D&D 5E do CRs seem a bit arbitrary?

[Tormyr]

Yeah, it's almost like we need a "z axis" to determine how difficult a monster could be, given certain external factors. For example, a cave bear is a cave bear. You can generally look at that creature and know what to expect from the stat block, and CR should be fairly accurate given nearly any sort of "typical" circumstance (finding one asleep or at the bottom of a pit notwithstanding). On the other hand, the gelatinous cube is either trivial or deadly-up-to-TPK, depending on the environment. It's almost like you could use a color-coded system on top of CR; something like green for "fairly straightforward," yellow for "can be more dangerous given an ideal setup or if your party lacks a key ability," and red for "given a relatively common setup or without a specific type of counter this monster can be far more deadly than its CR might indicate."

Of course, the more you try to codify this stuff, the more confusing it gets, the more words you have to dedicate to explaining it, and the cost of everything gets higher - possibly without much gain, since most experienced DMs would likely intuit this stuff anyway.

Such a system would almost seem unwieldy. There are just so many combinations of stuff in stack blocks versus situations that could occur. The CR system seems to be based on "everyone take your turns and do your most deadly thing. Anything outside the scope of that falls into encounter building where the CR is just an XP value, but the extra stuff can modify the encounter in fun and interesting ways.

 

[FormerlyHemlock]

Making a poorly calibrated, swingy (from GM-side understanding of it) encounter system into a tightly calibrated, predictable (GM-side) system is an extremely significant project. Doing the inverse is absolutely trivial.

Conversely, a traffic system and roadway built to handle poorly-calibrated, swingy drag racers will have absolutely no problem with careful drivers in Ferraris, but a roadway that assumes everyone is in Ferraris will go down in flames if everyone switches to cheap drag racers.

5E can handle both precisely-calibrated Ferraris and swingy drag racers.

 

[Manbearcat]

Not really. I offered up an extreme example knowing that the fact that if a threat represented no challenge, it would be well known that no XP is earned.

But I think it's less well appreciated that for any given creature, there is a continuum of challenge which depends on the environment in which the creature is encountered - and this includes situations where sans the creature, the environment itself wouldn't represent a hazard at all.

Consider the case of a dire tiger encountered in a gladiatorial arena. A very large portion of the potential threat of a dire tiger is neutralized, and the encounter is turned into a very predictable slugfest between the PC(s) and a brute low AC monster where you can do a calculation like, "Tiger does average X damage per turn. Party does average Y damager per turn. How many rounds pass before the tiger drops a party member, or the party drops the tiger?"

But consider the case of the same beast encountered in a vast tall grass pampas of thick grass 9-14' tall with a muddy mire and shallow pools beneath it.

Suddenly things like low-light vision, +11 bonus to hide in favored terrain, +11 bonus to move silently, and scent become huge factors in the monster's challenge rating. In particular, a RBDM will have such cats stalk the party, attack at night, with the intention of surprising the party when they are trying to camp and killing the character on watch or as they break camp (while the cleric/wizard is preparing spells, for example). The 100% concealment past 5' completely counters missile weapons and many magic spells. Lone characters could be quickly overwhelmed by the tiger's pounce (perhaps 36 damage on the first round) and decent chance of surprise. The +24 grapple bonuses plus improved grab and enough strength to carry a PC as a light burden, means that a cat can move the grapple and run away with a victim (even one not yet unconscious) forcing the party to try to chase in a situation where the cat's broad paws might incur no movement penalty while the party itself is forced to treat everything as difficult terrain. The cleric and fighter in their heavy armor might be useless while the tiger is 'running away' with another character in its mouth. After all, the cat probably isn't interested in fighting to the death. It's intending to jump one target and go.

Imagine 4 encounters of this sort in quick succession, attacking isolated members of the party. Four encounters per day is fair, right?

But if the same party is flying on magic carpets over that pampas? Now, the encounter if it occurs at all is again trivial.

I'll see your dire tiger anecdote and raise you a dire shark/kraken and a roc. Skrimishing swimmers and flyers are even more dangerous.

