Hierarchy of Saving Throws

[Moreb Benhk]

In practical terms: do they? The monster examples you used don't really fit most generic adventure designs. If you talked to ten different 5E D&D tables chosen at random and they told you that they never encountered a ghost or an intellect devourer in combat from levels 1 to 11, I'd find that plausible unless they were all playing Curse of Strahd. However, if HALF of the ten tables never encountered a spellcaster capable of casting 3rd level spells, I wouldn't find that plausible at all.

Not only are the effects rare in terms of gameplay, but the monsters who cause those INT/CHA effects are also rare in terms of how adventures are designed. If I gave a list of monsters to twenty new DMs that didn't have stats, just two-paragraphs of description, and then asked them to draw out a quest spanning from level 1 to 11 the only monster I'd expect to see more than once or twice is the ghost.

I think it's also possible around the shared thematics of things like Mindflayers and Intellect Devourers - in a campaign where one is present, the other might be quite likely to be present too. So they might be part of an all-or-nothing thematic element - leaving Int saves relatively important or even more side-lined, depending on the campaign, but seldom how they'd appear to be based off the straight Monster Manual statistical prevalence.

[thereaper]

If you do the math on Mind Flayer, a trivial calculation of CR makes it CR 4 or 5 if you don't factor in the stuns and just assume it uses Mind Blast (on 2 targets), then Tentacles, then Devour Brain. It becomes an even CR 7 if you, say, double its effective HP because half the party will be stunned, and add 3 to its effective attack bonus because it'll be attacking with advantage on 2 out of 3 turns. I'd say that CR 7 is an accurate representation of a Mind Flayer's challenge as a solo monster.

There's something like a 20% chance of one or more members of the party dying in that encounter. I don't think that can realistically be called anything but Deadly at that level.

[MrCharlie]

Not if you're fighting something highly damaging like, say, giants. While the party is sorting out their broken line, the monsters aren't just sitting there gloating. It's likely that someone, perhaps multiple people, is going to die in the time it takes people to recover their actions.

And this problem will get worse as the game goes on. Half of your party forfeiting a round or two against high-CR Team Monster can easily result in a surprise TPK. And TPKs are much harder to recover from than the barbarian being send to have a vacation in the Nine Hells.

I guess, to start, I'm confused at why your using a spell that explicitly makes people run away as a TPK vehicle versus something like hypnotic pattern which makes them politely wait to be killed and turns the awakening attack into a critical. In practice both can be solved via breaking the concentration of the caster, but still.

But to put it another way; If we assume that everyone has a 75-25 shot of making any given save a prior, then with a single target INT or CHA save that effectively flips a party member, like Enemies Abound, we've lost a players action, and the actions required to pacify and constrain them. Generously, let's assume that's a 1 to 1 ratio; one INT saved have a 75% chance of negating two actions, one of which is now an enemies action. If we represent allied actions as 1, inactions as 0, and an enemies actions as -1, we've got a net of -3 actions with a 75% chance of hitting.

If we have a fear spell, we've negated 3 people outright, their actions are now somewhere between wasted, actively bad, and somewhat positive, depending on how good running is. Let's assume that running is neutral-either there is nowhere to go, or running is equally bad and good depending on who is doing it. That's a -3 net as well, but it's better because the chance of hitting was already factored in.

Obviously, as the number of people in the party increases, fear becomes better. However we are assuming everyone is clumped, and the AOE can hit everyone too without hitting the massive giants you've got looming over everyone. For that reason, I think 4 people is a fine comparison; often you'll hit more, but often you'll hit less or hit allies as well.

So you're right that fear may have been action economy for the enemies if every save is equally likely to fail. However one key difference is that players and monsters aren't made the same way. Player characters typically deal much higher nova damage than a comparable NPC. A CR 8 frost giant might deal 56 damage, and giants are giant boatloads of HP and raw weapon damage for their CR. But a player might be able to drop their own fireball on the "enemies" surrounding them and deal the equivalent of 140 damage to this four person party once afflicted with enemies abound. And that's honestly a soft-ball; by the time you've got something fearing the party and a bunch of giants, you're dealing with worse than a friendly fireball. It all depends on whom is targeted, and what they do.

