D&D 5E (2014) The Glass Cannon or the Bag of Hit Points

[Celtavian]

No, tailoring is bespoke content creation and has zero in common with karma, which is a runtime metagame construct. In point of fact, it is how I completely obviate the need to tailor anything at all. It means there is always a price for failure, but that price is never unfair.

Why do you like using bespoke instead of tailored? It's pretentious. Karma sounds like tailored rules to suit how you want to run the game. Runtime metagame construct? You're using a video game reference for a TTRPG? You essentially run encounters from what I have read that the party has little ability to defeat without you constructing the encounter to make the enemy act in a disorganized and inefficient manner. I view that as tailoring. I'm assuming you find the imagery in your mind interesting or derive some pleasure from creating these scenarios otherwise you would not create them.

My games are crafted more like an author crafts a story. The DM is in charge of the setting, antagonist, secondary characters, and plot. The PCs are in charge of the main characters. The rules create a sense of randomness and fate during the combat/encounter resolution process that keeps everyone guessing as to the outcome.

I'm not even sure why we have discussions. Our perspectives are so foreign to the other as to be nearly incompatible. Story is my main interest. I want characters that are worth telling a story about. I wouldn't even play this game without some kind of story to play in. You and chriton are looking for a different experience apparently. I find sandboxes as they are called boring and trivial.

IWD2? I have no frame of reference. Is that a video game?

 

[Celtavian]

You're mistaken BTW about my motives for playing the beholders erratically. It's not about giving the PCs a break, it's just more realistic and more fun for me than a coordinated battlegroup led by a live hive queen. I want these beholders to be a static threat that mostly self-destructs instead of taking over the whole kingdom, ergo the hive queen is dead in the crash. This would be true if the PCs were first level or twentieth. And again, this while discussion arose from you asking how self-paced adaptive difficulty can exist in D&D. Well, that's how.

This is nonsense. 24 beholders should be able to destroy any party if played even somewhat intelligently. It is not "realistic" as you say. It is ridiculous. Successfully turning beholders into the equivalent of orcs is not what I refer to as "self-paced adaptive difficulty." It's a DM creating a ridiculous scenario, then running what is supposed to be one of the most fearsome creatures in D&D in a clown-like fashion to keep his PCs alive.

Whenever I read examples from your campaign, I can't take them seriously. I'm sorry. It's so ridiculous that I find it incredulous. I have zero interest in turning beholders into orcs. As I said, you and I are far apart in how we do things. These discussions seem to end up exactly the same in every thread we debate in and are unproductive.

I'm going to absent myself from the conversation. I have given my recommendations for how to challenge PCs. I know my method works. It doesn't require the improvisation chriton uses or the strange "runtime metagame" strangeness that makes your game work. It's good solid DM preparation based on PC capabilities within the context of the campaign story. It works nearly 100% of the time.

 

[FormerlyHemlock]

I'm going to absent myself from the conversation. I have given my recommendations for how to challenge PCs. I know my method works. It doesn't require the improvisation chriton uses or the strange "runtime metagame" strangeness that makes your game work. It's good solid DM preparation based on PC capabilities within the context of the campaign story. It works nearly 100% of the time.

Sayonara. Your One True Way is not the one true way you believe it to be. Sandboxes are a lot of fun for my table and the only kind of game I am interested in running or playing. If you like tailored plots then fine, but no thanks.

I have no clue why you equate beholders and orcs, but whatever. Apparently you've never played Spelljammer.

 

[AtomicPope]

When you tailor something, you are trying engage the players with something that is challenging or interesting. You take into account their capabilities because that is how you make something challenging or interesting. It does you no good to toss something out there that will utterly rip the party apart with no possible other outcome or to throw things in their way they walk over with no other possible outcome.

Tailoring encounters is specifically done to engage the players.

It's actually good that someone is bringing this up, and it's mentioned in Boxed Texted in the DMG when describing monsters, levels, and encounters. The specific mention is the Rakshasa, which has spell immunity to all spells that any party of equal level could cast. This goes against the standard encounter creation, especially in a party that is all casters, as its immunities would render the party powerless. The same could be said for flying creatures with Flyby Attack versus a HtH melee party.

The unwinnable situation goes both ways, and unless there's a storyline reason to make it happen (tailored) then there's no point in rolling the dice.

 

[FormerlyHemlock]

It's actually good that someone is bringing this up, and it's mentioned in Boxed Texted in the DMG when describing monsters, levels, and encounters. The specific mention is the Rakshasa, which has spell immunity to all spells that any party of equal level could cast. This goes against the standard encounter creation, especially in a party that is all casters, as its immunities would render the party powerless. The same could be said for flying creatures with Flyby Attack versus a HtH melee party.

