Monster math - building monsters like a nerd
By Harry ·
Contents
Hey all, this article is going to be a breakdown of the basics math behind monsters - this is extrapolated from about 580 official D&D 5.5 monsters.
There are gaps in the data as there are few to no monsters between CR25-29, but where possible I extrapolated the data with median regression analysis to fill in the blanks as best we can. Paul Hughes’ Blog of Holding was a huge help as he’s done some incredible work with breaking down the math and methodology behind building monsters (check it out here). I compared my results to his as a backstop for making sure things didn't go too wonky and it seems we’re playing in the right ballpark. His methodology was adapted for our calculations of Damage Per Round (DPR).
The hope with this article is to create a baseline guidance for anyone looking to build a monster at any CR, whether legendary or normal. If you want to skip the whole breakdown you can just have a look at the 2 tables below. We’ve split legendary and normal monsters as there seems to be some measurable difference between them when it comes to HP & DPR etc. To reiterate, this is a guideline, you can break it! But, it's a handy tool to check whether or not your monster is in the right ballpark when it comes to the individual building blocks. Building using our guidelines isn’t a guarantee of “encounter safety”, and doesn't account for the flavour you want to give a monster. So we’re all for going outside the bounds; push AC down, pump DPR up, mess with HP… this is all to say a good monster is what you want to make of it, we’re just offering a touch of support. I’ll take any feedback, I’m only just one guy doing this, so there’s likely an oversight here or there, but should be pretty good overall.
Quick check monster building tables
Table 1: Monster base stats
Normal monsters (non-legendary)
| CR | DPR¹ (±35% above CR ½) | HP² (±30% above CR ½) | AC³ (±2) |
|---|---|---|---|
| 0 | 1 | 3 | 12 |
| 1/8 | 4 | 8 | 12 |
| 1/4 | 6 | 13 | 12 |
| 1/2 | 8 | 20 | 12 |
| 1 | 11 | 30 | 13 |
| 2 | 17 | 45 | 13 |
| 3 | 23 | 60 | 14 |
| 4 | 29 | 75 | 14 |
| 5 | 35 | 90 | 15 |
| 6 | 41 | 105 | 15 |
| 7 | 47 | 120 | 15 |
| 8 | 53 | 135 | 16 |
| 9 | 59 | 150 | 16 |
| 10 | 65 | 165 | 17 |
| 11 | 71 | 180 | 17 |
| 12 | 77 | 195 | 17 |
| 13 | 83 | 210 | 18 |
| 14 | 89 | 225 | 18 |
| 15 | 95 | 240 | 19 |
| 16 | 101 | 255 | 19 |
| 17 | 107 | 270 | 19 |
| 18 | 113 | 285 | 20 |
| 19 | 119 | 300 | 20 |
| 20 | 125 | 315 | 21 |
| 21 | 131 | 330 | 21 |
| 22 | 137 | 345 | 21 |
| 23 | 143 | 360 | 22 |
| 24 | 149 | 375 | 22 |
| 25 | 155 | 390 | 23 |
| 26 | 161 | 405 | 23 |
| 27 | 167 | 420 | 23 |
| 28 | 173 | 435 | 24 |
| 29 | 179 | 450 | 24 |
| 30 | 185 | 465 | 25 |
¹ CR 0-½ (±2 DPR): data extrapolated from linear regression of median DPR for each CR. (CR 0-½ DPR uses the observed median).
² CR 0-½ (±4 HP) data extrapolated from linear regression of median HP for each CR.
³ Data extrapolated from linear regression of median AC for each CR. (CR 0-½ AC uses the observed median).
Legendary monsters
| CR | DPR⁴ (±33% above CR ½) | HP⁵ (±15% above CR ½) | AC⁶ (±2) |
|---|---|---|---|
| 0 | 1 | 3 | 12 |
| 1/8 | 4 | 8 | 12 |
| 1/4 | 6 | 13 | 12 |
| 1/2 | 8 | 20 | 12 |
| 1 | 9 | 65 | 13 |
| 2 | 17 | 75 | 13 |
| 3 | 26 | 85 | 14 |
| 4 | 34 | 95 | 14 |
| 5 | 43 | 105 | 15 |
| 6 | 51 | 115 | 15 |
| 7 | 60 | 125 | 15 |
| 8 | 68 | 135 | 16 |
| 9 | 77 | 145 | 16 |
| 10 | 85 | 155 | 17 |
| 11 | 94 | 165 | 17 |
| 12 | 102 | 175 | 17 |
| 13 | 111 | 185 | 18 |
| 14 | 119 | 195 | 18 |
| 15 | 128 | 205 | 19 |
| 16 | 136 | 215 | 19 |
| 17 | 145 | 225 | 19 |
| 18 | 153 | 235 | 20 |
| 19 | 162 | 278 | 20 |
| 20 | 170 | 320 | 21 |
| 21 | 179 | 363 | 21 |
| 22 | 187 | 405 | 21 |
| 23 | 196 | 448 | 22 |
| 24 | 204 | 490 | 22 |
| 25 | 213 | 533 | 23 |
| 26 | 221 | 575 | 23 |
| 27 | 230 | 618 | 23 |
| 28 | 238 | 660 | 24 |
| 29 | 247 | 703 | 24 |
| 30 | 255 | 745 | 25 |
⁴ CR 0-½ (±2 DPR): data extrapolated from linear regression of median DPR for each CR. (CR 0-½ DPR uses the observed median from non-legendary monsters).
⁵ CR 0-½ (±4 HP) data extrapolated from linear regression of median HP for each CR. CR0-½ uses observed median HP from non-legendary monsters. Obviously be super careful about HP for a legendary monster at Low CRs, as extra damage and HP could be lethal, and there are almost no official legendary monsters below CR3. The table values are more useful for higher CR monsters.
⁶ Data extrapolated from linear regression of median AC for each CR. (CR 0-½ AC uses the observed median).
