects of mercantile life, and allow her to take 10. She’s still going to fail half the time, and a quarter of the time will have a wildly inaccurate idea of appropriate prices. Skill Focus, class skill bonus and a high Intelligence will give her a +8 if she’s lucky, which means 10% of the time she doesn’t know the value of the goods she trades professionally. Just in order to be competent at selling household goods, she’ll need to be 3rd level. Anyone working in high-pressure environments or dealing more specialist merchandise needs to be several levels higher.
  • Unfortunately, because combat skills scale with level, a competent used-horse salesman can readily beat up a gang of bandits or low-level PCs.
  • If we stick to low-level NPCs in order to preserve the combat balance, then we quickly reach a ridiculous skill imbalance. A 10th-level fighter with an interest in religion (i.e. putting a point in every level) is more knowledgeable than the average professional priest. A skilled thief whose interest in engineering is limited to avoiding or breaking it can ace the entrance exam for any engineering school, advise on renovation projects, lecture on architecture and design siege weapons, any of which is beyond many full-time engineers. An adventuring wizard, whose experience is largely practical, is more adept at magical theory than stipendiary research wizards, and better at crafting magical items than someone who earns a living that way.
  • Hit Dice and skills vary wildly and sometimes bizarrely between skills. The Expert is marked out from the Commoner by a threefold increase in skill points, military-level competence with weapons and armour, a larger hit die, and a good saving throw; yet academics are not generally hardier than labourers, blacksmiths are not necessarily more mentally stable than fishermen, and hairdressers are not yet routinely trained in the use of tower shields.
  • And let us not forget - making NPCs is basically exactly as complicated as making PCs if you attempt to do it by the rules.

The solution staring me in the face seems very simple: NPC classes do not grant combat advancement. There’s really no reason why either base attack bonus or hit points should increase through training in most professions.

Revised NPC classes

My version of NPC classes would look something like this:


The commoner is an ordinary member of society, occupied in low-skilled work, subsistence farming, fishing or similar generalised livelihoods.

Alignment: any

Hit Points: no increase.

Class Skills: The commoner's class skills (and the key ability for each skill) are Craft (Int), Knowledge (Local), Perception (Wis), and Profession (Wis). They may choose two additional skills from the following list: Appraise (Int), Climb (Str), Handle Animal (Cha), Heal (Wis), Knowledge (Nature), Ride (Dex), Survival (Wis) and Swim (Str).

Skill ranks per level: 3 + Int modifier.

Saving throws: 1/3 level.

Base Attack Bonus: +0

Class Features:

The following are class features of the commoner NPC class.

Weapon and Armour Proficiency: The commoner is proficient with one simple weapon. He is not proficient with any other weapons, nor is he proficient with any type of armour or shield.

The commoner no longer has scaling hit points or base attack bonus. Farming, digging, carrying shopping and fishing do not help you survive swordfights or bolts of lightning. Their saving throws still scale, as general life experience should still improve their ability to predict danger, endure hardship or muster willpower.

I’ve given them a slightly more fluid selection of skills, because I’m unsatisfied with the standard ones. Knowledge (Local) is incredibly basic and I’m astonished that the designers didn’t include it. I was actually very tempted to make it scale with class level. Ride is actually fairly unlikely for commoners (they don’t tend to own horses), whereas appraising goods, handling illness (without paying for a doctor), understanding natural processes (for farming) and general survival skills are all very useful for the average peasant. It also allows this class to represent a broader selection of characters, including hunter-gatherers and subsistence farmers as well as ‘generic mediaeval peasants’. Skilled hunters, however, I would give levels of expert or even ranger.

It’s perfectly possible to tweak these commoners to represent background; they have feats, after all. Typically I’d see these being used for skills, but there are alternatives, especially given a couple of levels. Toughness and Great Fortitude make a very hardy labourer. Lightning Reflexes and Dodge would suit a petty thief. Alertness and bow proficiency would cover basic hunting. For someone with militia training, a proficiency or two would make them more viable.

I’m not sure what to do about the favoured class benefits, as these give a way to increase hit points. It’s basically up to the DM at that point; I’d tend to use them for skills, but someone in a physical profession might gain hit points to make them hardier when dealing with nonlethal damage, and someone with a more dangerous job might justify a few extra hit points.

Note that these builds are not gaining hit dice at all. This means feats that depend on hit dice (such as Defensive Combat Training or Toughness) will not help. I can’t decide whether that matters or not. A DM could easily rule that they have a d0 and still gain the stated benefits.

A default commoner is simple to build: simply give them one point in Profession and Knowledge (Local) per level, plus one other skill appropriate to their lifestyle. A reasonable guideline would be one skill point in each per four years of age, including the +3 bonus for class skills. A 60-year-old herdsman has Profession (Herdsman) +15, Knowledge (Local) +15 and Handle Animal +15. His 12-year-old niece has three points in the same skills. For the most part feats can be totally ignored.

One more point: in mediaeval Britain, at least, commoners were famously required to train regularly with the longbow. Similar laws, with varying arms and armour, have applied in many regions at various times. As such, it’s entirely reasonable to argue that commoners should in fact be proficient with one martial weapon suitable to the setting. Between these kinds of laws, the threat of robbery and banditry, a taste for physical sports, and the prevalence of things like knife fighting, you could reasonably argue that in many settings, commoners’ skill with weapons would tend to exceed that of the middle and upper classes. This is quite the reverse of the normal arrangement.


The expert is an specialist, focusing on specific skills that require considerable training. Artisans, scholars, hunters, spies and pickpockets are all experts of one kind or another.

Alignment: any

Hit Points: no increase.

Class Skills: The expert can select any 10 skills to be class skills.

Skill ranks per level: 3 + Int modifier.

Saving throws: 1/3 level.

Base Attack Bonus: +0

Class Features:

The following are class features of the expert NPC class.

Weapon and Armor Proficiency: The expert is proficient with one simple weapon. She is not proficient with any other weapons, nor is she proficient with any type of armor or shield.

Saving Throws: The expert can select one saving throw from Fortitude, Reflex and Will, representing her specialism. This base saving throw progresses at 2 + 1/2 level, rather than 1/3 level.

This expert is considerably weaker than the standard one, which is exactly how it should be. An expert is a specialist in some area of craft, study or skill; they are neither warriors nor polymaths. They don’t need a large number of skills at a high level, except for sages, who should have high Intelligence in any case. The class skill bonus helps them maintain several extra skills at a respectable level without investing lots of skill points.

I have removed the armour and weapon proficiencies, on the grounds that there is no earthly reason for them to have any. Like anyone else, they can take feats or multiclassing to represent any special training, such as military service.

To give some more flexibility to the class, I’ve also allowed the good saving throw to be chosen. I suspect the original idea was that willpower is necessary to gain real expertise in a field; however, my build is intended to cover practical experts as well as academic ones, so that doesn’t necessarily apply.

A blacksmith expert could take Endurance, Skill Focus and Master Craftsman, allowing him to make and repair simple magical equipment. He might well take the bonus hit points from his favoured class to help withstand the heat, and would want a good Fortitude save. A spy might take Persuasive, Deceitful, then move into Deft Hands and other professional areas, depending what sort of agent she is; she’d want a good Reflex save. A professional hunter might take Nimble Moves, Acrobatic Steps and Alertness. A scholar would have a high Int to give plenty of skill points, and would probably tend towards with Skill Focus, perhaps multiclassing a level or two of wizard or cleric to represent particular specialities.


There is really no need for an aristocrat class. They are simply another kind of expert, in having specialist skills. The core rules differentiate the two through giving aristocrats set skills (unnecessary) and proficiencies as a fighter, which is entirely unjustified by their core competencies of “having rich parents” and, um... nothing. Any aristocrat who actually has combat training or military experience can use feats and multiclassing to reflect this - levels of fighter, or even warrior, would do the trick.

