Rounds to Balance

[Chronos]

TLDR: Balance favors martials at first level, is close to even at 5th level, and hugely favors casters at 11th and 20th level.

Premise:
The primary way that D&D characters interact with the world mechanically is by dealing damage, and so the total amount of damage characters deal over the course of an adventuring day is a plausible way of comparing mechanical balance. In any given round, some characters (especially spellcasters) can significantly increase their damage output by expending resources, which is theoretically balanced by the non-caster classes dealing more damage when they're not expending resources. Hence, in a sufficiently-short adventuring day, when casters can expend resources in most or all of the rounds, balance will tend to favor casters, while in a sufficiently-long adventuring day, casters will be unable to spend resources in most rounds, and so the higher resourceless damage of the non-casters will be relevant, and balance will favor the non-casters. Hence the concept of this thread: "rounds to balance", the number of rounds of combat per adventuring day for different classes to deal the same amount of damage. Here, I'll be looking at four characters, a single-classed fighter, rogue, wizard, and cleric (other classes could be included, but I think this is enough for a start).

General assumptions:

*The party will get two short rests per adventuring day, and hence short-rest resources can be considered as long-rest resources by multiplying their number of uses by three.

*Since we're looking at long adventuring days, all party members are assumed to have used up all expendable resources by the time they take the appropriate sort of rest. This might still mean some resourceless rounds early on, so long as everything is used up eventually: In what order it's used up doesn't matter.

*Non-damaging contributions to a combat aren't included, for the sake of simplicity. If they were, this would favor casters, since casters have many more non-damaging options than martials do.

*Flexibility in when damage is dealt is also not considered. This would also favor casters, since when presented with an easy combat, casters can choose to conserve resources for more difficult fights, or go nova for more difficult fights, an option mostly not available to martials.

*Damage types (for resistances, immunities, or vulnerabilities) are also not considered. This would also favor casters, who have many more options for varying their damage type.

*Tactics that synergistically involve multiple party members, such as a battlemaster using Commander's Strike on a rogue or a wizard casting Haste on a fighter, are not considered for the sake of simplicity, since it is difficult to tell who to credit with the damage.

*Effects that deal damage over multiple rounds are assumed to last for three rounds, that being the typical length of a combat.

*Overkill is ignored, because I have no good way to model it.

*Comparisons will be made at 1st level, 5th level, 11th level, and 20th level. I would have gone with the starts of the tiers, but I wanted to give the fighter his fourth attack.

*Enemies will be assumed to have the average AC and saves (rounded to the nearest integer) of monsters in the Monster Manual with CR matching the party's level (data from a thread by Platypusbill (https://forums.giantitp.com/showthread.php?562144-D-amp-D-5e-MM-Monster-Defences-(AC-vs-Saving-Throws)))

*Since the rulebooks claim that the game plays just fine without magic items, martials will be assumed to not have magic weapons (except in cases where the class itself grants it, which doesn't occur in the four classes I'm looking at).

*Since I'm measuring by damage dealt, I will assume that all classes are attempting to maximize their damage.

*All racial abilities will be ignored, beyond the assumption that all classes can start with a 16 in their primary stat.

*Only core (PHB, DMG, MM) options will be considered, for simplicity.


Fighter-specific assumptions:

*At first level, the fighter has a 16 Str and wields a greatsword. Opportunity attacks are assumed to be negligible at this level.

*By fifth level, the fighter has specialized to battlemaster, taken Polearm Master and switched to using a halberd, for two d10 attacks and a d4 attack per round. With the additional options from Polearm Master, I'll assume that he's also able to make an attack as a reaction 50% of the time.

*The greatest damage from a superiority die is from Precision Strike used to turn a miss into a hit. I will assume that the opportunity for this happens sufficiently many times, and that the fighter can recognize when they are, so on the rounds where the fighter uses a superiority die, his damage is his resourceless damage plus the damage of one hit.

*By 11th level, the fighter has maxed out his strength and also taken Great Weapon Master. At average monster AC, it is worthwhile to use the -5/+10 option, so I'll assume that he's always using that. However, I will not be counting the bonus action attack option for a crit or downing an enemy, since it's difficult to model how often that'll happen, and even when it does, it's replacing the polearm master bonus attack, for a net increase of only 3 damage (adjusted for accuracy).