Table 2: Details of interest relating to all monsters (legendary & non-legendary)
| CR | Probability of at least 1 Resistance⁷ | Probability of at least 1 immunity⁸ | Probability of at least 1 Saving Throw Prof. (ST)⁹ | Median ST Proficiency (±1)¹⁰ | Probability of Multiattack¹¹ | Median uses in multiattack¹² | Probability of a Trait¹³ | Traits Trend (±2 up to 8, ±3 for 9+)¹⁴ | Ability Point Budget total (±15 points)¹⁵ |
|---|---|---|---|---|---|---|---|---|---|
| 0 | 12.50% | 9.38% | 6.50% | 1 | 0.00% | 0 | 62.50% | 1 | 48 |
| 1/8 | 8.33% | 4.17% | 9.80% | 1 | 0.00% | 0 | 50.00% | 1 | 57 |
| 1/4 | 13.04% | 17.39% | 11.50% | 1 | 8.70% | 2 | 54.35% | 1 | 59 |
| 1/2 | 13.89% | 19.44% | 14.50% | 1 | 19.44% | 2 | 66.67% | 1 | 61 |
| 1 | 30.43% | 30.43% | 19.60% | 2 | 32.61% | 2 | 54.35% | 1 | 70 |
| 2 | 23.08% | 29.23% | 29.10% | 2 | 50.77% | 2 | 56.92% | 1 | 73 |
| 3 | 14.58% | 41.67% | 38.00% | 2 | 85.42% | 2 | 62.50% | 1 | 75 |
| 4 | 22.86% | 34.29% | 46.10% | 2 | 91.43% | 2 | 57.14% | 1 | 77 |
| 5 | 40.00% | 37.50% | 53.50% | 2 | 92.50% | 2 | 75.00% | 1 | 80 |
| 6 | 16.00% | 32.00% | 60.10% | 2 | 92.00% | 3 | 48.00% | 1 | 82 |
| 7 | 23.81% | 33.33% | 65.80% | 2 | 95.24% | 3 | 52.38% | 1 | 84 |
| 8 | 31.03% | 27.59% | 70.80% | 2 | 100.00% | 3 | 62.07% | 1 | 87 |
| 9 | 37.50% | 62.50% | 75.00% | 2 | 100.00% | 3 | 68.75% | 2 | 89 |
| 10 | 23.53% | 41.18% | 78.70% | 2 | 100.00% | 3 | 58.82% | 2 | 91 |
| 11 | 35.29% | 41.18% | 81.70% | 2 | 94.12% | 3 | 82.35% | 2 | 94 |
| 12 | 44.44% | 44.44% | 84.30% | 3 | 100.00% | 3 | 66.67% | 2 | 96 |
| 13 | 44.44% | 66.67% | 86.50% | 3 | 100.00% | 3 | 100.00% | 2 | 98 |
| 14 | 0.00% | 100.00% | 88.40% | 3 | 100.00% | 3 | 100.00% | 2 | 101 |
| 15 | 13.33% | 40.00% | 90.00% | 3 | 100.00% | 3 | 100.00% | 2 | 103 |
| 16 | 37.50% | 62.50% | 91.30% | 3 | 100.00% | 3 | 87.50% | 2 | 105 |
| 17 | 50.00% | 75.00% | 92.50% | 3 | 100.00% | 3 | 100.00% | 2 | 108 |
| 18 | 100.00% | 75.00% | 93.50% | 3 | 100.00% | 3 | 100.00% | 2 | 110 |
| 19 | 50.00% | 100.00% | 94.30% | 3 | 100.00% | 3* | 100.00% | 2 | 112 |
| 20 | 50.00% | 75.00% | 95.00% | 3 | 100.00% | 3 | 100.00% | 2 | 115 |
| 21 | 50.00% | 83.33% | 95.60% | 3 | 100.00% | 3 | 100.00% | 2 | 117 |
| 22 | 33.33% | 83.33% | 96.20% | 3 | 100.00% | 3 | 100.00% | 2 | 119 |
| 23 | 40.00% | 100.00% | 96.60% | 3 | 100.00% | 3 | 100.00% | 2 | 122 |
| 24 | 0.00% | 100.00% | 97.00% | 3 | 100.00% | 3 | 100.00% | 2 | 124 |
| 25 | 100.00% | 100.00% | 97.40% | 3 | 100.00% | 3 | 100.00% | 2 | 126 |
| 26 | 100%* (assumed) | 100%* (assumed) | 97.70% | 3 | 100%* (assumed) | 3 (assumed) | 100%* (assumed) | 2 (assumed) | 129 |
| 27 | 100%* (assumed) | 100%* (assumed) | 97.90% | 3 | 100%* (assumed) | 3 (assumed) | 100%* (assumed) | 2 (assumed) | 131 |
| 28 | 100%* (assumed) | 100%* (assumed) | 98.20% | 3 | 100%* (assumed) | 3 (assumed) | 100%* (assumed) | 2 (assumed) | 133 |
| 29 | 100%* (assumed) | 100%* (assumed) | 98.40% | 3 | 100%* (assumed) | 3 (assumed) | 100%* (assumed) | 2 (assumed) | 136 |
| 30 | 100.00% | 100.00% | 98.50% | 4 | 100.00% | 4 | 100% | 4 | 138 |
⁷ Based on raw % of monsters with at least 1 resistance for each CR.
⁸ Based on raw % of monsters with at least 1 immunity for each CR.
⁹ Based on % of observed monsters with 1 proficiency per CR & quadratic trendline fit applied.
¹⁰ Based on median Save prof. In each CR: CR bands created as no statistically significant trendline fit could be produced for each CR vs bands.
¹¹ Based on % monsters in each CR w/ MA.
¹² Median uses within a multiattack. CR19* Median is 2, given uses around it, 3 is a plausible guideline.
¹³ Raw % of monsters with at least 1 trait.
¹⁴ Based on grouped medians CR0-9, CR10+, variation (±) captures 98.9% of monsters but some have more.