Essentially, the aristocrat feels like one of those classes that's over-specific. It seems to represent a vision of an aristocracy that's highly-educated, and has a lot of social graces, but also martial training that goes far beyond self-defence and duelling into the realms of the professional soldier. In practice, I don't think these ever really combined. Military aristocrats, those appointed for wartime exploits or used to the thick of combat, might have heavy armour training and be really good at killing, but are very unlikely to be cultured and erudite. Conversely, those who were highly-educated and versed in etiquette might sit on a horse giving orders, but were unlikely to know how to fight in armour. Regency aristocrats mixed smallsword and pistol training with polite society, running estates and amateur study, but wouldn't generally know how to handle actual warfare. The Renaissance scholar-types didn't fight. The aristocrat can be used to model any one of these types, but in my view gives too broad a set of competencies to all of them.

Warriors and Adepts

The warrior and adept are the most problematic of the classes.

The warrior, representing a professional soldier, can stay mostly unchanged. If we’re getting picky, I would personally remove Heavy Armour and Tower Shield proficiency, as these are specialist pieces of equipment, and have them bought with feats by any warriors who need them. I would probably also restrict their weapon proficiencies to Simple plus one Martial, again buying specialist equipment with feats. After all, what else do they need them for?

Adepts, though... it’s a very specific kind of class, and it’s doing several things. For one thing, it’s merging spell lists so that one class can cover ‘suitable’ spells from divine, primal and arcane magic. For another, it’s limiting spell progression so that you can have spellcasting NPCs that aren’t immensely powerful (though this can also be done by simply reducing their level). Finally, it strips out the complications and powerful secondary attributes of spellcasters: a druid’s companion, a cleric’s undead-turning and a wizard’s plethora of metamagic feats and spells.

Unfortunately, it has a couple of obvious drawbacks. One is that it’s aiming for a very specific hedge-witch niche that isn’t suitable for all settings or NPC roles. Another is that it has all the juicy BAB and hit points that come with levels, and this combines with the slow spellcasting progression to mean that any decently-skilled backwoods mystic can go toe-to-toe with a gang of bandits.

Adept options

The simplest solution is simply to strip out the BAB and hit point aspects of the class. This allows civilian communities to have cunning-men, wise-women and mystics without them becoming badass. They might be able to throw out a spell or two if bandits raid the village, but they’re not much of a threat.

What this doesn’t do is deal with the other types of civilian spellcaster. Not all wizards are going to explore dungeons or be sent to confront rebels, and not all clerics risk life and limb in the front lines. Any world with functioning magic is going to have spellcasters who just make things, who help build civic improvements, who wander in the woods communing with nature, who provide entertainment, and who generally focus on the everyday problems of the ordinary citizen rather than the desperate cases.

A fairly simple (but possibly heretical) solution would be to have either a Civilian template or parallel civilian spellcasting classes. They would have little or no hit point progression (favoured class points alone could represent the occasional dangers of their profession), and no BAB. This restriction applies to familiars and animal companions as well. The whole range of spells and spell levels would be open to them; however, most of their expertise would be spent on the theoretical or everyday spells that are not present in the rulebook, and thus their number of rulebook spells would be significantly restricted (probably to as little as one spell known per level they can cast). A tenth-level research wizard, as found in many large towns, is mostly occupied in highly technical divination spells with no combat application, and funds herself by crafting witchlights and enchanting farming implements. As a result, she only knows half a dozen spells that can be used to reproduce effects in the rulebook, and she’s probably not up to casting many of them in a day – too draining. This model allows you to have NPCs that can do just about anything necessary, without completely overshadowing the PCs. Sure, the town wizard can help track down the bandit lair, but she’s not going to hunt them down for you. The ancient hermit in the remote valley can tell you what’s spooking the wildlife and can calm the wolves, but he can’t have his pet bear eat the bullette, because he doesn’t have one.

