Level Inflation is a disease even clerics can’t cure.

I recently read a Twitter conversation about high-level cleric characters in D&D and their effects on the game, and I thought about writing something on that. But as usually happens, I can’t tell the difference between things I have already written or just things I have thought about, because as it turns out, I already had a post about this very same thing. But it was in my drafts folder, from July of 2016, unpublished. For reference, this is the tweet that triggered memories of that three-year-old dusty draft:

So here it is, with some minor variations, the post I wanted to write three years ago but never did for some reason:

Level Inflation in D&D is a disease clerics can’t cure.

If you want to get technical, I guess “level inflation,” closely related to “level scaling,” can be described as what happens when the distribution of character levels in an RPG setting, game, or campaign stops following a power-law distribution and is forced to follow something more like a normal distribution. In other words, when the medium (most common) level is not lvl 0-1 characters but tends towards aberrations in the 7-13 range.

level comparison
Charts make you look smart

I’m not, however, trying to get very precise here; I’m just using those images as a nifty visual help.

With extreme level scaling, your character is always at the center of the normal distribution, no matter how powerful he gets. So, at 1st-level, your average sewer has only rats, but at level 20 houses an entire hive of beholders. And no matter where you travel, the entire world seems to scale to the level of your competence.

In D&D, level inflation may be the consequence of not understanding the significance of level progression, what those levels actually mean and entail, and the common error of assuming that level 20 (or some other maximum) is something that has to exist in a setting just because it’s in a table in the book. In fact, level inflation seems to be a common illness for those people obsessed with recreating “Settings.”

For example, if someone is playing in a Conan or Star Wars setting, it is always assumed that Conan, Darth Vader, or Luke Skywalker should be quite close to the upper tiers, towards the lvl 20 range.

Unfortunately, that transforms those iconic characters into walking gods and, therefore, everything else around them has to go through a level inflation upgrade if the original source material is to make any sense at all. So, the soldiers or giant snakes that Conan kills in this or that story, or the villains Luke kills, are not low-level NPCs (which is what they actually were) but suddenly they have to be reinterpreted as level 10+ characters for them to be meaningfully threatening.

Level inflation and Old School

As a general rule, OSR systems tend to focus more on low-level action. Even if the D&D forefathers also designed the rules to allow for god-tier campaigns, as far as I know, they actually never played them. It was just there to sell more books and for keeping the people always asking for more busy killing gods and demons.

Unfortunately, level inflation seems to be something that RPGs caught quickly, and D&D 1st edition already had some outrageous examples (although not in the main books.) Not surprisingly, though, it is something that seemed to pop up when people tried to play in some preexisting setting. For example, there is an article in Dragon Magazine #26 detailing the stats for Cugel the Clever, and whoever wrote the article gave him these stats:

Cugel the Clever

14th level thief? +3 leather armor? +2 rapier?!

Compare that with how the Dying Earth RPG describes its three types of game campaigns: Cugel-level, Turjan-Level, and Rhialto-level.

“A Cugel-level series features character who, thought they might be quite competent in two or three areas, rely primarly on their wits and luck to prevail. Events in the game will resemble those in the books The Eyes of the Overworld and Cugel’s Saga.”

A Cugel-type character is the average (as in, most common) campaign tier in that game, but in the older D&D version, level inflation makes a 14th-level character the new average.

The same happened to poor Conan the Barbarian in the early days of the hobby. In Dragon Magazine #36, page 10, Gygax tried his luck making a D&D version of Conan, and these are the stats he thought would be adequate:

Conan level inflation

Now, to be fair, this is D&D first edition, and there levels go from 1 to 30+ but it’s still quite absurd. I’m not an expert on the Cimmerian, but from what I remember about his stories, he mostly fought big animals, a few ugly monsters, and common humans. However, with those stats, I can’t see how any giant spider, gorilla, huge snake, or a few conspirators trying to murder him would pose any threat… unless you up their character level to keep up with him.