¹⁵ Based on linear regression of the median ability points per CR, use with caution - see AS section below.
For full details of the above tables below is a quick runthrough of everything:
DPR: Damage Per Round
When building a monster that isn't over or underpowered it's important to understand how much damage that monster can put out per round. We used some similar assumptions to Paul Hughes when calculating this:
- Combat takes place over 3 rounds
- Monster always choose their highest damage attack
- All attacks hit
- All foes fail their saves
- All AOE attacks affect 2 foes
- Limited use: Recharge & 1/day spells etc. are only used once
- Legendary actions should be accounted for
Knowing how much DPR a monster puts out helps to balance the monster “kit” when it comes to damage, what traits, actions, Legendary actions, special weapons (Breath weapon), spells they have access to and their uses.
Methodology
So for DPR, I went through every monster, got the average damage of each of their damaging abilities, worked out the max Average damage they could do over 3 rounds (based on assumptions above) and then divided by 3.
There seems to be quite a difference between Legendary and normal monsters (as well as non-legendary dragons, usually hit harder) So I’ve split the analysis between normal and legendary monsters.
With every monster's average damage mapped, I took the median DPR for each CR as well as the upper and lower bound (25th and 75th percentile) to see the spread for every CR.
From the Median scores we calculated linear regression. And these were used as a basis to calculate a rounded formula for DPR for both and set standard levels across CR ranges. To note, for DPR for CRs below 1 we used observed median vs regression results as it built overpowered Low CR monsters - the same numbers were then applied to Legendary monsters at those CR.
DPR formulas
Simplified Formula (Normal):
DPR = 6 × CR + 5 [From CR1]
Simplified Formula (Legendary):
DPR = 8.5 × CR
Now for the variation:
Non-legendary monsters: below CR 1, ±2 DPR seems to capture most of the range of monsters, above that ±35% captures about 80% of monsters in the range.
Legendary Monsters: ±33% seems to capture about 86% of monsters analysed. There is no comparative range from below CR1 because the lowest level legendary monster is CR3 in the dataset and the regression analysis under-estimates DPR at lower CRs due to the regression line being steeper than for non-legendary monsters.
I’m probably going to say this a million times, but again, I want to reiterate that even with 580+ monsters there are very few monsters above CR19, so the data I have is sparse - this is doubly so for the limited amount of legendary monsters which usually sit at higher CRs. So even with these trendlines created it's always worth following your gut and own thoughts about monster creation when building one. On the other hand, remember that there are very very few Legendary monsters at lower CRs, so be very thoughtful about how you might create one, as the balancing of an encounter at low levels is much more important for player survivability. Especially at very low CRs 0-2 (though I've never encountered a legendary monster that low, but could be a fun idea), the data is less solid here as it is extrapolated downwards from much higher CR legendary monsters we have data for. I have not playtested this yet in combat at the table, so it's worth just keeping your thinking cap on.
Also to note that HP doesn’t correlate to DPR very strongly across all monsters, in fact it's quite weak. So it's not necessarily true that High HP = Low Damage or vice versa. With normal monsters that correlation is very weak, with Legendary monsters there was close to none. But probably best to weigh in with your own thoughts when building a monster, and maybe lean to the safe side. If a monster has high offences capabilities, probably best to balance that with defensive capabilities that aren't too intensive.
HP
Now to HP. I took the HP of every monster at every CR, found the median and looked at the bounds, similar to above.
Normal (non-legendary) monsters
HP formula, normal monsters:
HP = 15 × CR + 15 [From CR 1]
For normal monsters this formula captures the median HP pretty well across all non-legendary monsters. To capture the upper and lower bounds of HP at any CR I found ±30% seemed to capture about 85% of all monsters - so you can adjust accordingly, being conservative is safer unless you really know what you want to build. Below CR1 we used observed values for Median HP.
Legendary monsters
HP formulas, legendary:
HP = 10 × CR + 55 [From CR 1-18]
HP = 42.5 × CR − 530 [From CR19-30]
To fit the raw median data as closely as possible we have split the analysis into two formulas. From CR 19 onwards the HP increases much faster (about 43 HP per CR vs 10HP per CR). This two line regression had a better fit.
To account for variation in health:
Legendary Monsters: A variation of about a ±15% HP seems to capture about 86% of the dataset. If the range expands to ±20% above CR18 that goes up to 89%.
Normal Monsters: a variation of ±4 HP for CR 0-1/2, and 30% covers about 85% of monsters.
For Legendary monsters we have taken the observed HP medians of non-legendary/normal monsters and applied them to CRs below 1. It's also important to note that above CR19 there are very few monsters, so please take the CR19+ results with a pinch of salt and build to your liking.
This two line regression also means that the median HP values for legendary monsters from CR10-19 fall slightly below those of Normal monsters. The difference is quite small, and both have ± ranges which allow to change the distribution of those stats. And then there is the simple fact that though HP is lower in that range for legendary monsters, the median DPR is higher to reflect their legendary status.
AC
AC appears to not show much difference between legendary and Normal monsters, so AC can be calculated using one simple linear regression done on all the median values from 1-30, using observed Median AC below 1.
AC formula, all monsters:
AC = 0.4 × CR + 12.6 [From CR 1-30]
To account for variation in AC we have given a ±2 AC to show you the variation in AC that is well within conservative bounds to guide you. AC ±2 covers 80.8% of normal and 91.2% of legendary monsters. But again, flavour your monster how you like, there are some monsters who pump out a lot of damage and you may want to balance that out with a corresponding decrease in HP or AC or something of the like that fits the monster.
Other considerations
Other than the base raw stats of AC, HP and DPR, there are a few other things that can be inferred from the monsters that can help with building a monster.
There are:
- Damage Resistances
- Damage immunities
- Actions or traits that impose a condition (on self or target eg. Advantage, Prone, poisoned etc.)