*By 20th level, the fighter has also taken Martial Adept, to gain an additional superiority die. All other ASIs are assumed to be spent on feats with no or negligible impact on damage.


Rogue-specific assumptions:

*The rogue is using a light crossbow from range. I could have assumed dual-wielding, but with the bonus action spent on that, it's difficult to tell when the rogue has advantage or otherwise qualifies for sneak attack, and dual-wielding is only slightly higher damage than single-wielding with advantage.

*At first level, the rogue gets advantage (and hence also Sneak Attack) on one round out of three from being hidden at the start of a three-round combat, and then gets Sneak Attack (but not advantage) on one of the other two rounds, from ally positioning. I'm not sure how good this assumption is, but I have to assume something.

*By fifth level, the rogue has specialized to thief. This is an exception to my assumption that everyone is building to maximize damage, simply because it's too difficult to model how often the extra damage from assassin will apply.

*By fifth level, the rogue is using her bonus action every turn to successfully hide, and hence always has advantage (and Sneak Attack). Her Dex is now 18.

*By 11th level, the rogue's Dex is 20. All further ASIs are spent on feats with no or negligible impact on damage.


Wizard-specific assumptions:

*When not using resources, the wizard uses Chill Touch, for 1d8, 2d8, 3d8, or 4d8 damage of a reliable type. The riders are assumed to be irrelevant.

*At first level, the wizard uses her spell slots on Sleep. It doesn't technically deal damage, but it takes enemies out of the fight as measured by HP, which is close enough.

*The wizard chooses a subclass that doesn't have any significant improvement to her damage output. Core, this would only be Evoker, and some of the Evoker damage boosts are difficult to calculate.

*By fifth level, 1st-level spell slots are assumed to contribute negligible damage (she might be using them for Shield or other utility purposes), so the wizard is considered out of resources when she doesn't have any slots of 2nd level or higher.

*By fifth level, the wizard has 18 Int, and by 11th, she has 20. All further ASIs are spent on feats with no or negligible impact on damage.

*Second-level spell slots are used on Scorching Ray. Third through fifth level spell slots are used on Fireballs and/or Lightning Bolts, each of which is assumed to hit four enemies.

*Arcane Recovery is always used to recover as many Fireballs as possible.

*Sixth, seventh, and eighth-level spell slots are used on Sunbeam. Over a three-round duration (during which the wizard casts no further spells), it hits an average of three creatures per round (four on the first, three on the second, and two on the third, as the number of targets available diminishes).

*The ninth-level slot is used on Meteor Swarm. Due to its extreme range and independently-targeted spheres, it is assumed to hit 8 creatures.


Cleric-specific assumptions:

*On rounds when not using resources, the cleric uses Sacred Flame.

*At first level, the cleric has 16 Wis, and uses all of his spell slots on Guiding Bolt. The rider is assumed to be irrelevant (both because of my no-teamwork assumption, and because at this level, most monsters will die from a Guiding Bolt before the rider becomes relevant).

*The cleric is assumed to be a subclass that gets Potent Cantrips, but any other damage boost from subclass is assumed to be negligible. Channel Divinity features, in particular, are not considered.

*By fifth level, the cleric has 18 Wis, and by 11th, he has 20. All further ASIs are spent on feats with no or negligible impact on damage.

*By fifth level, 1st-level slots are assumed to contribute negligible damage (he might be using them on Healing Word or other utility purposes), and so the cleric is considered out of resources when he runs out of slots of 2nd level or higher.

*Second-level spell slots are used on Spiritual Weapon. Over a three-round duration, each round he attacks once with the Spiritual Weapon and casts Sacred Flame.

*Third and higher level spell slots are used on Spirit Guardians. Over a three-round duration, it hits an average of three creatures per round (four on the first, three on the second, and two on the third, as the number of targets available diminishes). Since maintaining the spell does not require an action, on the second and third rounds, he also casts Sacred Flame.


Spreadsheet: https://docs.google.com/spreadsheets/d/1WLI_s4EsfbL3byx-Q9s_pUxtzRT-zuNI2_KqnbFBDwE/edit?usp=sharing

(Note: I probably made some mistakes here. Part of the reason I did this in a spreadsheet is so that, once caught, any mistake fixes would propagate through the calculations.)

