Estimating CR

[Biggus]

I'd be interested to hear any formulas/tables/rules of thumb people used to estimate CRs. The two main ones I know about currently are:

Vorpal Tribble's calculator (https://forums.giantitp.com/showthread.php?192329-CR): I've occasionally used this but it's not very reliable, for one thing it doesn't include some important information (such as attack and damage bonuses) and also it tops out at 25HD (and even if you extend the formula it doesn't work well for high-HD creatures).

This table (https://forums.giantitp.com/showthread.php?623578-3-5-Average-Monster-Stats-UPDATED-TABLE&highlight=average+monster+stats) I've found more useful (or to be precise a smoothed version of it, the values jump about quite erratically in places) but again it's quite limited in what it includes, and at very high and very low CRs there aren't enough examples to make a meaningful average.

So, any suggestions? I know you ultimately have to eyeball it as many monsters have individual quirks that can't be easily captured in a table or formula, but anything which can provide useful information to feed into that decision would be helpful.

[Puke]

Hello,

estimating CR is very hard. This is kinda a die and retry thing. Throw some things at your players, note the CR of the encounter. If it is too easy, raise it.
The CR is an aproximate estimation. It ultimately depends on the PCs group.

A group with only magic users might have a really hard time facing a beholder, even at higher levels.
Same thing for a flying creature if none of your players can throw anything at it.

The CR of an encounter really depends of your group. So you have to know their abilities well in order to design good encounters. Remember it's always a bad thing to throw monsters who are immune to all PCs abilities, but you need to know what they can and cannot do in order to build something fun for them.

[pabelfly]

CR is a really rough ballpark at best. Start a group with a few different types of encounters at the same CR - groups, single enemies, ranged, melee, etc and see what your group deals with well and what they're not so great against.

When you start upping encounter difficulty to suit your player's strengths, have a backup plan to reduce the difficulty if your players need it. An important NPC arrives and diverts some attention from the enemy, the enemy suddenly weakens because they were under the effects of buff spells, stuff like that. You can even fudge the numbers a little.

[lylsyly]

CR is a really rough ballpark at best. Start a group with a few different types of encounters at the same CR - groups, single enemies, ranged, melee, etc and see what your group deals with well and what they're not so great against.

When you start upping encounter difficulty to suit your player's strengths, have a backup plan to reduce the difficulty if your players need it. An important NPC arrives and diverts some attention from the enemy, the enemy suddenly weakens because they were under the effects of buff spells, stuff like that. You can even fudge the numbers a little.

I used this ploy for many years in a campaign I ran. Had a half celestial template elf bard with a tibbit Sorcerer pretending to be his familiar. Kept them at the Groups level +3 at all times. Used them to bail the party out of trouble, give the party guidance on what the should do next, blah blah blah. Yes, it's kind of railroady but it worked out and the group always had fun (that IS why we play after all0. ;-)

[pabelfly]

I used this ploy for many years in a campaign I ran. Had a half celestial template elf bard with a tibbit Sorcerer pretending to be his familiar. Kept them at the Groups level +3 at all times. Used them to bail the party out of trouble, give the party guidance on what the should do next, blah blah blah. Yes, it's kind of railroady but it worked out and the group always had fun (that IS why we play after all0. ;-)

Not a bad idea since the group seems to like it. If it works it works.

[Biggus]

Thanks for the replies, but I already know all that. More what I was looking for is things like: has anybody ever compiled a list of average ability scores, total attack bonuses, damage etc by CR? Has anyone attempted a rating system for special abilities?

Or, another example: in the thread (https://forums.giantitp.com/showthread.php?636746-Which-cr-is-this-creature-(3-5)) that inspired this one, Fouredged Sword said they're wary of creatures which have a CR lower than their caster level. That's the sort of rule of thumb I was thinking of; it doesn't have to be mathematically precise, just a "bear this in mind when assessing CRs".

[Fizban]

Not usually in that level of granularity, but I guarantee there are some lists like that out there (or were, if they were on the WotC boards and got nuked). But I don't have any bookmarks because I find them useless.