- Saving Throws
- Multiattack
- Number of Traits (non-damage)
- Ability Scores
- Spellcasters
Damage resistances
For resistances the analysis looks at what % of monsters at a given CR have a resistance, and if so, what is the median as well as the variance. There’s no clean way to apply this unfortunately as there are quite a few outlying monsters with very high resistance counts at all CRs. But the chance of resistance is a good indicator for whether to apply it to your own monsters.
| CR | Monsters | With resistance | Prevalence | Median types if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 4 | 12.50% | 1 | 1 | 1 |
| 1/8 | 24 | 2 | 8.33% | 4 | 3 | 5 |
| 1/4 | 46 | 6 | 13.04% | 3 | 1 | 3 |
| 1/2 | 36 | 5 | 13.89% | 3 | 1 | 5 |
| 1 | 46 | 14 | 30.43% | 3 | 1 | 7 |
| 2 | 65 | 15 | 23.08% | 1 | 1 | 8 |
| 3 | 48 | 7 | 14.58% | 1 | 1 | 4 |
| 4 | 35 | 8 | 22.86% | 5 | 2 | 8 |
| 5 | 40 | 16 | 40.00% | 2 | 1 | 6 |
| 6 | 25 | 4 | 16.00% | 3 | 1 | 3 |
| 7 | 21 | 5 | 23.81% | 1 | 1 | 5 |
| 8 | 29 | 9 | 31.03% | 2 | 1 | 5 |
| 9 | 16 | 6 | 37.50% | 3 | 1 | 5 |
| 10 | 17 | 4 | 23.53% | 2.5 | 1 | 5 |
| 11 | 17 | 6 | 35.29% | 1.5 | 1 | 3 |
| 12 | 9 | 4 | 44.44% | 3 | 1 | 3 |
| 13 | 9 | 4 | 44.44% | 2 | 1 | 3 |
| 14 | 5 | 0 | 0.00% | — | — | — |
| 15 | 15 | 2 | 13.33% | 1 | 1 | 1 |
| 16 | 8 | 3 | 37.50% | 2 | 1 | 3 |
| 17 | 8 | 4 | 50.00% | 2.5 | 1 | 4 |
| 18 | 4 | 4 | 100.00% | 2 | 2 | 3 |
| 19 | 2 | 1 | 50.00% | 2 | 2 | 2 |
| 20 | 4 | 2 | 50.00% | 3 | 1 | 5 |
| 21 | 6 | 3 | 50.00% | 2 | 2 | 3 |
| 22 | 6 | 2 | 33.33% | 2 | 1 | 3 |
| 23 | 5 | 2 | 40.00% | 3 | 3 | 3 |
| 24 | 2 | 0 | 0.00% | — | — | — |
| 25 | 1 | 1 | 100.00% | 2 | 2 | 2 |
| 26 | 0 | 0 | — | — | — | — |
| 27 | 0 | 0 | — | — | — | — |
| 28 | 0 | 0 | — | — | — | — |
| 29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100.00% | 3 | 3 | 3 |
Damage immunities
Much like the above table the same applies to Immunities. Again, it's a good rule of thumb to use this when thinking of applying any immunities to your monsters.
| CR | Monsters | With immunity | Prevalence | Median types if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 3 | 9.38% | 2 | 1 | 2 |
| 1/8 | 24 | 1 | 4.17% | 1 | 1 | 1 |
| 1/4 | 46 | 8 | 17.39% | 1.5 | 1 | 2 |
| 1/2 | 36 | 7 | 19.44% | 1 | 1 | 2 |
| 1 | 46 | 14 | 30.43% | 1 | 1 | 2 |
| 2 | 65 | 19 | 29.23% | 1 | 1 | 2 |
| 3 | 48 | 20 | 41.67% | 1 | 1 | 2 |
| 4 | 35 | 12 | 34.29% | 2 | 1 | 4 |
| 5 | 40 | 15 | 37.50% | 1 | 1 | 2 |
| 6 | 25 | 8 | 32.00% | 1 | 1 | 2 |
| 7 | 21 | 7 | 33.33% | 1 | 1 | 2 |
| 8 | 29 | 8 | 27.59% | 1 | 1 | 2 |
| 9 | 16 | 10 | 62.50% | 1 | 1 | 3 |
| 10 | 17 | 7 | 41.18% | 1 | 1 | 2 |
| 11 | 17 | 7 | 41.18% | 2 | 1 | 2 |
| 12 | 9 | 4 | 44.44% | 2 | 1 | 2 |
| 13 | 9 | 6 | 66.67% | 1 | 1 | 2 |
| 14 | 5 | 5 | 100.00% | 1 | 1 | 3 |
| 15 | 15 | 6 | 40.00% | 1 | 1 | 2 |
| 16 | 8 | 5 | 62.50% | 1 | 1 | 3 |
| 17 | 8 | 6 | 75.00% | 1 | 1 | 2 |
| 18 | 4 | 3 | 75.00% | 2 | 2 | 3 |
| 19 | 2 | 2 | 100.00% | 1.5 | 1 | 2 |
| 20 | 4 | 3 | 75.00% | 1 | 1 | 2 |
| 21 | 6 | 5 | 83.33% | 2 | 1 | 2 |
| 22 | 6 | 5 | 83.33% | 2 | 1 | 6 |
| 23 | 5 | 5 | 100.00% | 2 | 1 | 3 |
| 24 | 2 | 2 | 100.00% | 1 | 1 | 1 |
| 25 | 1 | 1 | 100.00% | 2 | 2 | 2 |
| 26–29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100.00% | 2 | 2 | 2 |
Actions or traits that impose a condition on self or target
This data looks at the monsters and tries to understand actions or traits that have additional utility than just damage eg. Advantage on self, knocking someone prone with the attack etc. This is for us to get a grasp of extending a monster beyond just an attack.
What I looked at:
- Applies a condition to a target: Prone, Grappled, Poisoned, Restrained or Frightened.
- Grants Advantage to the monster; including attack rolls, saving throws and ability checks.