Neither of those is unwinnable. Vs. Rakshasa, either Magic Weapon or Conjure Fire Elemental can suffice. Vs. Flyby, a Readied Grapple will change Flyby to melee.

 

[Celebrim]

The characters in my campaign are now rapidly increasing in power (currently 11th level) and easily capable of dishing out huge damage to opponents.

So far I've tended to go down the bag of hit points route so enemies last longer and can dish out some punishment of their own however I'm starting to think that maybe another approach is needed.

High level play creates huge design challenges. I think simply increasing the hit points of the monsters won't create challenge without undesirable sideeffects, but neither do I think what you call the 'glass cannon' approach is viable.

There are two different but important issues when dealing with high level PCs:

1) They can deliver the beatdown. This is what you are talking about. If high level PC's can unload on a single enemy, that enemy has to be extremely durable to survive even a single round.
2) They can dominate the action economy. This is what you are not talking about that you need to. In addition to beating down an opponent directly, high level PC's have the ability to steal actions from their enemies using debuffs or inflicting statuses. Combined with the fact that PC's typically have 4-6 actions to a single foes 1, this means that high level single foes will often go down hard to just a few 'unlucky' rolls.

The problem with the bag of hit points approach is that eventually in the effort to control the PC's ability to beat down a foe, you elevate the importance of the action economy too greatly. In most editions, this would result in the PC's prioritizing save or suck (or save or die) attacks, and consequently elevating the importance of spellcasters.

The problem with trying to address this with a glass cannon approach is that you might still end up elevating the importance of the action economy while risking the problem of one shotting PC's, which is generally no fun tactically or thematically. Keeping PC's on the verge of death provides tension and allows them to make choices. Winning the initiative and then randomly killing a PC before they can act doesn't make for fun for anyone.

In general, the best way to challenge high level characters is with more foes. The biggest mistake I see most DM's doing is trying to challenge high level PC's with a single foe all the time, when in fact this should be a relatively rare situation and requires very delicate handling. One of the best things 4e introduced was recognition of this fact, and though I didn't always agree with the specific implementation, the fact that 4e was trying to introduce formal design approaches to make encounters and particularly encounters with single powerful opponents interesting (especially at high level) was a good thing. 5e to some extent follows on this approach. If you are using 3e, recognize that most high level foes are relatively badly designed and require very specific handling and/or redesigns to make them more interesting.

Having multiple foes deal with most of the issues of high level design handily. The PC's are no longer as easily able to dominate the action economy. The monsters are generally no longer as able to one shot a PC, and not only can the DM reasonably narrate why the NPC's are focusing fire on a single target (lack of coordination, each monster attacks its most accessible foe, etc.) but often as not simple geometry prevents that sort of focus and the PC's can use formations and terrain to prevent that focus. And the monsters generally have a big bag of hit points that makes them hard to kill quickly, often far more hit points than any one single foe would have. The are still some design challenges in selecting the foe(s) you wish to use to challenge a high level party: they should have a ranged attack, they should have some defenses against common area of effect attacks, they should have an attack that is reasonably level invariant (touch attacks, save for half, very high attack bonus, automatic damage, etc.), they should have enough HD to avoid HD based win buttons, and so forth but in general this is an easier approach in most additions.

For a single foe, you really need something that has multiple actions each round, and preferably at least one being an area of effect attack. The goal here is be able to inflict lots of damage, but also be able to spread it out. A monster with both an attack and a damaging aura or gaze attack is a good example, or a monster with both an attack and the ability to use a breath weapon or a spell like ability as a free action. Alternatively, the monster can have many attacks, like a hydra, but have the monster specifically limited to only a certain number of attacks on each foe.

Monsters with 'reactive' attacks that can take place on the PC's turn also qualify here - often these are attacks triggered by attacking the monster. Examples are, successfully hitting the monster sprays acid blood, or raises a cloud of choking spores, or causes the weapon to become stuck, or perhaps the monster is covered with spikes or flames so that an attack draws a counter attack. Alternatively, the monsters body could be super hot or cold, resulting in heat/chill metal effects when you touch them. In some editions like 3e, having Combat Reflexes and melee attack with reach is also a good example of being able to reactively counter the PC's. Another idea would be having reflective armor that bounced spells if the monster's Spell Resistance (or saving throw) succeeded.

The single monster also needs to be able to threaten to steal actions from the PC's either by forcing them to move, forcing them to aid other PC's, or taking the PC out of combat for a round or two. This can involve a monster that skirmishes successfully and forces PC's to spend rounds chasing, or it can be a monster that has short term area based debuffs, or it can be a monster that can create terrain or obstacles over the course of the fight.