Summay of results:

At 1st level, neither the rogue nor the fighter has any expendable resources, and so they balance after zero rounds. At 11th and 20th level, the fighter does more at-will damage than the rogue, and can expend resources to deal even more yet, while the rogue never gets any expendable resources. Thus, there is no number of rounds after which the rogue's damage balances the fighter's.

Similarly, at levels 1 and 5, the cleric lags behind the wizard in both at-will and resource damage, and so there is no number of rounds at which the cleric's damage balances the wizard's. However, once the cleric gets Potent Spellcasting, his at-will damage is greater than the wizard's, so he can eventually (after a very long time) catch up.

At level 1, the Sleep spell is ludicrously powerful: The rogue and fighter would need 4-6 three-round combats per day to catch up to the effectiveness of the wizard's three Sleep spells. However, a single combat would suffice for the martials to catch up to the cleric.

At 5th level, it is theoretically possible, but only just, for the rogue to catch up to the fighter. Rounds to balance between the martials and the casters, however, range from 16 to 25. Assuming that parties are expected to be able to handle roughly seven combats per day, this is close to being balanced.

At higher levels, even the best-case comparison between martials and casters, the level 11 fighter vs. the cleric, would still require 40 combat rounds per day before the fighter can catch up. I think it's safe to say that very, very few parties will ever see this many combat rounds in a day, and so the casters are far, far overpowered by this point.

[Waazraath]

To start with: interesting, thanks for the effort!

I don't have time to delve deep into the math, but based on your post (mainly the assumptions) a few comments. I'll quote you using italics instead of quote-blocks, to keep it a bit readable:

*Since we're looking at long adventuring days, all party members are assumed to have used up all expendable resources by the time they take the appropriate sort of rest. This might still mean some resourceless rounds early on, so long as everything is used up eventually: In what order it's used up doesn't matter.
--> I question how realistic this is. In a regular adventuring day, long rest dependent classes (and to a lesser extend short rest dependent classes as well) wil conserve resources, in case there will be a (though) fight ahead. This is dependent on playstyle, but assuming every resource has been used up at the end of an adventuring day introduces a strong bias towards long-rest dependent classes at the cost of others, mostly no-resource using classes. I'd say some part of resrources should be considered unused at the average adventuring day, based on experience somewhere in the 10%-30% bandwidth (and maybe using a different % for short rest and long rest resources - something like 10% SR unused and 20% long rest unused).

*Non-damaging contributions to a combat aren't included, for the sake of simplicity. If they were, this would favor casters, since casters have many more non-damaging options than martials do.
--> I don't think it would, because that would drastically lower the exact metric you are using here (rounds/damage) cause casters would do less damage in this case - given that the expendable resource they use for damage is the same they would use for non-damage.

*Flexibility in when damage is dealt is also not considered. This would also favor casters, since when presented with an easy combat, casters can choose to conserve resources for more difficult fights, or go nova for more difficult fights, an option mostly not available to martials.
--> a bit of a generalization, this goes for a lot of martials (using, ki, maneuvers, rage, action surge) as well. And the more resources are conserved, the bigger the chance that some are left at the end of the adventuring day (see my first point).

*Since the rulebooks claim that the game plays just fine without magic items, martials will be assumed to not have magic weapons (except in cases where the class itself grants it, which doesn't occur in the four classes I'm looking at).
--> fair for calculation; on the one hand, I could argue that the DM has a table with more or less assumed items at different levels (somewhere around p35 I think), on the other hand, that would also require giving appropriate rods staves and wands to casters which would make calculations more difficult and I don't know how much it would change in the end.

As for the rogue: why not assuming advantage with that tasha's ability, for the use of bonus actions, and use one of the tasha's classes (phantom or soulknife) to calculate damage instead of thief? It would lead to a higher max damage (as is intendent) without being that much more difficult to calculate (I think).

*At first level, the wizard uses her spell slots on Sleep. It doesn't technically deal damage, but it takes enemies out of the fight as measured by HP, which is close enough.
--> this seems very generous to me; if I look at the spreadsheet, am I correct this assumes to take out 3 monsters with an average of 7.5 hp? In my experience it's quite rare for a sleep to be this effective, and even on the rare cases where it is, there are often other enemies who spend a round to wake the sleeping target, leading to sleep being in practice more a delaying & damage reducing tactic than really taking out monsters out of the fight.

