In this answer I suggested that "the majority" of NPCs would be 0-level characters, and was told by our friendly moderator @BrianBallson-Stanton in a comment (since deleted) that D&D 4 didn't have (or wasn't supposed to have) level gaps.

In AD&D (2nd Edition) campaigns that I ran and played in with my friends, in the absence of any written canon guidelines, we made assumptions about the relative levels of the NPC population of the worlds that we called the "Fifty-Percent Rule"; that the majority (i.e. 50%) would be level 0 characters, and of the other 50%, 50% of them would be Level 1, and so on, so that:

Level 0 : 50%
Level 1 : 25%
Level 2 : 12.5%
Level 3 : 6.25%
Level 4 : 3.125%
Level 5 : 1.5625%
Level 6 : 0.78125%
Level 7 : 0.390625%
Level 8 : 0.1953125%
Level 9 : 0.09765625%
Level 10: 0.048828125%
Level 11: 0.0244140625%
Level 12: 0.01220703125%
Level 13: 0.006103515625%
Level 14: 0.0030517578125%
Level 15: 0.00152587890625%
Level 16: 0.000762939453125%
Level 17: 0.0003814697265625%
Level 18: 0.00019073486328125%
Level 19: 0.000095367431640625%
Level 20: 0.0000476837158203125%
and so on...

Referring to the answer I gave, for a D&D 4e caster to be able to cast Raise Dead, a Level 8 spell, they would need to be 8th level at minimum, i.e. in the top 0.390625% of the population. In a city with a population of 10,000 (Not unreasonable for a D&D-type world), this would be 39 individuals of any character type of Level 8+. Given that most of them (say, 90%) would be unable to cast the spell in question, there may be at most three or four (3.9) who could actually Raise Dead.

My question is:

Is this a reasonable breakdown of the population distribution of experience levels in a D&D campaign (any version)? Is there some canon reference in any version of D&D that states an actual level-population distribution?


3 Answers 3


Not for modern editions. It varies by edition assumptions and by the campaign tone, furthermore, if you're looking for a natural distribution, you should be using the power-law distribution instead.

At the end of the day, one of the most critical questions in a D&D campaign is "how high magic is this?" And, if the world-designer is being methodical about it, it boils down to a question of "how many high-powered individuals are there in the world?"

Changing the ratio makes for a very different "seeming" world, as for the earlier editions, the narrative of the world changed with power.

In 4e, my experience is that the system doesn't care about simulating the world. The narrative focuses on antagonists and challenges appropriate to the party level. Thus, skill checks scale as the narrative scales, which means that the narrative dictates if there are casters able to perform certain rituals. With that said, Dragon 397 with its hireling rules certainly suggests that a party can get hirelings of their level, and scales costs appropriately.

If you're looking for a naturalistic representation of XP, assuming anyone can earn XP, level distributions should follow a power-law curve.

  • \$\begingroup\$ Incidentally, the model the OP mentioned isn't too dissimilar to the one provided in the 3.5 DMG. Second edition AD&D didn't provide a detailed breakdown in the core, but it did mention the assumption that the majority of people in the world were 0-level characters with a profession-appropriate set of non-weapon proficiencies. I don't know what other editions did, though. \$\endgroup\$
    – GMJoe
    Commented May 9, 2014 at 6:29
  • 1
    \$\begingroup\$ @Brian, I like your reasoning, but your answer then begs the question: how, even in a general way, would a GM go about calculating a population breakdown using a power-law curve? How would the GM's decision about "how high leveled should this campaign be" be expressed numerically in a power-law calculation? \$\endgroup\$
    – Monty Wild
    Commented May 11, 2014 at 22:03
  • \$\begingroup\$ That's a system specific question. I recommend asking it as a separate question. \$\endgroup\$ Commented May 11, 2014 at 22:16

I'm not aware offhand of any such breakdowns in true (TSR/WOTC) D&D editions, but Adventurer, Conqueror, King (aka ACKS, a B/X-derived system) approaches this from an economic angle, considering costs of strongholds, tax or business incomes, etc. and comparing them to the "XP for gold" model and comes up with the following numbers:

Level  1: 1 in 20
Level  2: 1 in 50
Level  3: 1 in 150
Level  4: 1 in 375
Level  5: 1 in 1,000
Level  6: 1 in 3,000
Level  7: 1 in 8,000
Level  8: 1 in 20,000
Level  9: 1 in 60,000
Level 10: 1 in 160,000
Level 11: 1 in 450,000
Level 12: 1 in 1,200,000
Level 13: 1 in 3,250,000
Level 14: 1 in 10,000,000

(The table stops at level 14 because that's ACKS' level cap.)

There are also lists for such things as the likely level of a given area's leader/lord, expected numbers of NPCs of each level in realms of various sizes (ranging from a Barony of 800 people to an Empire of 28 million), etc.

Note that ACKS assumes that a substantial majority (a bit over 90%) of the population are level 0 and, in fact, are not even capable of becoming leveled characters. Your world may vary.


This is definitely an edition and milieu-specific question. Personally, I like the 50% rule you summarize here and could see it working just fine, but some editions expect more leveled characters among the NPC population (4e) and others even less (1e and 2e). For instance, here's what the 1e DMG has to say about "Typical Inhabitants" on pg. 88:

"The bulk of the people met on an adventure in an inhabited area - whether city, town, village, or along the roads through the countryside, will be average folk, with no profession as adventurers know it, and no special abilities for clericism, fighting, magic, or thievery. They are simply typical, normal people (as you define typical and normal for the milieu, of course)."

Following that is a mini-table showing hit point and combat ability stats based on sex and general classification (sedentary, active, or laboring). That write-up suggests even less than 50% of the population are leveled, although Gygax does allow for differences based on milieu.

  • \$\begingroup\$ I seem to recall the "Typical Inhabitants" section from early D&D editions. Even though I have never played later editions (except as computer games), the little bits I have read appear to suggest that in the latest editions, non-adventurers can have levels that represent their advancement in crafts, trades and the like. Would I be correct in that? \$\endgroup\$
    – Monty Wild
    Commented May 11, 2014 at 22:08
  • \$\begingroup\$ @MontyWild You would be correct, yes. Skill gain and max skill level are tied fairly strongly to level in post-2e editions, making it difficult to create a RAW expert diplomat who's not also got a ton of hit points and a greater-than-0 attack bonus. Whether that's a good thing or a bad thing is, of course, a matter of personal and playstyle preference. \$\endgroup\$
    – GMJoe
    Commented May 12, 2014 at 6:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .