D&D 5E DMG's definition of "Deadly" is much less deadly than mine: Data Aggregation?

[FormerlyHemlock]

I am actually coming to think WotC might be overvaluing the action economy of the monsters by just a bit. Maybe only give it the 1.5 once you hit more monsters than = party count, and x2 when you hit more than x3? something like that?

The basis for the x1.5 and x3 is simple arithmetic based on the assumption of single-target attacks, i.e. Lanchester's Squared Law. If PCs have area effects like Shatter, Web, Hypnotic Pattern, etc., x3 will tend to overvalue the monsters. Similarly, monsters with area effects will be deadlier against large number of PCs than their numbers indicate. You know how you read about groups of 6 7th level PCs taking down Balors with ease even though it is Deadly? I would dearly love to pitch an "Easy" fight of 14 Magma Mephits at such a party. Have fun killing 308 HP worth of enemies with a 98 HP breath weapon while you're under disadvantage and constant damage from Heat Metal...

I expect the PCs would win but I doubt they'd feel like it was an Easy fight. Against a Hard number of Magma Mephits (33 total) I expect they wouldn't win at all. Say hello to 231 HP worth of AoE on top of 66d8 fire damage per round, PCs...

 

[Henry]

Learn something new every day; namely, that there's jargon for the mathematics of murder. Shouldn't have been surprised, but it caught me off guard.

In the Mephit example, you're still taking the action economy to crazy levels at that point -- anywhere from a 3:1 to at 8:1 disadvantage, stuff breaks down. There's not a simple to use system that would account for all that. On the other hand, there's got to be some tweaking that could put the numbers a little closer to empirical outcomes - I applaud Quickleaf for attempting, anyway. Problem is it's also got to account for player tactical skill (my 5e group is a bunch of cool ladies and gents, but I think only one of them besides me has been playing more than two or three years), number of area effect users (as Hemlock notes), terrain (my players had a harder time with a fight with two CR2 Perytons than with the aforementioned orcs, but mainly because of ranged attacks limited to 60 feet or less, and so forth.

 

[cmad1977]

I think the balance is more about 'feel' than hard number.

 

[FormerlyHemlock]

Learn something new every day; namely, that there's jargon for the mathematics of murder. Shouldn't have been surprised, but it caught me off guard.

In the Mephit example, you're still taking the action economy to crazy levels at that point -- anywhere from a 3:1 to at 8:1 disadvantage, stuff breaks down. There's not a simple to use system that would account for all that. On the other hand, there's got to be some tweaking that could put the numbers a little closer to empirical outcomes - I applaud Quickleaf for attempting, anyway. Problem is it's also got to account for player tactical skill (my 5e group is a bunch of cool ladies and gents, but I think only one of them besides me has been playing more than two or three years), number of area effect users (as Hemlock notes), terrain (my players had a harder time with a fight with two CR2 Perytons than with the aforementioned orcs, but mainly because of ranged attacks limited to 60 feet or less, and so forth.

I'm only taking the mephits to crazy levels because they are so cheap from an XP budget standpoint that I can't resist. You could put six magma mephits up against six PCs and they would still do far more damage than their not-even-Easy rating indicates, due partly to their AoE damage. Honestly I don't think 3:1 odds against the PCs are even all that crazy. 3:1 odds were ubiquitous in the old AD&D Gold Box games (my only exposure to 1E) and pretty common in 2nd edition as well. In 5E, I rather feel that any mook fight that doesn't feature at least 3:1 odds hardly even counts as a mook fight, and 10:1 is better.

And there is a fairly simple system to account for it: vs. AoE, Lanchester's Linear Law applies, and vs. direct attacks, Lanchester's Squared Law applies, so do what modern militaries do in their simplest analyses and take the 1.5 exponent. Bam! Scales easily to any number of opponents, and predicts that 33 Magma Mephits are about 3.5 times as dangerous as 14 Magma Mephits, which honestly seems about right to me. I'm exaggerating the simplicity unfortunately because you still need to account for heterogeneity (e.g. 10 Magma Mephits and an Orc War Chief), and furthermore even if you could solve heterogeneity problem the overall tactical problem is still full of discontinuities, uncertainty, and nonlinearity. So yeah, general difficulty prediction intractable--but that's not because action economies are intractable. It's due to other factors.

And yeah, I realize I am agreeing with your ultimate conclusion ("difficulty is intractable") but I'm nitpicking some of the points along the way in hopes of clarifying the problem.

 

[I'm A Banana]

The intractability of difficulty is part of why I like the feel of an XP system that awards encounter-based XP based on what you actually did - that is to say, you get XP for a "Hard" encounter when one PC drops, or a "Medium" encounter when you use a daily ability, etc. While this does mean a relatively simple combat can give you a lot of XP if a group of tactical numbskulls flail at it, it does account for actual difficulty, rather than hypothetical projected difficulty.

It doesn't solve the problem of predicting how much a party can take before going down, however. But that's always going to be at best a rough guess - no estimate survives contact with the PC's.

 

[FormerlyHemlock]

The intractability of difficulty is part of why I like the feel of an XP system that awards encounter-based XP based on what you actually did - that is to say, you get XP for a "Hard" encounter when one PC drops, or a "Medium" encounter when you use a daily ability, etc. While this does mean a relatively simple combat can give you a lot of XP if a group of tactical numbskulls flail at it, it does account for actual difficulty, rather than hypothetical projected difficulty.

It doesn't solve the problem of predicting how much a party can take before going down, however. But that's always going to be at best a rough guess - no estimate survives contact with the PC's.

I don't think that does account for actual difficulty. What it does is account for actual injury. You could call it "masochism XP", and a direct consequence would be that it incentivizes you to play as stupidly as possible most of the time, while holding resources in reserve (e.g. Relentless Endurance, your best spells) that let you nova effectively even when you appear to have been mostly defeated. Healing spells would become more valuable because preventing damage is worse than taking damage and then healing it afterward. Diviners might use their abilities against the party to drop more PCs in hopes of getting more XP. Etc., etc.

The root of the problem is that you're assuming that outcomes are a rough proxy for difficulty, and you're using that as the basis for a system that rewards facing difficult challenges, which undermines the assumption your system is built on because it changes player incentives.

 

[Quickleaf]

[MENTION=2067]I'm A Banana[/MENTION]
"Degree of injury/death determining difficulty post-battle" is an interesting hypothesis, but ultimately one I'd reject.

Take a my group of players (six 2nd level PCs) who thru good use of tactics, readied actions, trap setting, ranged attacks, and focus fire were able to pound 3 carrion crawlers into the dirt. I asked them how it felt and the experienced players said "deadly, but we played smart, and we got lucky."

How would these players react if I said, "well, you may have played smart, but only one of you was paralyzed (and got over it), and only one of you was KO'ed (and got better), so this encounter will only be worth Hard XP not Deadly XP"?

You're damn right they'd be bitter or up in arms!

Of course, I could just not tell them, but that's not my style.

 

[I'm A Banana]

I don't think that does account for actual difficulty. What it does is account for actual injury. You could call it "masochism XP", and a direct consequence would be that it incentivizes you to play as stupidly as possible most of the time, while holding resources in reserve (e.g. Relentless Endurance, your best spells) that let you nova effectively even when you appear to have been mostly defeated. Healing spells would become more valuable because preventing damage is worse than taking damage and then healing it afterward. Diviners might use their abilities against the party to drop more PCs in hopes of getting more XP. Etc., etc.

The root of the problem is that you're assuming that outcomes are a rough proxy for difficulty, and you're using that as the basis for a system that rewards facing difficult challenges, which undermines the assumption your system is built on because it changes player incentives.

The thing is, is that if they play as stupidly as possible, then they actually do come close to having a TPK. Or put another way: if you respond to the incentives by edging as close to a TPK as possible without going over through dumb play, you still risked that TPK, and the XP you'd earn reflects the actual damage done to your party, even if it was mostly self-inflicted. And you'd have to play pretty smart to know how dumb you could be without being so dumb that everyone dies. Since spending resources would be part of what raises the difficulty bar, a maximizing party would be incentivized to USE those resources, early and often, because the weaker they get, the more XP they earn. It encourages a bit of recklessness - you are rewarded for taking on a risk and not quite making it - that interestingly follows real-life learning, where you only learn more if you're pushed to the limit of what you can do. If the party of 4 goblins isn't pushing you to your limit...well...hand me that blindfold, I'll learn more that way.

Certainly not a fully-developed concept, though, just something I kind of like the vibe of, bouncing around up there.