I think my message here is that it depends a lot on the circumstances, but I can easily envision a situation where enemies abound ends up being miles worse and does significantly more to the party than fear does. This doesn't mean that enemies abound is a better spell than fear, like most AOEs if you can hit everyone it is probably better, but that I don't think the comparison is as final as you think.

(This is ignoring the fringe case where the charmed person cannot make follow up saves on a 20, which is where spells like enemies abound become a new beast-the enemy can cast it then run, while fear requires some follow up to be deadly).

As for plane shift-Put simply, you're not getting why it's so dangerous.

The real risk is almost never TPKing. The party does not simply lose, at least not most of the time, and almost certainly not at a level where 7th level spells pop up. Hell, at that level, yes, you can easily recover from a TPK. Keep a lock of hair from everyone in a safe box with 10000 gold and a note to reincarnate these people if they don't check in after a month. Then go recover your magic items and kick your killers' ass.

Meanwhile, plane shift imposes consequences even if the party wins. Yes, you've won, yes, you've stopped the evil mage, but that won't bring Bob the Barbarian back. Bob is in hell now. Hope you guys like fiends, it's road trip time!

Because of this, it's dangerous for an entirely different reason than other spells; not because it might defeat the party, but because it might hurt them in victory. And given that most of the time the odds are stacked in the parties favor to eventually win, this is actually a big deal.

(Even ignoring that plane shift, as a one action and done spell, enables fun with hit and run tactics).

[Deathtongue]

INT Saving throw aside, Enemies Abound is just not a very scary spell in the hands of NPCs. It's one of the few effects where it's a bigger pain in the neck for NPCs than PCs. There are just too many ways to subvert it if you're a PC: namely area of effect spells (which don't have targets) and just closing your eyes before making the attack. Not to mention just piling on nuisance damage on the PC of concern -- like, say, the victim intentionally triggering an OA or walking through caltrops.

I guess, to start, I'm confused at why your using a spell that explicitly makes people run away as a TPK vehicle versus something like hypnotic pattern which makes them politely wait to be killed and turns the awakening attack into a critical.

The real risk is almost never TPKing. The party does not simply lose, at least not most of the time, and almost certainly not at a level where 7th level spells pop up.

Eh? Aside from level 1, levels 4-7 are the most dangerous ones for PCs. That's when your party can get turbo-nuked by stuff like Fireball and Lightning Bolt. It is absolutely vital for you not to lose actions at this level range, because it's the difference between one or two Mages getting off two 8d6+ AoEs with no counterplay and only getting one such effect while being backed up by Counterspell / Absorb Elements / Evasion / (Mass) Healing Word / etc. How do you feel about your odds when a Hobgoblin Devastator slams your party with an extra-strength Fireball aright before a Priest orders three of your party members to grovel with an upcasted Command?

I brought up Fear because it can screw you over hardcore in a way that a DM can't anticipate. If you're on an open field or, god forbid, your speed is reduced you might not just lose two actions, you might lose five or more unless the problem is dealt with. And Hypnotic Pattern, like Enemies Abound, is just not as scary of a spell for prepared PC parties. That said, a lot of PC parties are not strategy ninjas like we are and they're unprepared to have familiars start the wakey-wakey chain or have low-damage AoEs on standby so I can see how people would just sit there and take it.

[MaxWilson]

Eh? Aside from level 1, levels 4-7 are the most dangerous ones for PCs. That's when your party can get turbo-nuked by stuff like Fireball and Lightning Bolt. It is absolutely vital for you not to lose actions at this level range, because it's the difference between one or two Mages getting off two 8d6+ AoEs with no counterplay and only getting one such effect while being backed up by Counterspell / Absorb Elements / Evasion / (Mass) Healing Word / etc. How do you feel about your odds when a Hobgoblin Devastator slams your party with an extra-strength Fireball aright before a Priest orders three of your party members to grovel with an upcasted Command?