*Second-level spell slots are used on Scorching Ray. Third through fifth level spell slots are used on Fireballs and/or Lightning Bolts, each of which is assumed to hit four enemies.
--> again, this is really generous. In my experience, 2.5 creatures per fireball is more realistic. As for Lightning bolts, I don't think I've seen a player damage 4 critters with 1 bolt in several decades of D&D.

*The ninth-level slot is used on Meteor Swarm. Due to its extreme range and independently-targeted spheres, it is assumed to hit 8 creatures.
--> again, very generous. Of course, radius is big enough theoretically. But if I go through my published adventure books at high levels there are many fights that don't even include 8 enemies, let alone having them all placed so that they would be hit.

*On rounds when not using resources, the cleric uses Sacred Flame.
--> when optimizing for damage, why not a melee cleric (that's at least half of them) with a greatsword or the like, and later on a SCAG cantrip?

*Third and higher level spell slots are used on Spirit Guardians. Over a three-round duration, it hits an average of three creatures per round (four on the first, three on the second, and two on the third, as the number of targets available diminishes). Since maintaining the spell does not require an action, on the second and third rounds, he also casts Sacred Flame.
--> I'd probably put it at 2.5 creatures. The calculation doesn't seem to include the possiblity that concentration gets disrupted, is that correct?

My conclusion: as I said, interesting. Based on my own gaming experience there is a serious positive bias though towards the wizard due to the very generous way damage is calculated at every level (sleep, FB/LB and Meteor Shower), and a negative bias towards the rogue because pretty easy damage upgrades aren't included.

[Chronos]

I didn't include any of the options from Tasha's or SCAG or other sources, because I was keeping this analysis core-only, for simplicity. In any event, core does include a way for a rogue to use a bonus action to generate advantage, at all levels past first, and I was assuming that the rogue was using it.

I question how realistic this is. In a regular adventuring day, long rest dependent classes (and to a lesser extend short rest dependent classes as well) wil conserve resources, in case there will be a (though) fight ahead.
But that's the point: Casters have that option, while non-casters don't. Having options never makes a class weaker. They can, if they want, use up all of their resources.

I don't think it would, because that would drastically lower the exact metric you are using here (rounds/damage) cause casters would do less damage in this case - given that the expendable resource they use for damage is the same they would use for non-damage.
Again, it comes down to options: If a non-damaging spell will be more effective than a damaging spell, the caster will use the non-damaging spell. If a non-damaging spell would be less effective than a damaging spell, the caster will use the damaging spell. Having non-damaging options means that sometimes the caster will be more effective, but never means that the caster would be less effective.

a bit of a generalization, this goes for a lot of martials (using, ki, maneuvers, rage, action surge) as well.
Most classes (other than rogue) do have some expendable resources, true. And in fact I chose a fighter subclass with more expendable resources than usual for this analysis. It's still far less relevant for them than it is for casters.

this seems very generous to me; if I look at the spreadsheet, am I correct this assumes to take out 3 monsters with an average of 7.5 hp?
It assumes that it takes out 5d8 HP worth of monsters, because that's what the spell says. It could be one large creature, a few middling creatures, or a whole swarm of creatures with 1 or 2 HP each.

The number of targets per area-effect spell might be a valid criticism. My group doesn't use Fireball very often, so I don't have much data for typical cases for that. We do use Spirit Guardians a lot, and my figures roughly match my experience there, but then, I also don't know if our results are typical.

Oh, and I assumed a cleric using cantrips instead of a melee weapon (with or without Booming Blade), because I didn't want to have to worry about more than one ability score per character.

[Waazraath]

I didn't include any of the options from Tasha's or SCAG or other sources, because I was keeping this analysis core-only, for simplicity. In any event, core does include a way for a rogue to use a bonus action to generate advantage, at all levels past first, and I was assuming that the rogue was using it.

But that's the point: Casters have that option, while non-casters don't. Having options never makes a class weaker. They can, if they want, use up all of their resources.

Again, it comes down to options: If a non-damaging spell will be more effective than a damaging spell, the caster will use the non-damaging spell. If a non-damaging spell would be less effective than a damaging spell, the caster will use the damaging spell. Having non-damaging options means that sometimes the caster will be more effective, but never means that the caster would be less effective.