And this is on top of the monster needed to be durable enough to survive to use these abilities. At some point, this means more than just a bag of hit points but also the ability to resist or shake off debuffs, mitigate against spells, and possibly even mitigate against bad luck with some sort of 'destiny point' mechanic or some sort of 'on failure' mechanic.

A single foe doesn't necessarily need all of these abilities, but the higher the level of the PC's the more of these abilities a single foe has to have in combination to be an interesting challenge.

If you are playing 4e/5e, the designers tend to have gone at least halfway to where the monster needs to be with concepts like 4e's 'solo' foe or 5e's 'legendary' foe. If you are playing 3e or earlier, you are probably going to have to do a bit more thought when thinking about how to present a single tough foe.

 

[Celebrim]

In the current fight the party is facing a half dragon gladiator amongst others. My approach was I wanted it to hang around for a few rounds to threaten the PCs so I boosted its hp by nearly 70 extra giving it well over 150.

My thinking now is it should have have around 50hp but there should have been 3 or 4 of them!

Solo's are a unique problem due to the Action Economy. They'll never compete with the PCs if they are all able to focus on it so do you provide minions to help it or massively boost hp and threats so it's a challenge? One example is dragons, I've found all they can do as written is skirmish a party as a toe to toe solo is a slaughter.

Well, it sounds like you are aware of the challenges and are on the right path.

I've got a case study for single foes I can provide when I get home to give you an idea how in 3e I try to go from typical 3e design toward what I feel is successful monster design. Granted, this case is not intended for as high of level of play as you are currently at, but I think it is typical of what you need to do to make a single foe a more interesting encounter as it illustrates some of the points I made in the longer post above.

I think if I could sum up my point in one thing though it would be solo foes need increased damage dealing abilities but not increased ability to focus fire. Something like claw/claw/bite/rake/rend and so forth tends to result in combats where the only interesting aspect is who wins initiative.

 

[FormerlyHemlock]

Excellent analysis, Celebrim. I hope you don't mind if I quibble with a minor side-point:

The problem with trying to address this with a glass cannon approach is that you might still end up elevating the importance of the action economy while risking the problem of one shotting PC's, which is generally no fun tactically or thematically. Keeping PC's on the verge of death provides tension and allows them to make choices. Winning the initiative and then randomly killing a PC before they can act doesn't make for fun for anyone.

It can be fun if the players know that's a possibility (perhaps due to prior exposure). They can then enact countermeasures such as Death Ward, Cutting Words (decrease enemy initiative), stealth, and antimagic/antilife shells.

Besides, it's not like default 5E makes it difficult to overcome regular PC death. "Randomly killing a PC" = someone must spend an action and 300 gold on Revivify, which might be a totally fair price to pay for a victory. Disintegration and Finger of Death are a different story however.

However, I fully agree with you that "more foes" is an excellent way to challenge high-powered characters, whether that means "high-level" or "decked out with magic items" or "has lots of minions". 5E scales really well with quantity. The DM has to learn a few tricks for managing large combats, but it's not hard to do.

P.S. The only monsters I've found so far to make good solos are the ones with class levels. Specifically, I've used dragons with multiple levels in Dragon Sorcerer. Getting a dragon's regular attacks and legendary actions plus a Counterspell or Shield reaction plus a bonus action Hold Person V or Mass Suggestion plus any precast spells like Darkness (for the blindsight combo) or Mirror Image--that's enough of a boost to the dragon's action economy for them to take on a whole party at something approaching parity.

 

[Celebrim]

It's actually good that someone is bringing this up, and it's mentioned in Boxed Texted in the DMG when describing monsters, levels, and encounters. The specific mention is the Rakshasa, which has spell immunity to all spells that any party of equal level could cast. This goes against the standard encounter creation, especially in a party that is all casters, as its immunities would render the party powerless.

While this is an important thing to keep in mind, there is a significant risk of over tailoring to the PC's as well. Taken to an extreme, this approach results in railroads where the DM is deciding the outcome of encounters, and even in the less extreme case is probably just as likely to result in loss of dramatic tension as it is to increase it. At the very least, it tends to result in treating PC's with kid gloves and discourages the players from developing technical skills as players. As someone else rightly pointed out, just because direct attacks have been rendered difficult or impossible doesn't mean that the players ought to feel like they are out of options. A player of consummate skill is continually aware of what would challenge his party and is working to mitigate against that.