Eh? Why single out levels 4-7? What makes the PCs immune to Fireball and Lightning Bolt at levels 1-3? Just DM pity?

What makes the party immune to Fireball and Lightning Bolt at level 8+? 5 Flameskulls is a Hard encounter at level 10, not even Deadly. What makes 5 Fireballs at level 10 so much less problematic than Fireball at level 4?

I wonder where you're picking these numbers from. They seem arbitrary to me--more about Deathtongue personal DMing style than about 5E per se.

[MrCharlie]

INT Saving throw aside, Enemies Abound is just not a very scary spell in the hands of NPCs. It's one of the few effects where it's a bigger pain in the neck for NPCs than PCs. There are just too many ways to subvert it if you're a PC: namely area of effect spells (which don't have targets) and just closing your eyes before making the attack. Not to mention just piling on nuisance damage on the PC of concern -- like, say, the victim intentionally triggering an OA or walking through caltrops.

Eh? Aside from level 1, levels 4-7 are the most dangerous ones for PCs. That's when your party can get turbo-nuked by stuff like Fireball and Lightning Bolt. It is absolutely vital for you not to lose actions at this level range, because it's the difference between one or two Mages getting off two 8d6+ AoEs with no counterplay and only getting one such effect while being backed up by Counterspell / Absorb Elements / Evasion / (Mass) Healing Word / etc. How do you feel about your odds when a Hobgoblin Devastator slams your party with an extra-strength Fireball aright before a Priest orders three of your party members to grovel with an upcasted Command?

I brought up Fear because it can screw you over hardcore in a way that a DM can't anticipate. If you're on an open field or, god forbid, your speed is reduced you might not just lose two actions, you might lose five or more unless the problem is dealt with. And Hypnotic Pattern, like Enemies Abound, is just not as scary of a spell for prepared PC parties. That said, a lot of PC parties are not strategy ninjas like we are and they're unprepared to have familiars start the wakey-wakey chain or have low-damage AoEs on standby so I can see how people would just sit there and take it.

If your PCs are cheesing enemies abound, then you need better PCs. I mean, seriously man, someone intentionally hurting themselves to break an effect they don't know is on them??? If we're to account for that level of metagaming than anything that has charm on it should be discounted by the PC will lawyer their way out of anything and everything! Hell, you can cheese Fear by closing your eyes and staring at the ground, the enemy is no longer in your "line of sight"!

The text is clear; everyone is an enemy. If they want to metagame that with AOEs, then they should be targeting the biggest cluster, which means the party is likely suffering AOE as well.

The chances that hypnotic pattern or fear actually last until ending normally is very low regardless. Both are concentration, and the standing rule in combat is kill the guy concentrating. The same is true of enemies abound, but in the meantime the players are killing each other. But regardless, Hypnotic pattern keeps you in melee, Fear doesen't and your entire point was that disabling people for the giants is what kills them, so...huh? And I've rarely seen fear do anything like what your suggesting; most of the time it's a simple turn waster, some of the time it leads a couple PCs to safety, but basically never does it disable PCs for five turns. Hypnotic pattern rarely does either, but neither have I seen it be easily nullified (how are familiars making their saves? How do party members know to hit allies with AOE to awaken them?)

But given that there are a couple of non-concentrating monster ways to charm using charisma, and this isn't limited to spells, we're kinduve off target anyway.

Also, level 7 spells aren't showing up at levels 4-7, the earliest CR I know of with that level spell is CR 9 (and it's non-combat) so a party isn't really supposed to be dealing with them at level 4 by any means. By the time you're routinely facing level 7 spells your into later tier 2 and early tier 3, at which point PCs start breaking the balance rapidly. And the most common TPKs are level 1, 2, and 3, and always have been. **** is simply swingy.

And if the party is facing multiple casters of course it's a bad day. In general, casters are poorly designed as enemies and should be used sparingly, for the same reason you don't design characters based on PC rules. If the DM is throwing two mages at a level 4 party, they lose. If the DM is throwing two mages at a level 6 party, he has no idea how to design encounters, and initiative is going to determine who wins unless the group is optimized. Hell, I guaranteeI can TPK virtually any party by intentionally rigging encounters by abusing the CR rules to create what seems like a reasonable challenge while being anything but! Stacking AOE is one of those ways, and says little about how good the AOE actually is.

Eh? Why single out levels 4-7? What makes the PCs immune to Fireball and Lightning Bolt at levels 1-3? Just DM pity?

What makes the party immune to Fireball and Lightning Bolt at level 8+? 5 Flameskulls is a Hard encounter at level 10, not even Deadly. What makes 5 Fireballs at level 10 so much less problematic than Fireball at level 4?

I wonder where you're picking these numbers from. They seem arbitrary to me--more about Deathtongue personal DMing style than about 5E per se.

Encounter design. Fireball throwing enemies are above the CR recommended for levels 1-3. If you want to TPK parties at level 1-3 without any mercy, you use magma memphits instead. CR 1/2, heat metal and fire breath, ability to fly, they appear to be a reasonable challenge but have two unavoidable damage sources, one of which is AOE, and even explode when they die. Two magma memphits is a "hard" challenge 4 level 1 PCs, but they are liable to drop half the party in one round and kill the rest with their death throes if the party is clumped up. It's not fireball but it works, at at CR 1/2 you can liberally season with them at higher levels.

But yeah, in general cherrypicking an encounter to a particular spell or situation is going to bias opinions of it. I could make my point look better by saying "Well, imagine if the PCs are in a box, then one of them gets hit with enemies abound from an archmage. Someone dies in this box!" or "Imagine an archmage swoops in, casts plane shift, and then his contingency dimension doors him away. Now he just does that four times. Boom, plane shift is overpowered!" Exaggeration and stupidity aside, you see my point.

Damaging AOE gets better as you add more overlapping instances of it. You can throw 14 hobgoblin devastators against 4 level 20 PCs, which is 112d6 of fireballs + potentially 28d6 of arcane advantage, and then say it was only a "hard" encounter if the PC's do get dropped by that nonsense. Anything can look nutso when you look at it in a vacuum at extremes.

Deathtongue-all rants about encounter design aside, what this was actually an argument that Fear was more dangerous than many of the spells and abilities which charm or pseudo-charm PCs with INT or CHA saves. I'm still baffled by the idea that you can honestly tell me that an AOE that typically moves part of the party to safety is an ideal TPK tool, but there are other options. The one area where your point is absolutely correct is that it's AOE, which is one typical difference between INT/CHA and WIS saves (other than mind-flayers, which hit beyond their CR in practice anyway and are an outlier). So while the effect is less dangerous, the fact that it works in an area is more dangerous. I think we can agree that the AOE matters if nothing else. I'd put that as a check mark in making WIS saves more "common" more than more dangerous effect wise, but it still widens the gap in terms of how often they are rolled, and at the end of the day that makes WIS more important than INT or CHA even without accounting for it.

[lall]

Thus, Wis, Cha, Int, Con, Dex, Str is the effect ranking.

This would be my ranking. Not positive of the exact order of the mentals, but I try to be proficient in all three so that it doesn’t matter. Flesh to Stone (CON) is nasty, but you have time to dispel it.

[Wizard_Lizard]

This would be my ranking. Not positive of the exact order of the mentals, but I try to be proficient in all three so that it doesn’t matter. Flesh to Stone (CON) is nasty, but you have time to dispel it.

also lets not forget that if you're a caster.. CON is useful for concentration.

[lall]

True. I personally don’t worry about proficiency in CON saves though as 1. I’d rather be proficient in the mentals and 2. I rarely have concentration spells I intend to use in combat (as Dispel Magic exists).

[prabe]

There seem to be two hierarchies of saving throws, based on whether you're targeting or being the target. It has seemed as though most of the discussion has been on the latter (your saves are being targeted), but I think I've seen some discussion of the former (you're targeting someone else's saves). I wonder how different they are--maybe there are datamongers who can wrangle that answer.