The problem here lies in the fact that we're moving outside of the combat resolution mechanics. You're going to be breaking out the AD&D Evasion and Pursuit rules which are different than the 3.x Evasion and Pursuit rules which are different than 4e Group Checks or 4e Skill Challenges which are different than 5e Chases (but are the same as 4e Group Checks as they were ported directly over!). The common theme of each of these rules subsystems is that, as is pretty much the legacy with D&D, noncombat resolution mechanics are siloed away from the combat resolution mechanics. There is a different level of zoom, abstraction, and very different mechanical apparatus (relative to each other and relative to the system's own combat resolution mechanics) from subsystem to subsystem.

D&D's siloed combat mechanics just make the gameplay a very different experience than a game such as Dungeon World where the entirety of the game's mechanical infrastructure is unified so you're seamlessly able to just follow the fiction and go from combats to chases (etc) without having to toggle back and forth between subsystems (one of which has a tight action economy and all kinds of component parts that the other doesn't possess...and there is little to no overlap in potency between abilities brought to bear in one versus the other). For D&D's purposes, I thought 4e had a great answer to this. The conflict with such a creature would either be created as a Lurker Hazard/Trap (of relevant xp value where the danger is represented by its capacity in the "Lurker" role) or it would be an n complexity SC where n represents the danger the creature brings to bear with hit and run tactics (and the stakes are PC or NPC, better still, is dragged away and devoured...or the creature is caught and the combat resolution mechanics are then consulted for what happens next).

So in total, I'm in agreement with, I think it was, @Tormyr (?). In order for something akin to CR to be of value to a GM, it needs to be premised upon the deployment of a standard combat routine for the creature within the venue of the combat resolution mechanics. Its capacity outside of the combat resolution mechanics is very useful to know, but you don't need the same sort of precision in calibration (you do need calibration however...just not the same type) that you do in the D&D combat mechanics minigame (which are governed by all kinds of mechanical bits, bobs, units, and widgets that must interface coherently...because the stakes are generally rather high)

 

[Manbearcat]

Conversely, a traffic system and roadway built to handle poorly-calibrated, swingy drag racers will have absolutely no problem with careful drivers in Ferraris, but a roadway that assumes everyone is in Ferraris will go down in flames if everyone switches to cheap drag racers.

5E can handle both precisely-calibrated Ferraris and swingy drag racers.

I promise I'm not being willfully obtuse here, but I don't understand your meaning and how it relates to whether the 5e CR/encounter budget system consistently bears out predictable results for GMs (or not) when they're building encounters. And I'm not sure how it relates to the point I was disputing by KM that it is just as easy (or it is fair to expect GMs to do the work) to tightly calibrate a wonky system as it is to perturb a tightly calibrated system.

Can you help me out?

 

[FormerlyHemlock]

I promise I'm not being willfully obtuse here, but I don't understand your meaning and how it relates to whether the 5e CR/encounter budget system consistently bears out predictable results for GMs (or not) when they're building encounters. And I'm not sure how it relates to the point I was disputing by KM that it is just as easy (or it is fair to expect GMs to do the work) to tightly calibrate a wonky system as it is to perturb a tightly calibrated system.

Can you help me out?

KM can only speak for himself of course, but the reason I mentioned systems built for Lamborghinis is that my sense is that is closer to what he meant when he said that balance is deeply embedded into the 4E system: it's not a system that can handle destabilization, whereas 5E is designed to do so. I don't know 4E well at all, but 5E certainly is tolerant of heterogeneity. You can run a party in 5E with PCs running from 8th level down to 2nd level, with varying levels of optimization, and all can contribute. My understanding is that you can't do the same thing in 4E. Thus the analogy to systems calibrated only for perfectly balanced Ferraris vs swingy drag racers. In the analogy, 5E PCs are swingy drag racers.

Does that help?

 

[Ristamar]

I get the feeling that most people who strongly favor "Combat as Sport", particularly those that enjoy 4th Edition, are never going to be happy with the method for budgeting encounters in 5th Edition. It's a loose guideline in a game with a lot of combat related mechanics that rely heavily on DM adjudication.

And while I appreciate an entry for common tactics within a monster's description, I don't necessarily want that hard coded into it's CR or whatever label might be used to determine it's threat level. I'm not looking to run encounters as if they were an automated script.