The dangers, if not silliness, of high-level characters were also obvious from the very beginning, which is why it was common practice among the original creators to retire characters when (and if, which rarely happened) they got to level 10-14 or so. In Dragon #45, Lewis Pulsipher wrote an article called “Ways to handle high-level headaches” and, in the end, he says this:

“But AD&D is a better game, and more believable, when the characters are not superpowerful and visitations of gods and demons are rare if not unknown. Somehow you have to neutralize the characters who, by hard work or good luck, have reached the rarefied heights of twelfth or fifteenth level.”

Neutralize. Not that you should adapt your campaign or your “world” to them, he says that you should neutralize them.

Two cars in every garage and a Jesus in every town.

What happens when level inflation is the norm and high-level characters are not neutralized? Well, “A Jesus in every town” is what happens because in every village or hamlet you will find a high-level cleric to resurrect your heroes for very reasonable rates. The problem is that high-level characters should be an anomaly, with each level increase reducing not only the probability of that specific character existing but also the odds of a similarly powerful character existing somewhere else.

In other words, if you want to keep the illusion of playing in an organic, somewhat “realistic” setting, it becomes increasingly more difficult to find a good challenge at the top of the pyramid. Hence the need (in video game-like campaigns) for level scaling, to artificially inflate the challenge so your “Chosen One” can go from 1 to 20, always doing the same thing (kill -> plunder –> sell stuff. Rinse and repeat, forever and ever.)

Now I’m going to run some numbers to calculate what are the odds of a cleric of a given level existing in a certain campaign. Now, first of all, we need the base rate. In other words, how many 1st-level clerics are among the general population. I’m feeling generous, so I’ll say 1 in 200. In any campaign that aims to recreate a sense of peril and adventure, clerics should be way more uncommon, though, but right now that doesn’t matter.

Now, it seems common sense to assume that there are more 1st-level clerics than 2nd-level clerics, and that there are more 2nd-level clerics than 3rd-level clerics. And so on. But what is the average ratio between levels? In other words, for every 2nd-level NPC, how many 1st-level NPCs exist? Any number could be used, but I think 1:4 is a simple but valid ratio. So, if there is one first-level cleric for every 200 (adult) people, that means there is a 2nd-level cleric for every 800.

The formula to calculate the ratio for any specific level is not difficult: 1/(200*4^l-1 ), where l-1 is the level of the character minus 1. So a level 2 character would be: 1/(200*4^1) or 1/800, as mentioned above. So, how many clerics with powers of resurrection are out there? It depends on the spell, whether we are talking about Raising the Dead or Resurrection, and the game edition, but I believe you needed to be at least level 9-10 to cast Raise Dead (I’ll choose 9), therefore:

1/(200*4^8) = 1/13,107,200

[I know that the real odds of finding a cleric able to raise the dead are slightly higher because you also have to estimate and add the odds for higher-level clerics (10th, 11th, etc.), but I want to keep this simple]

That’s almost one for every 13 million people. If the setting were a bit more conservative, like 1:1000 instead of 1:200 for 1st-level clerics, that number would be 1:65,536,000. To understand what that number means, around 400 or so million (including women and children) is the number of people believed to have been alive in the entire world in the 15th century. This means that, if you are lucky, there could be at most three or four people able to raise the dead in the entire world (and I’m sure they are not hanging around whatever craphole your adventures are currently adventuring in.)

[Note that this applies to clerics of all alignments and religions. So if you want to find a high-level cleric of a specific deity or alignment you may be out of luck.]

Now, out of curiosity, what are the odds for a 20th level character in that 1:200*4 setting?

Well, one 20th-level cleric for every 5.49755814×10^13 humans, or 1 for every 55 trillion people. That’s more or less 7,000 times the current population of Earth.

Note that this also applies to enemies. When you are level 2, finding 4th-level villains is not impossible. When you are at level 10, finding a 15th-level Archmage would probably require travelling to a distant planet.