Most classes (other than rogue) do have some expendable resources, true. And in fact I chose a fighter subclass with more expendable resources than usual for this analysis. It's still far less relevant for them than it is for casters.

It assumes that it takes out 5d8 HP worth of monsters, because that's what the spell says. It could be one large creature, a few middling creatures, or a whole swarm of creatures with 1 or 2 HP each.

The number of targets per area-effect spell might be a valid criticism. My group doesn't use Fireball very often, so I don't have much data for typical cases for that. We do use Spirit Guardians a lot, and my figures roughly match my experience there, but then, I also don't know if our results are typical.

Oh, and I assumed a cleric using cantrips instead of a melee weapon (with or without Booming Blade), because I didn't want to have to worry about more than one ability score per character.

- regarding the rogue: fair enough.
- regarding that casters have the option to spent everything: theoretically correct, but that means that an underlying assumption is that either a) casters know exactly when to spend their spells or b) casters spent all their resources as soon as possible. The first one isn't realistic, the second one is bad tactics (for instance, the boss fight where the caster wants its big guns available is usually at the end). "More options is always better" is in this case only true because the bad tactics lead in this case to a better outcome in this models, given its specifications. If the purpose of a model is to simulate reality (which it usually is, and I think is the case with this one) the model you made will be imo be better if you build in the (reallity grounded) assumption that a number of resources won't be spent at the end of the day.
- as for sleep: ah, understand how you calculated this now. But this remains extremely generous imo. If there is one large enemy with 5d8 hp, there's an aprox 50% chance that sleep will do nothing, when rolling below average. In addition to the fact that the enemies aren't really taken out of the game but are often awakened 1 or 2 rounds later, the damage attributed to the spell remains far too high.
- as for the other AoE spells: am curious to experiences of other playgrounders here, I'm also only being N=1
- as for the melee cleric: I understand the choice, but it does lead to a small underestimation of damage imo (which might be compensated by a slightly optimistic amount of damage from Spirit Guardians, so oh well).

[Anymage]

One guy doing tons of single target damage while another guy is an aoe specialist isn't unbalanced, even if the aoe guy has a much higher damage cap. Doubly so because of the realities of many fights; a third level slot to counter an enemy fireball is probably a very good tactical choice even if it does result in less damage done over the day than just casting that fireball yourself.

The issue is that casters have tactical options (the rogue doesn't have the same breadth of options to counter/dispel enemy magics, to inflict statuses, or reshape the battlefield), as well as the fact that caster options just work in noncombat situations while nonmagical options have a much fuzzier resolution system on top of slower scaling and the uncertainty of the d20's flat probability curve. Muggles need options in those areas a lot more than they need more dpr added.

[Hael]

Interesting analysis.

But I think this analysis still amusingly favors martials in the sense that equalizing dpr isn’t ultimately that relevant, and it’s not on this aspect where the disparity is at its most stark.

What should be the criteria, is the relative win probability per resource/action used. In other words, how many ways can a fight go wrong after an action is taken. When the fighter hits one of two ogres hard with a stick, he/she has a chance to kill it, effectively leaving only half of the threat left. Thats obviously a good thing, and his dpr number and relative win percentage after an action will show it as such. Note that there are still many ways the fight can go poorly.. All the way up till the end.

However, when the wizard uses force cage, the fight is over with 100% probability on round1. At that point, no more resources are necessary. So even though it might take many rounds of poking the thing through a slit to do the job the fighter did many rounds earlier, the fight was a guarenteed win on the very first turn.

That enormous win probability swing (particularly in the first few rounds of combat) is basically everything in DnD combat. The tail end of long adventuring days, or the later rounds of combat (where the wizard doesn’t cast) are basically a miniscule fraction of the total relative win percentage. (Eg a slightly stronger axe for mop up on round 4, is usually like the difference between a 99.99% win chance and a 99.999%).

[Chronos]

Quoth Hael:

But I think this analysis still amusingly favors martials in the sense that equalizing dpr isn’t ultimately that relevant, and it’s not on this aspect where the disparity is at its most stark.
That was kind of my purpose in doing this analysis: Even if we throw the fighters a bone by focusing on their specialty, casters still eat their lunch. A caster who chooses to never do anything but maximize their damage is deliberately handicapping themselves, but even with that handicap, it's still almost impossible for the fighters to catch up (as opposed to most of the non-damage things that casters can do, where you can cross out the "almost", because fighters can't do them at all).