 

[Manbearcat]

KM can only speak for himself of course, but the reason I mentioned systems built for Lamborghinis is that my sense is that is closer to what he meant when he said that balance is deeply embedded into the 4E system: it's not a system that can handle destabilization, whereas 5E is designed to do so. I don't know 4E well at all, but 5E certainly is tolerant of heterogeneity. You can run a party in 5E with PCs running from 8th level down to 2nd level, with varying levels of optimization, and all can contribute. My understanding is that you can't do the same thing in 4E. Thus the analogy to systems calibrated only for perfectly balanced Ferraris vs swingy drag racers. In the analogy, 5E PCs are swingy drag racers.

Does that help?

That does help, thank you. Have some xp. However, while related, it is a bit of a different vector. Basically a product of bounded accuracy and flattened power curve through the levels; the PC analog for the designer's intent to keep low level creatures a threat to the PCs at higher levels. It is interesting, nonetheless.

Certainly a facet of the rule system (bug or feature depending on who you might ask), but rather orthogonal to the issue of CR, encounter budgets (in a system that primarily calibrates for the adventuring day - top down - rather than the encounter - bottom up) and the predictability (GM-side) of challenge/tempo/climax for GM's who are seeking specificity/precision when creating situations likely to be resolved by violence.

 

[aramis erak]

Yeah, it's almost like we need a "z axis" to determine how difficult a monster could be, given certain external factors. For example, a cave bear is a cave bear. You can generally look at that creature and know what to expect from the stat block, and CR should be fairly accurate given nearly any sort of "typical" circumstance (finding one asleep or at the bottom of a pit notwithstanding). On the other hand, the gelatinous cube is either trivial or deadly-up-to-TPK, depending on the environment. It's almost like you could use a color-coded system on top of CR; something like green for "fairly straightforward," yellow for "can be more dangerous given an ideal setup or if your party lacks a key ability," and red for "given a relatively common setup or without a specific type of counter this monster can be far more deadly than its CR might indicate."

Of course, the more you try to codify this stuff, the more confusing it gets, the more words you have to dedicate to explaining it, and the cost of everything gets higher - possibly without much gain, since most experienced DMs would likely intuit this stuff anyway.

It would be really nice if the MM listed the OCR and DCR components...
A high OCR tends to indicate potential lethality.
A high DCR, especially if it's from AC, can mean a very long drawn out fight.

A high OCR mostly from damage indicates a heavy hitter... likely to kill.
A high OCR mostly from to hit bonus indicates a constant but lesser damage threat.
A high DCR mostly from a high save DC means you are unlikely to resist its spells/powers, and the spells/powers determine the nature of its threat... too variable.

A high DCR from AC means it's a long battle no matter who is facing it.
A high DCR from HP after lowering for a low AC means a monster that is easily hit but tough - your damage output is more important than your hit rate.

 

[Sunseeker]

Conversely, a traffic system and roadway built to handle poorly-calibrated, swingy drag racers will have absolutely no problem with careful drivers in Ferraris, but a roadway that assumes everyone is in Ferraris will go down in flames if everyone switches to cheap drag racers.

5E can handle both precisely-calibrated Ferraris and swingy drag racers.

That's not technically correct because you're not accounting for the fact that both, and everyone else, is on the motorway at the same time. That's where poor calibration becomes a problem, you're not just accounting for Bob racing his rice-rocket down main street. You're accounting for the fact that while Bob is doing that, traffic lights are going off randomly, other drivers are on the road, pedestrians are trying to cross and there's the occasional emergency vehicle roaring through.

A single kobold or a single dragon might be well balanced and fitting for its CR against an "ideal" and prepared party. But once you start throwing in bugbears, efreets and lair effects, CR becomes unwieldy because a "challenge" is not a linear, mathematical concept. A single kobold may be a CR 1/4 but 6 kobolds could be a CR2 because the tactics they can employ and effects they can take advantage of increase non-linearly with their numbers.

That's why its important to have a good transportation system, because you'll never know if the guy driving the big-rig is a kobold or a salamander.

 

[FormerlyHemlock]

A single kobold or a single dragon might be well balanced and fitting for its CR against an "ideal" and prepared party. But once you start throwing in bugbears, efreets and lair effects, CR becomes unwieldy because a "challenge" is not a linear, mathematical concept. A single kobold may be a CR 1/4 but 6 kobolds could be a CR2 because the tactics they can employ and effects they can take advantage of increase non-linearly with their numbers.

I wasn't claiming that CR is robust. I'm just claiming that 5E is built so you can ignore CR and wing it successfully. It's the system that is robust (partly through being easy), not the CR metric.