I seriously considered disallowing most class features, to avoid some serious problems. A druid’s animal companion is extremely powerful at middle or higher levels, and a sorcerer’s bloodline abilities could give a high-level academic sorcerer abilities well out of sync with expectations. The abilities are too disparate to restrict in any straightforward way: a familiar or the sorcerer’s ray abilities aren’t particularly unbalancing, but an animal companion’s hit dice and BAB have the same problem as anything else, and the dragon sorcerer’s breath ability is dangerous in a way that other bloodline abilities aren’t. However, it’s not actually any more dangerous than equivalent-level spells and DMs ought to be making sensible decisions.

Also bear in mind that characteristic restrictions will severely limit the spellcasting capabilities of most NPCs. They tend to have lower stat arrays and are unlikely to reach much above 5th-level spells in the general way of things.

Are you experienced?

That seems to deal with the combat and skill aspects of the problem. A proficient NPC is now simply good at their job, without becoming a physical threat to the PCs, and can reasonably have high enough skills to overcome likely challenges. But how does an NPC gain these levels, if not through combat?

Well, the simplest option seems to be basing it on time. For simplicity, let’s call for a 100-year career. Obviously, most people won’t reach anything like that (at least most humans), which is just fine. However, someone who does is probably an expert in their job, whatever it may be. So we could simply say that after 100 years, the character is a 20th-level NPC. We can divide the necessary XP by 100, and say that earn that much every year by just plain doing their job. Unfortunately, the learning curve is far too curvy. An NPC would reach 8th level in their second year, 11th in the third year, then slow to an unbearable crawl. It would take seven years to reach 12th level, nine to reach 13th, twelve to reach 14th, twenty to reach 16th, fifty to reach 18th, and a hundred to reach 20th. This means that NPCs would be mostly clustered in mid-levels, but still far above many PCs. Not satisfactory.

Another option would be a bit more complicated. You could rule that NPCs each year gain experience equal to 10x their Intelligence. In the average progression, an average NPC would reach 2nd level after two years on the job, 5th level after fifteen years, and very few would survive to reach 10th level. Naturally, individual NPCs can have their experience adjusted however seems suitable; border-guarding warriors or particularly scheming aristocrats might gain experience at twice that rate. However, it’s functionally impossible to achieve anything above 13th level, even for dwarves and elves.

A third option is to work directly from age. Assume that achieving each level requires a number of years equal to that level. So a new recruit reaches 2nd level after two years, 3rd after five, and 4th after nine. Very few humans would reach 10th level, having to serve for 54 years and being about 70 years old at the time; however, it’s not beyond a dwarf or elf to reach 20th level.

A demographic option could also work here. A very simple rule, which models high mortality to some extent, is halving. About half the population are inexperienced young adults and children of 1st level. Half the remainder are 2nd level, and so on, rounding down each time, until you reach the maximum. This neatly limits level by community size, which is fairly likely if we assume that expertise tends to migrate to cities and that statistically you’re more likely to find an octogenarian in a city than in a hamlet. A small town of 1000 people would have a single 9th-level character, 62 of 4th level, and five hundred of 1st level. In contrast, a thorp of fifty wouldn’t exceed level six. You won’t find a twentieth-level character outside a city of half a million people, or at least a region with that many inhabitants. On this system, we could roughly assume that a world of fifty million people (given that much of it is wasteland and monsters) has about a hundred twentieth-level characters all told, many of whom are high-level peasants, politicians or merchants. This system is much more restrictive than the core rules, but I don’t know that that’s a problem. What this doesn’t do is specify the level of any specific individual.