DnD monsters can be highly specialized, just like PCs. Should those specialists count towards an average? Is an average even useful? And what books are being included? In addition to the overpowered outliers, early splat monsters also include some laughably weak outliers. And then there's a massive break in power levels between the earlier and later monsters, where MM 3 and FC 1-2 are trying to complete with The Internet and MM 4 and 5 start soft-testing various gimmicks leading into 4e.

So an average or table of everything is useless, and a table of only X/Y/Z books is useless unless you're using those exact books. Instead what you need is a list of good examples for their CR- and it has to be a list you personally picked to match your desired/party's power level, or else it still won't matter. The best an outsider could do is try to find say, the best, the worst, something in the middle, and their own (clearly marked) personal preference, for each general category at each CR, whatever categories they're using.

But I don't think any single book every has enough variety at each level to fill out such a list on its own, so any comprehensive CR example list will be forced to use multiple books, and then present fluctuating top and bottom ends depending on which books had the needed type of monster at the needed CR.

If you want some general rules in words, off the top of my head:


At CR 3 bruisers could have nearly 50hp (see 45hp Dire Wolf), but could also have closer to half that (see 29hp Ogre).
Monsters with humanoid capability (hands, thinking) are expected to be weaker than dumb bruisers, which are expected to be weaker than mindless bruisers.
Monsters with lots of SLAs can sometimes look suspiciously like "Clerics" with HD roughly equal to CR along with decent armor and weapon attacks that exist but aren't great.
Monsters with HD significantly below their CR are probably busted by simply being too fragile- significantly higher than their CR and magical abilities means they're going to start swinging the other direction towards "Dragon power."
Templates which reduce CR are a lie, particularly when applied to high-HD monsters (see Skeleton/Zombie, and in particular Skeletal/Zombie Dragon)- they may reduce the effect sufficiently, or they may not: an eyeballing modification to an unknown eyeballed monster is a recipe for failure.
Templates which increase CR may or may not be a lie.
"Menu Monsters" such as True Dragons, which don't have any default statblock because the DM must write it, are never a good judge of CR: their power is always dependent on how the DM builds them.
Corollary: It should go without saying that the given CR is for a monster's statblock is as written, not including whatever equipment/feat/spell/etc changes a DM considers making to them.
Monsters often have a line that says they use their treasure if appropriate: this is a statement for logical verisimilitude, not a license for the DM to char-op their monsters, and if the DM does so they need to re-evaluate the CR which they have now changed.
The idea that CR=Character level is a lie. It may be what the DMG says to base XP rewards on, but any amount of testing shows that PC-classed NPCs with chosen gear never line up with monsters: they're more fragile, and can range from less powerful to one-rounding the party. PC-classed NPCs are never a benchmark for CR (and indeed, should only give XP and be used as appropriate for their actual combat effect compared to benchmark monsters)

"Good Guy" monsters, particularly Good Outsiders, are generally better than their Evil counterparts: the Good guys are supposed to win, and their power level should not be used as a benchmark for new Evil guys (this also cuts off a ton of excuses for just giving away piles of PC casting on top of thickk bodies and piles of SLAs).
MM1 has a creature with an instant-kill gaze attack at CR 8, the Bodak- and that's just about all the Bodak has. If your monster has instant kills at CR 8 or lower, and is ridiculously tough on top of it, it's probably busted (I'm looking at you Drowned)
MM1 monsters with gaze attacks in general don't have special screw-yous for fighting them blind: you're expected to be able to close your eyes after all.
No monster of CR X should be able to apply lasting penalties or effects which a Cleric of level X cannot remove (MM1 does in fact bring in petrification before it's appropriate, but Shadows can be fought with Cure spells and Lesser Restoration after you've run away and prepared the next day [Allips are busted]).
If a monster has been printed in multiple books, check every version: it may have been massively ludicrously powered up (the Nycaloth from MoP to MM3), or merely mildly broken by an edit based on a different reading (the Wyste from The Speaker in Dreams to MM2), or completely unchanged- and then when you're done, check it against whatever you're actually using for your benchmark.
If a monster is from MM3, FC 1, or FC 2, it's going to be way stronger than MM1. MMs 4 and 5 are more gimmicky but still generally stronger.
PF has tons of little changes including basic monsters (they cut a HD off the Dire Wolf for example), so don't assume they're the same as 3.x.
Any 3.0 product that calls for a skeleton or zombie is expecting the 3.0 flat statblocks rather than the 3.5 templates.