[MaxWilson]

There seem to be two hierarchies of saving throws, based on whether you're targeting or being the target. It has seemed as though most of the discussion has been on the latter (your saves are being targeted), but I think I've seen some discussion of the former (you're targeting someone else's saves). I wonder how different they are--maybe there are datamongers who can wrangle that answer.

Int wins, hands down. Almost no monsters are proficient, and many of them have low Int stat on top of that.

Cha is good against low CR, seriously falls off against CR 10+ IIRC.

Against high CR, Dex is pretty good, and Str is often surprisingly good too even against strong monsters just because proficiency is rare.

Low CR monsters remain relevant, but as groups, so for high-level play you'd want to analyze a mix of AoEs (low CR) and single-target (high CR). I know how to do that analysis but I'm honestly not sure how to show that in a way that will be easy to read and analytically justify. I suppose I could just use the Xanathar tables to choose a random monster appropriate for each level, then crudely assume an AoE hits all of your "share" of that monster (per Xanathar's tables), then do that thousands of times and report the "best" saves you target against e.g. the biggest Xanathar mob allowed, the highest-CR solo, the median mob, and a random-sized mob (uniform probability for each column on Xanathar's table for that level range).

I'll edit this post when I have my results. May take an hour or more.

Edit: here's an interactive web app for viewing the results: https://maxwilson.github.io/Simple-S...owGraphsFor5E/

I intend to let it analyze Xanathar's- and DMG-generated encounters as well as just individual monsters, but my brain needs a rest. There are definitely issues with the web app (the JavaScript file for the app is painfully big and the UX is not terrific, since I stink at CSS) but this should suffice to let you spot some interesting patterns in the data, e.g. I never realized before how weak liches are against Charisma-based attacks like Banishment.

[prabe]

That was quick.

Thanks for the answer, and I look forward to seeing the future results, but I'm not attempting to exert pressure for speed or anything. This is mostly idle curiosity: I figure picking saves to target is only part of picking spells (and I'm not playing a spellcaster with lots of choice in that regard, anyway). Seriously, don't let working on this distract you from anything important to anyone.

[Moreb Benhk]

It seems there is a decent correlation on both fronts then? Aside from DEX and CHA saves at higher levels, it sounds like strong player saves roughly correspond to strong monster saves and vice versa. I guess the play-impact of this is inverse though, as you (the player) decide what saves to target vs encountering the saves at 'random' - it makes sense to invest in weaker saves offensively (by choosing spells that fit these) and stronger saves defensively.

[Deathtongue]

If your PCs are cheesing enemies abound, then you need better PCs. I mean, seriously man, someone intentionally hurting themselves to break an effect they don't know is on them???

What do you mean, don't know is on them? You can still make spell identification checks as a reaction post-Xanathar's. May not help you for the first round of Enemies Abound. And intentionally hurting yourself to break illusions is a well-established action-adventure trope. Don't make me break out the Naruto and Bleach.

The text is clear; everyone is an enemy. If they want to metagame that with AOEs, then they should be targeting the biggest cluster, which means the party is likely suffering AOE as well.

1) That contradicts the text of the spell. Enemies abound has you targeting enemies at random. 2) What's the nature of an enemy that compels you to target the biggest cluster with an AOE? It's an undefined game-term and if we're breaking out RAI/narrative intent, it'd be weird and genre-breaking for someone to focus on Dr. Doom's hated village-destroying minions over Dr. Doom if both of them are in the same room.

The chances that hypnotic pattern or fear actually last until ending normally is very low regardless. Both are concentration, and the standing rule in combat is kill the guy concentrating.

If you can. Spells like Fear and Hypnotic Pattern make that 'oh, just beat on the spellcaster' standing order much harder, because they take out multiple concentration-breakers for a period of time.

The same is true of enemies abound, but in the meantime the players are killing each other.

Enemies Abound targets one person. One. A player is attempting to kill people at random (which may not even put the other players at risk depending on encounter composition) and it's too easily subverted to last for more than a round.