If a party as a whole decides to go all in on a single strategy, the DM needs to keep that in mind, but he also should avoid rewarding that approach by removing from the list of possible encounters things that hard counter that strategy. After all, for every such encounter, there is another encounter where the PC's strategy overwhelms the monster. If a particular monster is virtually immune to magic, there is another monster that is particularly vulnerable to it. If a particular monster is strong versus melee attacks, there is another that is weak when the PC's can get to melee range against it. If the DM is deciding that it would be unfair to present a party with a challenge that works against their character's strengths, should he also not present the party with challenges that do work to their strengths?

And that gets to the crux of the problem. When a player invests resources to being good at a particular thing, then he has a reasonable expectation that in actual play he will be in fact good at that thing. What I've actually observed in 100% of the cases where a DM takes a tailored encounter approach, is that the difficulty of the encounter scales to the level of resources that the party has invested. The trivial example is a rogue who has invested in being able to pick locks consistently, finds that in practice he picks locks no more regularly than a rogue that isn't is tremendous lockpicker because the DM sets the DC of opening a lock according to the PC's bonus to do so. The player that invests in great perception, finds that the more perceptive he becomes the more hidden and subtle all the clues become. This goes on and on, so that the player is a great archer finds that all his foes are either virtually invulnerable to arrows or just as great of archer as he is, and the wizard that excels finds that all his foes now have a great deal of magical resistance and immunity.

The way that a DM avoids all these subtle traps and maintains a proper neutrality is not over tailoring the world to the PC's, but instead sets a reasonable baseline for what is out there and then leaves it up to the PC's how to deal with those challenges irrespective to what he knows about the PCs. If the DM is always tailoring encounters based on what he knows about the PC's, he's metagaming and as a result the player's agency is being eroded to a lesser or greater degree.

There are far more subtle ways to tailor encounters to maintain a reasonable range of challenge (and yes, challenge in encounters should always have a range and not be tightly zeroed in on the PCs abilities) than to metagame off of the PC's abilities. A better and approach in my opinion is to tailor to the player's skill level by increasing complications in the encounter - increasingly complicated terrain, increasingly disfavorable situations, increasing cunning and preparation by the foes, hostages to rescue, fog of war, and so forth. Good encounters are challenging in the same way good puzzles are, and not merely because they chip off a certain number of resources.

 

[Celebrim]

Yuck, those games sound awful.

It's actually worse than he even makes it.

It's possible to implement perfect adaptive difficulty. You can set it up so that every challenge in the game is not only scalable but is exactly scaled to the level of the PC.

You can fix the math completely. Imagine a simple system where a 1st level character has 50 hit points and does 10 damage per attack, and he fights orcs that have 10 hit points and do 5 damage per attack. Later, the character is 10th level and now has 500 hit points and does 100 damage per attack, and now he is facing orcs that have 100 hit points and 50 damage per attack. The net result of level up has been, exactly nothing. The game play doesn't change in the slightest. Four orcs is just as much of a challenge at 10th level as at 1st level. From one perspective, this is great; we've solved the problem of the difficulty curve and regardless of what order the player takes the challenges, the level of challenge will be predictable. If we want the game to have a difficulty curve, we can have early encountered challenges be 1 or 2 orcs, and then graduate up to 3 orcs, and eventually 4 orcs. The hardest encounter in the game can be 5 orcs, and regardless of how much the player grinds, that encounter will always require some degree of 'skill'.

But in other ways, this is terrible design since really we are leveling up to no purpose at all except to make the numbers bigger. Games like this use leveling the way pinball games that multiply all scores by 100 or 1000 before displaying them to the player to make bonuses seem big and feel special use scores.

Indeed, if leveling is optional, video games like this often have degenerate strategies in never leveling at all as often as not the designer tried to solve the difficulty curve by actually skewing the encounter difficulty by making things ever so slightly harder as you level up. There are even cases where the optimal strategy in adaptive difficulty games is if you want to be good at something, invest your resources in being good in the things you don't want to eventually be good in. An equivalent situation in an PnP RPG is a party that wants to be good at avoiding traps, decides to have no rogues in the party because they know that their DM - upon observing that the party has no defenses against traps - will avoid placing traps in the environment or make them easy to avoid. The worst case result here is Schrodinger's Trap, which exists only if searched for and is more difficult to find and disarm the better that PC is at finding and disarming traps, so that the party with rogues suffer's more from traps than the one without it.

And this is not a hypothetical. I've seen this occur in play, so that the DM gives monsters extra hit points when facing the fighter, and gives dungeons extra traps when the rogue is present all in the name of 'challenging encounters' as if the player by being good at something is signaling that is precisely the area that they want to be gimped and frustrated. The results of course are exactly backwards, as you'd expect having a fighter in the party would make you strong in melee combat but vulnerable to traps, where as the reverse would be true when the rogue is present but the fighter isn't. But if tailoring is in play, who knows?