- Imposes Disadvantage on a target's rolls, or on attacks directed against the monster.
- Gives the monster a beneficial condition, such as Invisible.
I’ve bundled them all together here just for a general sense of what you have access to when building your own monster.
| CR | Monsters | With feature | Prevalence | Median types if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 6 | 18.75% | 1 | 1 | 1 |
| 1/8 | 24 | 9 | 37.50% | 1 | 1 | 1 |
| 1/4 | 46 | 19 | 41.30% | 1 | 1 | 2 |
| 1/2 | 36 | 17 | 47.22% | 1 | 1 | 2 |
| 1 | 46 | 31 | 67.39% | 1 | 1 | 3 |
| 2 | 65 | 35 | 53.85% | 1 | 1 | 3 |
| 3 | 48 | 30 | 62.50% | 1 | 1 | 2 |
| 4 | 35 | 21 | 60.00% | 1 | 1 | 2 |
| 5 | 40 | 31 | 77.50% | 1 | 1 | 3 |
| 6 | 25 | 18 | 72.00% | 1 | 1 | 3 |
| 7 | 21 | 16 | 76.19% | 1 | 1 | 3 |
| 8 | 29 | 17 | 58.62% | 1 | 1 | 3 |
| 9 | 16 | 13 | 81.25% | 1 | 1 | 2 |
| 10 | 17 | 15 | 88.24% | 1 | 1 | 3 |
| 11 | 17 | 14 | 82.35% | 1 | 1 | 3 |
| 12 | 9 | 8 | 88.89% | 1 | 1 | 2 |
| 13 | 9 | 8 | 88.89% | 1 | 1 | 7 |
| 14 | 5 | 2 | 40.00% | 1 | 1 | 1 |
| 15 | 15 | 13 | 86.67% | 1 | 1 | 2 |
| 16 | 8 | 7 | 87.50% | 1 | 1 | 2 |
| 17 | 8 | 7 | 87.50% | 1 | 1 | 2 |
| 18 | 4 | 4 | 100.00% | 1.5 | 1 | 2 |
| 19 | 2 | 2 | 100.00% | 1.5 | 1 | 2 |
| 20 | 4 | 3 | 75.00% | 3 | 1 | 4 |
| 21 | 6 | 4 | 66.67% | 1.5 | 1 | 3 |
| 22 | 6 | 4 | 66.67% | 1 | 1 | 2 |
| 23 | 5 | 4 | 80.00% | 2.5 | 1 | 3 |
| 24 | 2 | 1 | 50.00% | 1 | 1 | 1 |
| 25 | 1 | 1 | 100.00% | 2 | 2 | 2 |
| 26–29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100.00% | 4 | 4 | 4 |
Saving throw proficiencies
Here we look at what % of monsters at a given CR have a saving throw proficiency, and of those who do, what is the median score with bounds attached.
What interesting here is that there seems to be some form of trend through the probabilities (Prevalence) of a monster at a given CR having a saving throw - but given WOtC would have to release monsters at a statistically appropriate rate to keep it in line, I’m not sure if it makes any sense. But, it does make life easier when figuring out if your monster might have a saving throw so ill put it here and throw it in the monster builder.
| CR | Monsters | With feature | Prevalence | Median if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 2 | 6.25% | 1.5 | 1 | 2 |
| 1/8 | 24 | 6 | 25.00% | 1 | 1 | 1 |
| 1/4 | 46 | 4 | 8.70% | 1 | 1 | 1 |
| 1/2 | 36 | 5 | 13.89% | 1 | 1 | 2 |
| 1 | 46 | 7 | 15.22% | 2 | 1 | 3 |
| 2 | 65 | 22 | 33.85% | 2 | 1 | 4 |
| 3 | 48 | 16 | 33.33% | 2 | 1 | 4 |
| 4 | 35 | 17 | 48.57% | 2 | 1 | 3 |
| 5 | 40 | 14 | 35.00% | 2 | 1 | 4 |
| 6 | 25 | 18 | 72.00% | 2 | 1 | 4 |
| 7 | 21 | 14 | 66.67% | 2 | 2 | 4 |
| 8 | 29 | 24 | 82.76% | 2 | 1 | 4 |
| 9 | 16 | 9 | 56.25% | 3 | 2 | 4 |
| 10 | 17 | 14 | 82.35% | 2.5 | 1 | 5 |
| 11 | 17 | 14 | 82.35% | 2 | 1 | 5 |
| 12 | 9 | 9 | 100.00% | 3 | 2 | 4 |
| 13 | 9 | 7 | 77.78% | 3 | 2 | 4 |
| 14 | 5 | 4 | 80.00% | 2 | 1 | 4 |
| 15 | 15 | 15 | 100.00% | 3 | 2 | 5 |
| 16 | 8 | 6 | 75.00% | 3 | 1 | 4 |
| 17 | 8 | 8 | 100.00% | 2 | 1 | 4 |
| 18 | 4 | 4 | 100.00% | 4 | 2 | 4 |
| 19 | 2 | 2 | 100.00% | 3 | 2 | 4 |
| 20 | 4 | 4 | 100.00% | 2 | 2 | 4 |
| 21 | 6 | 5 | 83.33% | 2 | 2 | 4 |
| 22 | 6 | 6 | 100.00% | 3 | 2 | 4 |
| 23 | 5 | 5 | 100.00% | 2 | 2 | 3 |
| 24 | 2 | 2 | 100.00% | 2 | 2 | 2 |
| 25 | 1 | 1 | 100.00% | 2 | 2 | 2 |
| 26 | 0 | 0 | — | — | — | — |
| 27 | 0 | 0 | — | — | — | — |
| 28 | 0 | 0 | — | — | — | — |
| 29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100.00% | 4 | 4 | 4 |
Multiattack
Here again we look at % of monsters at a given CR with multiattack and then the median amount of actions they have access to through it. Essentially to know if a monster in a CR likely has a multiattack and then how many actions they have access to. This ties in with DPR quite closely, as I found in the analysis of DPR that the use of multiattacks (always choosing the highest damage option of the action you have access to with it) on a majority of instances came very close to the DPR of the creature (and then doubled for legendary monsters - as their Legendary actions [3 uses mostly] mirrored the multiattack damage.