But regardless, Hypnotic pattern keeps you in melee, Fear doesen't and your entire point was that disabling people for the giants is what kills them, so...huh?

Did you forget that a typical party consists of more than one person? Yes, you run away and are 'safe' for 2-4 rounds. Congratulations on you getting yours. What about the rest of your party? What happens to you after the enemies gank them and go after you?

And I've rarely seen fear do anything like what your suggesting; most of the time it's a simple turn waster, some of the time it leads a couple PCs to safety, but basically never does it disable PCs for five turns.

Disabling PCs for five turns is a reasonable but uncommon edge case; for example, if it's being used in a featureless plain or out on the desert. I'd expect two rounds in a tightly-packed dungeon, yes, but having more rounds than that is not out of the question for, say, PCs stuck on a castle parapet or on a frozen-over lake.

Hypnotic pattern rarely does either, but neither have I seen it be easily nullified (how are familiars making their saves? How do party members know to hit allies with AOE to awaken them?)

I'm not a big fan of Hypnotic Pattern. I've been subject to too many 'monster conga lines of awakening' to view that spell as more than a verbal component-free curiosity.

Also, level 7 spells aren't showing up at levels 4-7,

That's generally true, but technically they are in a scary post-Volo's world. Warlock of the Great Old One (14th-level spellcaster), Diviner, Abjurer... not that you need 7th-level spells at that range to cause a TPK. A good roll on a Cone of Cold or a 5th-level Fireball can definitely cause one, and those enemies start showing up at CR 5.

And if the party is facing multiple casters of course it's a bad day. In general, casters are poorly designed as enemies and should be used sparingly, for the same reason you don't design characters based on PC rules. If the DM is throwing two mages at a level 4 party, they lose. If the DM is throwing two mages at a level 6 party, he has no idea how to design encounters, and initiative is going to determine who wins unless the group is optimized.

I'm sympathetic to the idea that Volo's and Mordenkainen's foes are way too hard for typical D&D parties, but it is what it is. The game designers expected monsters to show up at that CR, so we're just going to have to deal.

[Tanarii]

CR5 enemies are just about right for level 8.

[MaxWilson]

I edited my post upthread but also wanted to bump the thread to bring this to folks' attention: https://maxwilson.github.io/Simple-S...owGraphsFor5E/

As I said upthread, I'm still working on encounter construction, but you can at least look at individual monsters, filter for specific sourcebooks and creature types, and spot patterns in the data. I never realized before how weak liches are to Charisma-based attacks, or that the Tarrasque is proficient in Int saving throws.

[prabe]

I edited my post upthread but also wanted to bump the thread to bring this to folks' attention: https://maxwilson.github.io/Simple-S...owGraphsFor5E/

As I said upthread, I'm still working on encounter construction, but you can at least look at individual monsters, filter for specific sourcebooks and creature types, and spot patterns in the data. I never realized before how weak liches are to Charisma-based attacks, or that the Tarrasque is proficient in Int saving throws.

That is quite the thing. My English-major-ish dropout brain reels, trying to comprehend the thinking (extent and kind) that went into it. Most sincere thanks for sharing!

[Corran]

I edited my post upthread but also wanted to bump the thread to bring this to folks' attention: https://maxwilson.github.io/Simple-S...owGraphsFor5E/

As I said upthread, I'm still working on encounter construction, but you can at least look at individual monsters, filter for specific sourcebooks and creature types, and spot patterns in the data. I never realized before how weak liches are to Charisma-based attacks, or that the Tarrasque is proficient in Int saving throws.

Preface: I know nothing about programming. I can only think of how I would like to see your application expand. So please keep that in mind if my suggestions are too much work for their worth, or too much work in general (I suspect that they might be).

1) Have an option to exclude creatures with legendary resistances. I presume this might be easier to do if you make it part of the 'set filters' option? Though it would be nice to present it as a question just like you did with magic resistance. So, something like ''Are your attacks affected by legendary resistance?'', followed by ''Yes. Exclude legendary creatures from the results.'' and ''Include legendary creatures in the results and count them as 0% effectiveness.'', and ''No, my attacks are not affected by legendary resistance''. Or something of that sort.

