CR HP AC
0 1-6 13
1/8 7-15 13
1/4 16-49 13
1/2 50-70 13
1 71-85 13
2 86-100 13
3 101-115 13
4 116-130 14
5 131-145 15
6 146-160 15
7 161-175 15
8 176-190 16
9 191-205 16
10 206-220 17
11 221-235 17
12 236-250 17
13 251-265 18
14 266-280 18
15 281-295 18
16 296-310 18
17 311-325 19
18 326-340 19
19 341-355 19
20 356-400 19

One of the classic ways to put a higher-difficulty monster into a module for lower-level characters is to have it be already injured before they fight it, giving it fewer HP. Meanwhile, GMs often figure that if they just set a monster higher in its HP range (by default, they take the average dice rolls) that’s still fair but might make the fight a little more challenging.

If you haven’t looked into the monster math for 5e (starting on page 273 of the DMG), your intuitions about altering HP might be wrong. If you’ve come from previous editions where the total number of HD was equivalent to the monster’s level, that is no longer the case. The hit dice for a monster in 5e are basically completely arbitrary. The designer was trying to hit a certain HP total, and adjusted the number of HD until the average value of those dice plus Con mod got into the right range.

This means, if you try to adjust the monster’s HP, your intuition might be wrong for how much that lowers or raises the difficulty. Even a monster with extremely low HP might obliterate a lower-level party if it wins initiative (especially if it also has a high AC), since damage scales so high. And maxing out a monster’s HP could functionally double it, which could push it multiple CRs higher in some cases.

The relevant table is on page 274 of the DMG, but the important parts are to the right. Note that most of the published monsters are close to this system, but not exact, as they were likely tweaked in playtesting. But this is the official guideline for how monsters are supposed to be assigned a CR, and it’s at least in the right ballpark most of the time.

Essentially, to figure out a monster’s defensive CR, look up its HP on the table to get the basic CR. Then consult the target AC for that line. If the monster’s AC is different, move the CR up or down by 1 for every 2 points of AC. For example, if the monster has 110 HP and 15 AC, that puts it at CR 3, which has target AC 13. Since it’s 2 AC higher, its defensive CR is effectively 4. Special abilities might also increase this: if it has a lot of resistances/immunities or special defensive actions, it might bump up another couple.

Some monsters are glass cannons (high offense, but low defensive CR) and others are tanks (the opposite), so have their offensive CR and defensive CR averaged. Basically, if you’re looking at a monster before adjusting its HP, and its defensive CR comes out higher or lower than its listed CR, it may have more or less offense to compensate. Just keep that in mind.

To adjust the HP, figure out what CR the new HP total would put it at, then step up or down if the monster’s AC is not in line. Figure out how much that differs from the original defensive CR, and adjust the monster’s CR by half the difference (because you’re averaging the offensive CR).

Examples

The Inert Golem

This is basically what got me thinking about the problem in the first place. The Planescape module Doors to the Unknown‘s first chapter features a battleground where variant iron golems have gotten to 0 HP but not been smashed up, which means that if they take any fire damage, this causes them to heal HP and reanimate, attacking the party. They’re likely to acquire less than 20 HP from the various reduced-damage fireball effects in the area, but how high of a CR is an iron golem with only a few HP?

The base iron golem in the MM has 210 HP, AC 20, and a CR 16. 210 HP is right around CR 10, which expects 17 AC, so the 20 AC bumps that up a couple of CRs to 12. We can probably also assume that the golem immunities and special defensive abilities are good for a couple more points of CR, so about 14.

What if it only had at most a dozen HP from a stray 2d6 mini-fireball? That puts it at CR 1/8, which targets AC 13. The extra 7 points of 20 AC are now four steps up the chart, landing at 2, and the same immunities and such put that two more steps up to CR 4. That’s a 10 point difference, halved to 5, so lowers the CR to 11.

Is 11 too high for something that would go down in a single attack? Maybe? If nobody manages to get through its 20 AC (or its magic resistance, or targets an immunity accidentally) before it acts, it’s going to likely put out about 45 damage when it gets to act. Probably still doesn’t make it better than a Stone Golem (CR 10). It could wreck a low-level party that can’t manage to hit (especially if they don’t have magic or adamantine weapons), but this is an example of why you don’t want the offensive and defensive CR to get too far apart. It would probably make more sense to have a scenario where there’s a full-on fireball barrage that would probably get it to at least another CR threshold or two.

The Two Wolves

Winter Wolves are HP 75, AC 13, CR 3 with no real special defensive abilities other than immunity to cold. That HP total puts them right at CR 1, and the AC is correct for that CR. Which means most of its CR is coming from offense, presumably.

Maybe you want a threat for a first-level party, so try to have an injured wolf. Dropping it to half HP puts it in the CR 1/4 range, which doesn’t alter the AC, so it’s two steps down. That, halved, reduces it to a CR 2. It’s still, offensively, much more powerful: that cold breath could wipe out multiple low-level PCs if they get too close, and its bite might take out one every round. If there are multiple, Pack Tactics can be a big threat as well.

Conversely, you want a really chonky boi that’s been eating well and is at maximum for his (arbitrary) 10d10+20 HD: 120 HP. That puts him up at CR 4, and now his AC is a little low, so maybe a CR 3. That’s two points higher than the base for an averaged +1, and makes him a final CR 4.