| CR | Monsters | With Multiattack | Prevalence | Median amount if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 0 | 0.00% | 0 | 0 | 0 |
| 1/8 | 24 | 0 | 0.00% | 0 | 0 | 0 |
| 1/4 | 46 | 4 | 8.70% | 2 | 2 | 2 |
| 1/2 | 36 | 7 | 19.44% | 2 | 2 | 3 |
| 1 | 46 | 15 | 32.61% | 2 | 2 | 4 |
| 2 | 65 | 33 | 50.77% | 2 | 2 | 5 |
| 3 | 48 | 41 | 85.42% | 2 | 2 | 3 |
| 4 | 35 | 32 | 91.43% | 2 | 2 | 4 |
| 5 | 40 | 37 | 92.50% | 2 | 2 | 5 |
| 6 | 25 | 23 | 92.00% | 3 | 2 | 3 |
| 7 | 21 | 20 | 95.24% | 3 | 2 | 4 |
| 8 | 29 | 29 | 100.00% | 3 | 2 | 5 |
| 9 | 16 | 16 | 100.00% | 3 | 2 | 4 |
| 10 | 17 | 17 | 100.00% | 3 | 2 | 3 |
| 11 | 17 | 16 | 94.12% | 3 | 2 | 3 |
| 12 | 9 | 9 | 100.00% | 3 | 2 | 4 |
| 13 | 9 | 9 | 100.00% | 3 | 2 | 3 |
| 14 | 5 | 5 | 100.00% | 3 | 2 | 3 |
| 15 | 15 | 15 | 100.00% | 3 | 2 | 4 |
| 16 | 8 | 8 | 100.00% | 3 | 2 | 7 |
| 17 | 8 | 8 | 100.00% | 3 | 3 | 3 |
| 18 | 4 | 4 | 100.00% | 3 | 3 | 3 |
| 19 | 2 | 2 | 100.00% | 2 | 2 | 2 |
| 20 | 4 | 4 | 100.00% | 3 | 3 | 4 |
| 21 | 6 | 6 | 100.00% | 3 | 2 | 3 |
| 22 | 6 | 6 | 100.00% | 3 | 2 | 3 |
| 23 | 5 | 5 | 100.00% | 3 | 2 | 3 |
| 24 | 2 | 2 | 100.00% | 3 | 3 | 3 |
| 25 | 1 | 1 | 100.00% | 3 | 3 | 3 |
| 26 | 0 | 0 | — | — | — | — |
| 27 | 0 | 0 | — | — | — | — |
| 28 | 0 | 0 | — | — | — | — |
| 29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100% | 4 | 4 | 4 |
Number of traits
This essentially breaks down the prevalence of traits at every CR. For the most part traits don’t cause any damage - but keep traits in mind when building a monster if it has some sort of aura or AOE damage that applies automatically as it will need to be part of your DPR.
| CR | Monsters | With trait | Prevalence | Median amount if present | Minimum | Maximum |
|---|---|---|---|---|---|---|
| 0 | 32 | 20 | 62.50% | 1 | 1 | 2 |
| 1/8 | 24 | 12 | 50.00% | 1 | 1 | 3 |
| 1/4 | 46 | 25 | 54.35% | 1 | 1 | 2 |
| 1/2 | 36 | 24 | 66.67% | 1 | 1 | 3 |
| 1 | 46 | 25 | 54.35% | 1 | 1 | 2 |
| 2 | 65 | 37 | 56.92% | 1 | 1 | 3 |
| 3 | 48 | 30 | 62.50% | 1 | 1 | 3 |
| 4 | 35 | 20 | 57.14% | 1.5 | 1 | 3 |
| 5 | 40 | 30 | 75.00% | 2 | 1 | 5 |
| 6 | 25 | 12 | 48.00% | 1.5 | 1 | 3 |
| 7 | 21 | 11 | 52.38% | 1 | 1 | 3 |
| 8 | 29 | 18 | 62.07% | 1 | 1 | 3 |
| 9 | 16 | 11 | 68.75% | 2 | 1 | 4 |
| 10 | 17 | 10 | 58.82% | 2 | 1 | 5 |
| 11 | 17 | 14 | 82.35% | 2 | 1 | 5 |
| 12 | 9 | 6 | 66.67% | 2.5 | 1 | 4 |
| 13 | 9 | 9 | 100.00% | 2 | 1 | 4 |
| 14 | 5 | 5 | 100.00% | 1 | 1 | 2 |
| 15 | 15 | 15 | 100.00% | 1 | 1 | 7 |
| 16 | 8 | 7 | 87.50% | 3 | 1 | 3 |
| 17 | 8 | 8 | 100.00% | 2.5 | 1 | 4 |
| 18 | 4 | 4 | 100.00% | 2 | 1 | 4 |
| 19 | 2 | 2 | 100.00% | 3.5 | 3 | 4 |
| 20 | 4 | 4 | 100.00% | 3 | 1 | 7 |
| 21 | 6 | 6 | 100.00% | 2 | 1 | 4 |
| 22 | 6 | 6 | 100.00% | 2.5 | 1 | 4 |
| 23 | 5 | 5 | 100.00% | 2 | 1 | 3 |
| 24 | 2 | 2 | 100.00% | 1.5 | 1 | 2 |
| 25 | 1 | 1 | 100.00% | 4 | 4 | 4 |
| 26 | 0 | 0 | — | — | — | — |
| 27 | 0 | 0 | — | — | — | — |
| 28 | 0 | 0 | — | — | — | — |
| 29 | 0 | 0 | — | — | — | — |
| 30 | 1 | 1 | 100% | 4 | 4 | 4 |
There seems to be some statistical significance between your CR and the number of traits that the monster has available to them, so as a general rule of thumb I will add this to the monster builder for simplicity.