2) The next BIG step would be to account for condition resistances and immunities. For example: without accounting for the effect and the cost involved, how much upside should I expect if I choose a wis-targeting untyped effect (eg confusion) over a wis-targeting typed (eg charm) effect (eg hypnotic pattern)? Sure, the results wont say the whole story, but they would be very helpful in making an informed decision, particularly if I don't know what adventure I will be playing (cause with a smaller sample, the comparison might not be beneficial to begin with, while with a very large sample it is easier to justify limiting your choice to the best one overall; or at least trying to figure out combinations that complement each other optimally). This is probably a nightmare to produce in a single diagram, so I am thinking that the best way to do it would be by adding additional questions that would produce results for a single spell (in a general context, ie cha charm effect, or con restrain effect, or int untyped effect, etc; and this actually could be helpful for less experienced DM's who want to balance homebrew spells, and I am cunningly mentioning this because I think it might spark your interest). Then you run the app as many times as you want, and you compare the different diagrams produced.

ps: I am guessing the option to choose between single target and AoE, and the option to adjust the AoE's effectiveness, are there for the encounter construction part that you are working on, right? It sounds very promising. We could certainly have a few games with this once you complete it. Eg, someone starts a thread and declares a challenge, of which we don't know all the details. For example, we could know the number of creatures involved, and their respective CR's, and we know what level of caster we are supposed to be playing. So other people post their caster builds. Then the encounter is revealed, and we try to determine which poster has the best chance of contributing more to their group's efforts.

pps: I'd like to help you if I can. I don't think I'll be of any help with the programming stuff, but I could certainly do some of the boring job (make this or that list of monsters) should you need someone to.

[MaxWilson]

Preface: I know nothing about programming. I can only think of how I would like to see your application expand. So please keep that in mind if my suggestions are too much work for their worth, or too much work in general (I suspect that they might be).

1) Have an option to exclude creatures with legendary resistances. I presume this might be easier to do if you make it part of the 'set filters' option? Though it would be nice to present it as a question just like you did with magic resistance. So, something like ''Are your attacks affected by legendary resistance?'', followed by ''Yes. Exclude legendary creatures from the results.'' and ''Include legendary creatures in the results.'', or something of that sort.

Would you really want to exclude them, or would you like them factored into the effectiveness percentage somehow? Technically it's easy to determine whether/how many legendary resistances a given creature has--I just can't think of any useful questions to ask/answer with that data. The way I think of it right now is that "effectiveness" has to hit at least 4.0 before you can count on affecting any creatures with legendary resistances, so even if you choose a 100% effective spell (like Mental Prison against an Astral Dreadnought) it will of course always take 4 spells to affect it.

I can't think of any way to include legendary resistance in the picture that wouldn't make it more confusing, but maybe there's a way. Edit: I guess you could define Effectiveness as (average % failed saves) / (Legendary Resistances + 1).

BTW, "are your attacks affected by legendary resistance" is a redundant question. All saves are affected by legendary resistance, and no ability checks are. But you could ask whether you want to see effectiveness scaled down for LR... hmmm, I think I'll do that right now. One sec.... Edit2: [45 minutes later] Done.

2) The next BIG step would be to account for condition resistances and immunities. For example: without accounting for the effect and the cost involved, how much upside should I expect if I choose a wis-targeting untyped effect (eg confusion) over a wis-targeting typed (eg charm) effect (eg hypnotic pattern)?

I actually have this raw data already (including which creatures have advantage against specific conditions like being poisoned or charmed). What would be an appropriate UX? Maybe a selector asking you to optionally choose a type of effect from a list of conditions?

I also have raw data on damage resistances/immunities/etc., and again I don't know how to do anything useful with it, in the current UI. Hmmm, I guess I could make "damage: <dice> of <type> for <dice amount> save for <half or none>" an option for you to select when you choose a spell, instead of picking a condition. And in that case, "Effectiveness" can be total HP of damage inflicted, instead of average % of failed saves.