Of course, you can increase your odds for higher-level characters by making level progression much easier and common, something like a 1:2 proportion, but that still would make a very high-level character relatively uncommon:

1/(200* 2^21) = 419,430,400

Here’s a table with the odds of a character class with a rate base of 1/500 in two assumptions,  fast progression (*2 multiplier) and slow progression (*4 multiplier.) I have also market numbers close to specific population milestones for a high-fantasy yet somewhat demographically depressed sword-and-sorcery setting: important city* (50,000 population), kingdom** (500,000), large kingdom*** (2 million), large empire***E (10 million), continent***C (30 Million), world***W (90 million,) and Modern Earth Equivalent-MEE (7,7 billion)

L41/40001/32,000 *
L61/16,0001/512,000 **
L71/32,0001/2,048,000 ***
L81/64,000 *1/8,192,000 ***E
L91/128,0001/ 32,768,000 ***C
L101/256,0001/ 131,072,000 ***W
L111/512,000 **1/ 524,288,000
L121/1,024,0001/ 2.097×10^9
L13 1/2,048,000 ***1/ 8.389×10^9 MEE
L141/ 4,096,000 1/ 3.355×10^10
L151/ 8,192,000 ***E1/ 1.342×10^11
L161/ 16,384,000 1/ 5.369×10^11
L171/ 32,768,000 ***C1/ 2.147×10^12
L181/ 65,536,000 ***W1/ 8.590×10^12
L191/ 131,07,2000 1/ 3.436×10^13
L201/ 262,144,000 1/ 1.37438953×10^14

Note: This is only for humans. It does not include elves, liches, demi-gods, transdimensional aliens, or whatever you decide to come up with.

You can play with the two main variables (base rate and multiplier—the latter is the most significant of the two—) to see what happens. For example, if you want a setting where magic users are very uncommon yet the possibility of powerful magicians is not ruled out, assume a very high base rate and a very low multiplier (e.g. 1 magic-user per 200,000 but a multiplier of 1.5.)

2019 edit: I actually checked this with a popular computer RPG, Path of Exile, which includes a permadeath league (the Hardcore League) and has public stats for any character above level 45. One league started on the 13th of this month, and I found that at the top levels (100-85) the average multiplier was 1.5 (there was one 98th-level dude but 205 85th-level players,) but I’m sure that at lower levels, the multiplier is close to 1.1 Otherwise, you’ll need almost as many players as stars in the Universe for someone to get to the top. But it’s doable with an average multiplier close to 1.1 that increases just a tinny bit at every level after a certain point.

To find the number of lower-level characters given a higher level, raise the multiplier to the power of (the original level – desired one) So, in the fast column, if you have a population of half a million with one 11th level character and you want to know how many 6th-level are there: 1*2^(11-6) = 32. An unintended consequence of what I have just said is that, if you want to have a Jesus in every town, there’s no escape: either going from level 1 to 15 is pretty much a given for everybody (so a multiplier almost identical to 1—or even lower!) or half the population of your world is made of first-level clerics.

Of course, you can do whatever you want in your campaign, and what I have said here is not a law, just a suggestion and a consequence of power progressions. It’s also for background context and back-of-the-envelope world-building; your PCs can be the exception to the rule, and you are under no obligation to make a “realistic” setting (or any setting at all.) But generally speaking, even with the most generous assumptions, high-level NPCs should be uncommon and characters should travel a lot (even to other worlds) to find the next big challenge (at least if fighting human villains.) These numbers I have shown here just explain why.

2019 edit: So, to answer the question of the effect of high-level clerics on your game. You probably won’t have to worry about that because it’s unlikely the players will ever come across one unless that cleric happens to be in their own group. And if he is, what the hell is he doing mucking around with a bunch of murderhobos instead of setting up the next world-conquering religion?