The Third Way

The thing is, there is an alternative approach that bypasses most of these objections, which is not to faff about giving NPCs anything. It very rarely matters. NPCs without an adventuring or military background are not a physical threat, which is really where the complexity comes in. The important thing is that the average citizen, and even scholars, sages and civilian magic-users, have low HP and BAB. It doesn’t matter whether you assign a nominal 1st level or a nominal 10th level, and there’s no real reason to worry about it.

Thinking fairly logically, someone who devotes their attention entirely to natural history, magical theory, trading or farming is going to be decent at it. They are going to be much better than someone whose attention is split between that and mastering fifty different weapons, or learning ways to set people on fire, or studying the tenets of the gods. There really isn’t a huge problem if the village peasants have a +10 to their Knowledge (Local), or the wizard’s apprentice has 5 ranks in Spellcraft and your wizard only has one. That means they’ll pass slightly more skill checks than you. It’s not a big deal.

Now in some ways a weakness of D&D and its kin is that they don’t model past experience, which means a character whose background says they’ve been a mercenary for years doesn’t have any mechanical reflection of that. However, that’s just something we have to swallow, otherwise things will get massively overcomplicated or over-restrictive.

So my third way suggestion would be something like this. Note that these suggestions only cover generic characters. If you need a specific character with specific traits then just stat them out properly. This is for quick and dirty calculations; and remember, all we’re fundamentally doing is modifying a die roll slightly.

  • Where it actually matters, characters representing soldiers continue to take warrior levels.
  • The plebs have about a +10 total bonus (including ranks, stats and nominal feats) on skills to do with their role, and about a +2 on other things. On average they have about a +2 on saving throws, go down in a single hit (one way or another) and have a +1 attack bonus (including all modifiers).
  • Specialists have a higher bonus on relevant skills, varying with their alleged competence, and a higher appropriate saving throw. They won’t have a better BAB unless they use weapons, and they won’t have more hit points unless they need them. A really expert specialist with access to magic items might well have a +30 or +40 bonus in something, just like any high-level PC with magic items.
  • Spellcasters can cast a small number of spells appropriate to their interests, plus various nominal spells that don’t do anything mechanically. They have a higher Will save.


  1. It occurs to me that the sole reason NPC classes are in D&D 3.X/Pathfinder is because of the skill system; since the skill system is present, the designers felt you needed some means of assigning skills to NPCs, NPC classes were the solution they picked when they were putting together 3.0 and has stuck since.

    In pre-TSR versions of D&D you didn't have this issue. Random plebs were 0th-level randoms. Soldiers of more than baseline competence were fighters (because seriously, how hair-splitty do you need to get to have distinct "fighter" and "warrior" classes?), clergy didn't need a level assigned to them unless they had the divine nod, in which case they got some levels in cleric, wizards' apprentices were low-level wizards (kept in line by their master's housecat familiar, no doubt), and so on. If you needed to work out whether an NPC accomplished something which the game didn't have a subsystem for, you just made a judgement call; yes, this person is a peasant farmer, so there's no particular reason why they would fail to harvest this wheat.

    Then in the wake of Wizards taking over TSR there was a 2nd Edition supplement called Sages & Specialists, which really pushed the NPC class idea to the hilt. There'd been some classes published previously which were presented as NPC-only, but typically those were for the construction of villains or other character types not really suitable for being PCs but not exactly random goons either, and there'd been rules for how NPC sages go off and do research in the 1E DMG but Gygax never saw a need to bother assigning a level to those dudes. Sages & Specialists gave a ridiculous number of NPC classes, each with special abilities and their own experience table, and is the sort of supplement where you look at it and wonder whether anyone ever bothered actually using it in a campaign, because the amount of bookkeeping it called for was just daft - it might have been useful to craft one or two notable NPCs, but trying to actually give every blacksmith the PCs meet a level in Weaponsmith would just be impossibly tedious, and the provision of XP tables bizarrely implied that the GM would actually be tracking XP for NPCs, which is a level of work nobody who isn't a computer has time to do.

    In retrospect it was probably test-driving some of the ideas for 3rd edition, but I wish they'd just went with the 1E DMG approach of "here's some specialists the PCs might hire, here's some mechanics for what those specialists do". Your rules of thumb for handing PCs seem like they'd work fine in practice both in terms of adventure prep and actually running the dang game.

    1. I’m pretty sure you’re right about the design mentality.

      The Sages and Specialists thing sounds particularly OTT. I wonder if it’s just one of those blue-skies projects, where someone started thinking about how you could handle professions and low-combat campaigns, and ended up turning it into a full-blown supplement? Or maybe someone ran a campaign where everyone was just some guy, instead of an adventurer?

      I mean the thing here is, I can just about see how a certain type of person (not entirely unlike me) would value something for crafting “balanced” NPCs. Let’s say you’re running the classic “you were raised by the village blacksmith, and now this random bearded dude says you are special, also monsters are attacking” hook – if the blacksmith is going to be important and may well be doing some fighting, then maybe you want to model that in a way you find convincing. Or if the whole village gets dragged into a siege with goblins, where skillsets and NPC hit points might become significant, maybe that seems important to do in a consistent way.

      But yeah, especially in an AD&D framework, I’d be looking for something far more skeletal. Some guidance could still be useful though. Say:
      * Blacksmith: primary Blacksmith, Weaponsmith, Armoursmith, Fire-Building, Concentration; secondary Animal Handling, Smelting
      * Apprentice Blacksmith (20%/10%)
      * Journeyman/Village Blacksmith (40%/20%)
      * Town Blacksmith (60%/30%)
      * Master Blacksmith (80%/40%)

      And if you really want, a quick list of modifiers for particularly simple and difficult projects. Or maybe you do it the other way, having a base percentage chance and a list of modifiers by experience.

      Then you could (again, if wanted) offer some basic variation like:
      * Manual workers have slightly more HP as they’re used to physical hardship, maybe better saves
      * Older NPCs have a better chance to know facts, legends, people or places

      Basically, even for a non-combat D&D campaign, I'd be inclined to do most of this stuff with a simple set of rules that can be overlaid for variation. Maybe they just had fun doing it..?

    2. The other thing that occurs to me is, I wonder whether the whole “NPC class” thing didn’t actually emerge from PCs kicking off on NPCs, and some GMs being unhappy about it.

      So PCs are most likely to come into conflict with either civic leaders or town guards. They’re also the ones charged with keeping unruly PCs in line. At low levels this is fine because a handful of well-equipped guards are a serious threat to a whole party, and you know there’s a population behind them, so don’t kick off too much. Once you pick up a few levels, though – particularly with spells in the equation, and more so in 3rd ed. than earlier editions – the same old guards are a pushover. Clearly, the solution is to level them up, right?

      Except, in 3rd edition fighters gain feats pretty quickly, so a 6th-level guard would have about five feats. It’s probably not a balance problem, but it does make them fiddlier to run. Making them “warriors” cuts down on the amount of stuff they have. I can sort of see that. If you’re sold on a class-based model for everything, it makes some sense. Similarly, you can have “adepts” to offer a bit of healing and give the impression of village cunning folk who have some magical talent, helping with the fantasy feel, without filling the place with proper wizards. It avoids the awkward questions about why the town wizard doesn’t just handle the goblins.

      But they also like picking arguments with toffs, bartenders and the like, so it’s handy to have some idea of how tough they are. Wait, how tough is a baron? Baron von Graaf shouldn’t be a pushover for this bunch, he’d have, like, military training and stuff. And plenty of hit points, because reasons. And he’d know all about nobly sorts of things, so he needs some skills...

      And if you’re wedded to the idea of consistency, maybe it doesn’t occur to you that you can do this kind of stuff without allocating extra hit dice that grant more skill points.

  2. That's another nifty idea. I did always find it funny as well but I just *handwaviumed* it away.