Sure, the results wont say the whole story, but they would be very helpful in making an informed decision, particularly if I don't know what adventure I will be playing (cause with a smaller sample, the comparison might not be beneficial to begin with, while with a very large sample it is easier to justify limiting your choice to the best one overall; or at least trying to figure out combinations that complement each other optimally). This is probably a nightmare to produce in a single diagram, so I am thinking that the best way to do it would be by adding additional questions that would produce results for a single spell (in a general context, ie cha charm effect, or con restrain effect, or int untyped effect, etc; and this actually could be helpful for less experienced DM's who want to balance homebrew spells, and I am cunningly mentioning this because I think it might spark your interest). Then you run the app as many times as you want, and you compare the different diagrams produced.

Whoops, I should have finished reading.

Would it be helpful to show multiple graphs at the same time?

ps: I am guessing the option to choose between single target and AoE, and the option to adjust the AoE's effectiveness, are there for the encounter construction part that you are working on, right?

Correct. I've now removed them from the "Monster (by CR)" path--there's no reason to ask about AoE in that case, it's irrelevant.

It sounds very promising. We could certainly have a few games with this once you complete it. Eg, someone starts a thread and declares a challenge, of which we don't know all the details. For example, we could know the number of creatures involved, and their respective CR's, and we know what level of caster we are supposed to be playing. So other people posts their caster builds. Then the encounter is revealed, and we try to determine which poster has the best chance of contributing more to their group's efforts.

pps: I'd like to help you if I can. I don't think I'll be of any help with the programming stuff, but I could certainly do some of the boring job (make this or that list of monsters) should you need someone to.

The most helpful stuff for me is actually the design work: what should it ask you, how should it behave, what should it show, how to avoid clutter, etc. I might also ask for some help with data entry at some point, although for right now I have all the data I need for to answer questions about how best to attack monsters. I don't have complete data on monster spells though or monster attacks/damage--in order to answer questions about how best to defend against monsters I'd need more data entry.

However, if you're interested in collaborating we should probably start a separate thread to avoid cluttering up this one. Just post a link in this thread.

I don't promise to do a ton of work on this project (I find pure analysis less interesting than combat simulations) but the suggestions you've made so far are low-hanging fruit that would be pretty easy to do.

[Corran]

However, if you're interested in collaborating we should probably start a separate thread to avoid cluttering up this one. Just post a link in this thread.

Not sure how helpful that would be to you, but my curiosity is piqued, so I made this thread.

[MaxWilson]

I made a couple of small enhancements to the saving throw analyzer: https://maxwilson.github.io/Simple-S...owGraphsFor5E/

(1) You can now analyze randomly-generated encounters as well as just monsters.

(2) You can now set tags like "Fire Damage, Save for Half" or "Charmed" on your attacks to see how that changes overall effectiveness, since some monsters are vulnerable, resistant or immune to certain attacks.

For example, if I tell it to generate a bunch of undead-only encounters using Xanathar's rules, a typical Wisdom save graph will look something like this:

Spoiler: Image: Undead Wisdom Saves

Show

But if I'm really interested in casting Hypnotic Pattern, then I want to make sure undead that are immune to charm are counted as immune, so I hit the Effect tags button and tag it as "charmed". You will see the purple line drop substantially.

Spoiler: Image: Undead Wisdom Charm Saves

Show

If I want to compare it to Fireball instead I can switch to looking at Dex with effect tags "fire" and "save for half".

Limitations: the UI is still kind of ugly. Sorry. I'm not a design expert, and I'm focused more on functionality than on getting all of the text to line up prettily on each line.

Furthermore, for technical reasons, effect tags are not available on the overview screen--you have to drill down on a specific attribute (Str/Dex/Con/Int/Wis/Cha) in order to apply specific effects. I don't intend to fix this because this app has grown to a point where it's unwieldy and difficult to extend further, and because you should normally know what attribute you're interested in if you're applying a specific effect like "fire, save for half" (probably Dex).