7 thoughts on “Level Inflation is a disease even clerics can’t cure.

  1. Mary

    There’s always disallowing spells that raise the dead. 0:)

    I’m playing around with a Gamelit setting where magic — both clerical and wizard — is common enough to start making serious inroads on being sufficiently advanced to be indistinguishable from technology, but — the dead stay dead.


  2. 1 some people only play for the levels – who gets to level 20 first with all the cool stuff. Very much a computer game like Neverwinter. It’s OK but I’m not into it.
    2 The GM sets the tone. Just don’t stuff a village with every variety of restockist, unless you are aiming for a type 1 game. Let your characters fail.
    3. Play RM in some format or Grim and Perilous where accidental death is a real risk despite being lvl 20
    4 quit the level rush and the XP scrabble for a non level system like Runequest or home brew it.

    All thoughts from my MERP blog. Like you Emp: I think it’s the story that matters.


  3. I’ve always assumed that the PCs are anomalies in the game world and that most of the population never goes above 1st level, with bosses like bandit chiefs and thief lords maybe as high as 3rd. That’s what makes monsters monstrous–a bog standard troll or a few ghouls could easily wipe out an entire village.

    So A PC who gets to 5th or 6th level is a legendary hero–think “The Magnificent Seven” in terms of the power chasm between PCs and NPCs. And in a game I run, reaching that level would represent years of play and either good strategy or a lot of luck.

    As PCs get higher level the monsters they face get more dangerous. As written, monsters with a save versus death attack (poison, petrification, and the like) have a good chance of pulling off a TPK even with a group of relatively high level adventurers. So survivability is actually less at higher levels, and I think the game is designed that way for a reason.

    I see 3rd to 6th level as the optimum window for PC adventurers–powerful enough to defeat foes that would annihilate commoners, but still risking death every time they delve.

    I’ll admit, I run low level games and prefer to play in them. In my Sunday night game the highest level PC, an 8th level archer, died last week from an area effect poison cloud–save or die, and I missed the saving throw, So I rolled up a first level character (3d6 in order of course) and started over.

    Liked by 1 person

  4. der Nicht Kluge Hans

    Speaking of video-game – like campaigns, this issue that you describe for D&D reminds me of a certain Elder Scrolls game.

    Because Oblivion was a game that levelled NPCs with the player, and because the player would cease to level his combat skills beyond a threshold, the player would continue to level to a point at which nearly every enemy would become a nearly unkillable terminator. This was why two of the most useful mods for Oblivion were a mod that prevented the player from levelling beyond level thirty and a mod that levelled quest rewards from earlier levels to match the player’s current level.


  5. Pingback: Blog Watch: Risk Averse Marines, Level Inflation, Demonic Easter Bunnies, and Man Plots | Jeffro's Space Gaming Blog

  6. Pingback: Sensor Sweep: The Night Land, Fritz Leiber, Scott Oden, Viking Prince – Herman Watts

  7. D&D has a couple of relevant problems with its structure, which go back all the way to Gygax-era TSR:

    1) Leveling is linear progression. The only way forward is to gain more levels, and even when dual-classing became a thing it was gated so that you had to have unusually high Ability Scores to even qualify. While I think that classes provide structure that classless systems often go horribly wrong without (detail bloat, for example), I think that after a while progression should transition into being more about broadening character capabilities rather than raising the ‘level’ ever higher. No hard cap to leveling, but diminishing returns and incentivizing branching out.

    1a) If ‘char-op’ can be a thing, there aren’t enough diminishing returns.

    2) By AD&D, there were 9 levels of spells. Only way to get up to that 9th level was to reach Level 18. There were no alternate means provided to cast higher level spells (perhaps via groups of magic-user minions employed via ritual, or artificial and external valences to hold spells for the boss magic user, etc.), so players had the expectation that they would be allowed to get their characters up to the ‘end-game content’.



Leave a Reply to der Nicht Kluge Hans Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.