The standard expectation of an encounter is that neither the party nor the monster has a particular advantage, however, it is also that the monster actually has to be fought: thus "dumb bruisers" are expected to be able to get in range and take some swings (so an environment that merely lets them do so is not an advantage), and the party is expected to be able to run away (because not being able to run is a clear disadvantage).
-Corollary: An ambush monster should never be capable of 1-rounding anyone on pure damage: pouncers and flyers (who can dive for double claw damage), are expected to do those things in normal combat, not be pass-or-die checks on "detection" abilities (which are not required).
-Corollary: No monster should be capable of 1-rounding the meatshield on pure damage output (assuming the meatshield is meatshielding rather than pretending they're "DPS"), or if so, not until at least the point where instant-kill abilities are common. Their entire reason for existence is to be able to take those hits, and if you're using martial monsters so powerful the toughest person in the party can't survive them, you've changed the fundamental nature of combat entirely. Pick your benchmarks accordingly.
-Corollary: Monsters with bypassable defenses are expected to have those defenses, at least sometimes. These are the textbook Easy if Handled Properly encounter, and if your party casually bypasses DR and regeneration, then those monsters are now over-CR'd, unless the party is paying handsomly in build resources to achieve it (and of course if they can never bypass it the CR is too low).

[Saintheart]

I can't really add profitable formulae or calculations, but I can think of a couple of rules of thumb I tend towards or have tried out. These aren't really about assessing CR, but :

(1) Assume most 3.5 combats will last about four rounds. After that the outcome is usually clear unless you reinforce or the party has hit on a solution that will kill the monster via attrition, e.g. kiting something that has no ranged options, in which case you fiat that the party kills the monster immediately and move the hell on ... or introduce a new opponent, at which point it becomes a new encounter entirely.

(2) Despite the panoply of monster abilities, when you get right down to it they generally do one of four things: raise the monster's hitpoints in effect, raise the DC of its special attacks or the attack bonus of its weapons, raise its average expected damage per round, or raise its AC. DR, for example, in effect just increases a monster's hitpoints in most circumstances, because it reduces the rate at which it loses hitpoints in those four rounds. Without this being an absolute rule, this personally does help me simplify down a monster's stats to work out how dangerous it really is to the party. High DPR can make something dangerous even if its hitpoints are low, and consequently something that raises a high DPR creature's attack bonus makes it more dangerous in rough terms.

(3) If I'm looking to have a monster last reasonably, I try to set its AC somewhere in the range that the party's primary damage dealer will by default hit about 50% of the time, absent any assistance from feats or buffs from fellow party members. This gives the player enough of a sense of progress that they don't feel like it's pointless to close to melee, and it makes the Power Attack choice sort-of meaningful.

(4) Honestly, if the party is not a pack of new players, I'll tend to go above and not worry about CR, although not absurdly so. Players tend to be more inventive than you think and I don't pretend to be some great tactical genius of third edition, so I'll tend to allow higher CR and let the players figure out how to survive. In PbP I don't often schedule easy fights because combat's time-consuming and already hard enough to maintain players' excitement over without spending two weeks on a fight where they just steamroll the opposition. Also, in PbP you have more time to think through your options and come up with ideas and strategies. But as said, that's in the PbP context, Ftf is quite different obviously.

(5) Corollary being, I build for significant challenge in most of my fights, but I pull my punches if needed. I don't really want to TPK my players, so I will throw lifelines with more frequency than usual. I admit that most of the time the players know a lifeline is being given to them when this happens, but at least by that point they've experienced the fear of loss of their characters, and thus I've gotten some sort of emotional reaction out of them, which is a big point of playing a RPG. This also means I tend to steel myself against whining that the fights "aren't fair", because if their players survive them, I tend to think a fight is fair ... and when calculating XP and treasure, it's right down the line on the book's own calculations, I don't adjust treasure downward. If they wind up with higher WBL than appropriate, frankly I don't care and often the logistics of finding above-level gear put a crimp in their quest for nuclear weapons.