Ability scores
I looked at the distribution of ability scores across all CRs, to find a basic idea of the median number of ability points above (or below) having the standard array of 10 in each ability, for each CR. Making it a bit easier to distribute ability points when creating a monster. We’ve built this into the Monster creator. Much like “buying” points when you make a character the system works similarly here. It should at least keep you in the right ballpark for a monster of that CR.
For the purposes of the example below that means every monster starts with an array of 10 points in STR, DEX, CON, CHA, WIS, INT - (we do this because the monster builder comes standard with this array) then its about buying (or selling) points to get to an array that is the median for that CR. So the chart demonstrates the ability points above or below the standard array of 60 to be at the median for that CR.
The interesting thing to note is our 2 very high CR monsters have quite “low” scores, mainly due to them having incredibly low INT, WIS, CHA, DEX: due to their size and nature. So it's incredibly important to note that it's not necessarily true that an incredibly high CR monster has to have a monstrous amount of ability points. But applying some linear regression shows a trajectory that's far above what the most powerful 2 monsters are.
So from 0-½ ability scores increase very quickly at about an average of 24 points, then much more slowly, around 2.5 points per CR thereafter. So we’ll use the observed medians for low CR, then a linear regression formula CR1 and above.
Ability scores (AS) formula, all monsters:
AS = 2.4 × CR + 7.85 [From CR 1 up]
For capturing variability ±15 ability points, CR 1 and above, seems to capture a majority of monsters in a way that doesn't seem to overstep the bounds too much. Though as seen from the ACTUAL monsters - the CR 25 and CR 30 monsters would be 36 points and 45 points lower than what our regression shows! Which is huge. So these should only be treated as ballparks.
Spellcasters
I thought I’d do a little bit of analysis on spellcasters to see if we could draw out any lessons or trends. This part is much more for fun than deeper analysis, but could be good fuel for thought when approaching spellcasting.
At the bare minimum I wanted to know:
- At a given CR how many spells a spell caster has (median) then the absolute min and the max.
- Of those spells what % is damage vs non-damage.
- What is the median spell level for each CR.
- What is the median highest level spell for that CR.
But in the process of looking at spellcaster I found that of the monsters that do have spellcasting 41% have spell kits that don't do any damage (as in 0) - these are utility spell kits. Then when you look at the 59% of spellcasters with damage in their kit: only 25% (on average) of their spells are spells that do damage. These numbers are obviously aggregated overall, but it seems like an interesting divide.
So for the sake of this analysis i’ll divide up the non-damage spellcasters and damage spellcasters.
For a quick overview of the types of spellcasters that fall into these two groups (when looking at SRD)
Non damage: Priest Acolyte, Cultist Fanatic, Unicorn, Oni, Deva, Cloud Giant, Rakshasa, Planetar, Solar Djinni
Damage: All Dragons, Druid, Priest, Mage, Hags, Lamia, Incubus, Couatl, Lich, Mummy Lord
So for tables, I used median values for all spellcasters - first lets look at damage casters:
Damage spellcasters (spellcasters with at least 1 spell that does damage)
Amount of spells & spell levels they have access to
| CR | Median amount of spells | Min spells | Max spells | Kit damage % | Median spell level | Median highest spell level | Absolute highest spell level |
|---|---|---|---|---|---|---|---|
| 1/4 | — | — | — | — | — | — | — |
| 1 | 2 | 2 | 2 | 50% | 1 | 1 | 1 |
| 2 | 6 | 3 | 7 | 26.90% | 1.5 | 2 | 3 |
| 3 | 5 | 3 | 5 | 23.10% | 2 | 2 | 3 |
| 4 | 5 | 3 | 9 | 26% | 2.5 | 3.5 | 9 |
| 5 | 4 | 3 | 5 | 33.30% | 3 | 6 | 7 |
| 6 | 4 | 3 | 9 | 22.20% | 3 | 5 | 7 |
| 7 | 3 | 2 | 5 | 30% | 3.5 | 5 | 6 |
| 8 | 5.5 | 3 | 8 | 25% | 3 | 5 | 6 |
| 9 | 9 | 9 | 9 | 11.10% | 3 | 7 | 7 |
| 10 | 7 | 4 | 10 | 25% | 5 | 7.5 | 8 |
| 11 | 6 | 2 | 10 | 22.20% | 3 | 6.5 | 7 |
| 12 | 8 | 6 | 14 | 14.30% | 5 | 8 | 9 |
| 13 | 7 | 6 | 15 | 21.40% | 3.5 | 8 | 9 |
| 14 | 5.5 | 5 | 6 | 27.30% | 3.5 | 6.5 | 9 |
| 15 | 7 | 3 | 13 | 21.70% | 3.5 | 8 | 9 |
| 16 | 6 | 6 | 7 | 26.30% | 3.5 | 6 | 9 |