Take a my group of players (six 2nd level PCs) who thru good use of tactics, readied actions, trap setting, ranged attacks, and focus fire were able to pound 3 carrion crawlers into the dirt. I asked them how it felt and the experienced players said "deadly, but we played smart, and we got lucky."

How would these players react if I said, "well, you may have played smart, but only one of you was paralyzed (and got over it), and only one of you was KO'ed (and got better), so this encounter will only be worth Hard XP not Deadly XP"?

You're damn right they'd be bitter or up in arms!

Why would they be up in arms? They just had a Hard fight! Nobody died! That's a good thing! Yeah, they didn't learn as much from it as they might've if they came inches from death, but clearly they already know what they're doing, clearly. They don't need another level to handle these things!

Think of the bragging rights when they tackle Orcus at level 3 because they're just that good.

 

[Iosue]

My question is, how exactly are defining "deadly"? The DMG defines deadly as "could be lethal to one or more players." Every time this subject comes up, the encounters described strike me as being pretty deadly by that definition. Reading through the thread, it seems like you're defining it as "a decent chance of a TPK." Would that be accurate?

 

[FormerlyHemlock]

The thing is, is that if they play as stupidly as possible, then they actually do come close to having a TPK. Or put another way: if you respond to the incentives by edging as close to a TPK as possible without going over through dumb play, you still risked that TPK, and the XP you'd earn reflects the actual damage done to your party, even if it was mostly self-inflicted. And you'd have to play pretty smart to know how dumb you could be without being so dumb that everyone dies. Since spending resources would be part of what raises the difficulty bar, a maximizing party would be incentivized to USE those resources, early and often, because the weaker they get, the more XP they earn. It encourages a bit of recklessness - you are rewarded for taking on a risk and not quite making it - that interestingly follows real-life learning, where you only learn more if you're pushed to the limit of what you can do. If the party of 4 goblins isn't pushing you to your limit...well...hand me that blindfold, I'll learn more that way.

I'm okay with the idea of gaining extra XP through handicapping yourself, fighting blindfolded/left-handed/whatever. (In fact, I do something similar when I let my players skip over multiple similar encounters, e.g. hunting down 35 ropers on their island, if they can beat one tough encounter, e.g. 3 ropers making every roll at disadvantage while PCs make every roll at disadvanage. Although the logic I use there is not about handicapping but rather "this was the hardest fight.")

I just don't think that measuring outcomes is a good way to measure how much PCs were handicapped/challenged. I disagree very much, for instance, that you have to be very smart to play dumb enough to take "heavy" damage against light opposition which is nevertheless not heavy enough to actually endanger you. If I'm fighting eight goblins, I charge through them, taking opportunity attacks, until I'm down to about 50% health and then I smack a goblin and charge back behind friendly while a new PC takes my place. Due to the way HP work in 5E, as long as somebody saves a few Healing Words (a minor resource), the risk is minimal. Once everyone is moderately damaged, we kill half the goblins and do it again to get down to 25% health or so. Now the fight counts as Deadly since almost everybody took extremely heavy damage in the process and used up a lot of resources (maybe the wizard blows his spells on inefficient spells like Chromatic Orb V), so we get a ton of XP. Meanwhile the PCs are secretly hoarding whatever resources you haven't managed to account for when you designed the difficulty guidelines, which could be Lucky rolls or Shield spells or Rages or something else to make that last 25% of HP far more robust than the first 75% was.

What ends up happening looks far less like "learn to fight while blindfolded" than "learn to fight like the keystone cops." And like Quickleaf's players, I would absolutely hate that system--it's the worst kind of treadmill from my perspective in that it penalizes skill with pathological incentives. I can enjoy non-combat D&D, but that system would motivate me to avoid all XP progression and combat altogether, as much as possible, as a reaction. I'd wind up playing D&D like GURPS (i.e. focused primarily on skills and social interactions), in which case there's really not much point in it being D&D at all vs. freeform roleplaying. It doesn't sound completely unpleasant but it does sound pathologically counterproductive--I'd have to be really good friends with the DM and other players to enjoy spending time on such a relative waste of time.

Why would they be up in arms? They just had a Hard fight! Nobody died! That's a good thing! Yeah, they didn't learn as much from it as they might've if they came inches from death, but clearly they already know what they're doing, clearly. They don't need another level to handle these things!

Think of the bragging rights when they tackle Orcus at level 3 because they're just that good.

Think of the anger when they die to Orcus at level three because they didn't play more stupidly against the hook horrors.