| 17 | 4 | 4 | 10 | 33.30% | 3 | 8 | 9 |
| 18 | 10.5 | 10 | 11 | 14.30% | 3.75 | 8 | 9 |
| 19 | 9 | 9 | 9 | 44.40% | 5 | 8 | 9 |
| 20 | 7 | 7 | 7 | 14.30% | 3.75 | 8.5 | 9 |
| 21 | 10.5 | 6 | 15 | 35.70% | 4.5 | 9 | 9 |
| 22 | 9 | 4 | 11 | 33.30% | 5 | 9 | 9 |
| 23 | 7 | 6 | 8 | 21.40% | 4.5 | 7 | 9 |
| 24 | 5.5 | 5 | 6 | 36.40% | 4 | 7.5 | 9 |
Spell uses
| CR | Casters | Cantrips | At-will (all) | At-will leveled | 1/day | 2/day |
|---|---|---|---|---|---|---|
| 1/4 | 0 | — | — | — | — | — |
| 1 | 1 | 1 | 1 | 0 | 1 | 0 |
| 2 | 5 | 2 | 2 | 0 | 3 | 0 |
| 3 | 3 | 2 | 2 | 1 | 0 | 0 |
| 4 | 10 | 1 | 3 | 1 | 2 | 0 |
| 5 | 3 | 0 | 0 | 0 | 2 | 2 |
| 6 | 5 | 2 | 2 | 0 | 2 | 0 |
| 7 | 3 | 0 | 0 | 0 | 1 | 0 |
| 8 | 6 | 1 | 3 | 1.5 | 2 | 0.5 |
| 9 | 1 | 1 | 5 | 4 | 4 | 0 |
| 10 | 4 | 1 | 2.5 | 1 | 5 | 0 |
| 11 | 6 | 1 | 2 | 1 | 3.5 | 0 |
| 12 | 3 | 2 | 3 | 3 | 4 | 2 |
| 13 | 3 | 1 | 5 | 4 | 3 | 0 |
| 14 | 2 | 0.5 | 3.5 | 3 | 2 | 0 |
| 15 | 11 | 2 | 5 | 3 | 2 | 0 |
| 16 | 3 | 0 | 4 | 3 | 2 | 0 |
| 17 | 5 | 0 | 3 | 3 | 2 | 0 |
| 18 | 2 | 2 | 6 | 4 | 2 | 2.5 |
| 19 | 1 | 1 | 4 | 3 | 2 | 3 |
| 20 | 2 | 0.5 | 4 | 3.5 | 2 | 1 |
| 21 | 4 | 1.5 | 6 | 4.5 | 3.5 | 1 |
| 22 | 5 | 1 | 5 | 4 | 2 | 0 |
| 23 | 2 | 0.5 | 4 | 3.5 | 3 | 0 |
| 24 | 2 | 0 | 3 | 3 | 2.5 | 0 |
The big takeaway apart from the obvious increase in amount and level with CR, is that even for monsters who do damage with their spells, on average about 75% of their spell kit is non-damage.
Non-damage spellcasters (spellcasters with no spells that do damage)
Amount of spells & spell levels they have access to
| CR | Median amount of spells | Min spells | Max spells | Median spell level | Median highest spell level | Absolute highest spell level |
|---|---|---|---|---|---|---|
| 1/4 | 2 | 2 | 6 | 1.5 | 2 | 3 |
| 1 | 2.5 | 2 | 5 | 1.5 | 2 | 4 |
| 2 | 3 | 2 | 5 | 2 | 2 | 4 |
| 3 | 2 | 2 | 4 | 1.5 | 2 | 3 |
| 4 | 2 | 2 | 2 | 2 | 3 | 3 |
| 5 | 5.5 | 4 | 7 | 2.25 | 4.5 | 8 |
| 6 | 2.5 | 2 | 3 | 3 | 3 | 3 |
| 7 | 3 | 2 | 6 | 1.75 | 2.5 | 8 |
| 8 | 2.5 | 2 | 5 | 2.5 | 2.5 | 7 |
| 9 | 4.5 | 3 | 6 | 2.5 | 6.5 | 8 |
| 10 | 4 | 4 | 5 | 4 | 6 | 9 |
| 11 | 10 | 10 | 11 | 3 | 7 | 7 |
| 12 | 4 | 3 | 9 | 4 | 7 | 8 |
| 13 | 6 | 3 | 9 | 3.25 | 7.5 | 8 |
| 14 | — | — | — | — | — | — |
| 15 | 9 | 9 | 9 | 2 | 7 | 7 |
| 16 | 4.5 | 4 | 5 | 4 | 7.5 | 8 |
| 17 | 7 | 7 | 7 | 2.5 | 6 | 6 |
| 18 | — | — | — | — | — | — |
| 19 | — | — | — | — | — | — |
| 20 | — | — | — | — | — | — |
| 21 | 6 | 5 | 7 | 4.5 | 7.5 | 8 |
| 22 | — | — | — | — | — | — |
| 23 | 7 | 7 | 7 | 5 | 7 | 7 |
| 24 | — | — | — | — | — | — |
Interesting findings
- Overall 41% of spellcasters have no damaging spells.
- Overall only about 25% of spells in a damage dealers spell kit actually do damage.
- Damage capable spellcasters usually have a slightly larger spellkit on overall medians, by about 2 more.
- Of 1/day spells in a damage dealing spellcasters kit (usually highest level spells in the kit), overall only 58% of those spells are damage dealing spells. Obviously this varies, for some its 100% others its 0% and everything in between across all CRs, but it just goes to show its not always damage.
- From CR12 up, for damage spellcasters the median level of spell they have access to is about level 4.
- the 10 most common spells in non-damage spellcasters are: Mage hand, detect magic, detect thoughts, minor illusion, plane shift, detect evil and good, dancing lights, thaumaturgy, animal friendship, gaseous form, prestidigitation and Tongues
- The 10 most common spells for damage spellcasters are: Detect Magic, Mage Hand, Detect Thoughts, Invisibility, Scrying, Prestidigitation, Mage Armor, Thaumaturgy, Fireball, Shapechange.
- The most common spells that do damage are: Fireball, Lightning Bolt, Mind Spike, Ray of Sickness, Thunderwave, Scorching Ray – (then after those. All in equal measure:) – Befuddlement, Blight, Cone of Cold, Disintegrate, Finger of Death, Flame Strike, Guiding Bolt, Power Word Kill and Shatter.
So that’s that! Thanks for sticking with it! A bit of an abrupt end, but hopefully this helps a little when you build a monster! Obviously these are all only guidelines and recommendations, but monsters are much much more than just basic stats! Hopefully to budding GMs looking to take their first steps into the realm of monsters this can at least be a good first step that feels comfortable!
Any glaring issues please do give us a shout! We’d love to hear peoples thoughts!
Wishing you all the best on your next adventure,
Harry