[Waazraath]

That was kind of my purpose in doing this analysis: Even if we throw the fighters a bone by focusing on their specialty, casters still eat their lunch. A caster who chooses to never do anything but maximize their damage is deliberately handicapping themselves, but even with that handicap, it's still almost impossible for the fighters to catch up (as opposed to most of the non-damage things that casters can do, where you can cross out the "almost", because fighters can't do them at all).

Ah. That explains both the obvious bias in favour of casters in the analysis, as well as the reluctance to adress it. If you need a skewed white room analysis to prove a point, you're probably not making the point you think you are making though.

[Chronos]

I'm confused... the analysis was biased in favor of the martials, not the casters, and the conclusion was that, even with that bias, the casters still came out far ahead. If I take out the bias, the casters will be even further ahead.

[Waazraath]

I'm confused... the analysis was biased in favor of the martials, not the casters, and the conclusion was that, even with that bias, the casters still came out far ahead. If I take out the bias, the casters will be even further ahead.

No the analysis and the assumptions are heavily favored towards casters. As I've argued in the post above. The fact that you state the contrary makes that your post reads in all honesty as an analysis designed towards an outcome (to prove a point), instead of an objective analysis to discover something.

I mentioned some things already in my posts above, but to be more thorough:
- Sleep is assumed to take enemies out of the fight, while in fact it's often just a speed bump until another enemy awakens the target;
- Sleep does 0 with bad role, which isn't taken into account in the analysis;
- Casting Sleep on one big critter increases the chance of it doing 0, casting it on several smaller ones requirers them to be conveniently together - assuming you will be able to target the average of 5d8 with every casting is irrealistic (and thus bad modelling).
- All of this leads to overestimating what Sleep does in a real game enormously, 22.5 hp of enemies 'taken out' is in my experience 2 to 3 times more than should be expected;
- Fireball / Ligtning Bolt are assumed to hit far more targets than realistic (should be 2.5 on average at most)
- Sunbeam hitting 3 creatures per turn on average is likewise an almost bizarre overestimation. 3 creatures on average every round with a 5 ft line? Come on...
- Sunbeam is concentration, the possiblity of it failing isn't modelled
- Meteor Shower hitting average 8 creatures is an increadible overestimation as well
- While understandable from a modelling point of view and limiting yourself to the PHB, taking a rogue subclass without any damage increases paints an irrealistic picture if you compare it with 3 max damage subclasses;
- The cleric's Spirit Guardians has the same problem as Sunbeam (concentration, and requiers to be in melee range to boot)
- as mentioned in my first post, assuming all resources spent is irrealistic and a bias in favor of long rest users
- additionally, the elephant in the analysis which has not been mentioned yet is that this only looks at offense. Defense is at least as important (staying alive, staying alive), and where the Fighter, (to a lesser extent) the Rogue have those resource-less build into their class (hp, armor, and class features like uncanny dodge, evasion, second wind), especially wizards are depending on exactly the same long rest resource they are using for their offense. Resources MUST be spent on these to stay alive, and a % of slots should be allocated for this for a realistic assessment. The cleric could arguably do without (though assuming some slots being spent on Aid or the like to keep up would not be a stretch).
- the same goes for utility. The rogue has unlimited uses of it's expertise and reliable talent, where a wizard that actually does something with utility will automatically do less damage. So "Non-damaging contributions to a combat aren't included, for the sake of simplicity. If they were, this would favor casters, since casters have many more non-damaging options than martials do." this is simply not true, taking this into account would cost the wizard (and cleric) in damage and would cost the rogue nothing (and the BM fighter short rest resources, if the right maneuvers are taken).

All of this leads to a very biased outcome, underselling the rogue and with the wizard doing, at least 200% more damage than in a (more realistic) model that doesn't grossly overestimate area spells, doesn't disregard concentration and doesn't assume all resources spent. (And that's even disregarding the fact that some classes get utility and defenses for free while other classes have to pay for them with spell slots!) Yeah, the wizard come out on top that way, quelle surprise (not).

[LudicSavant]

- Fireball / Ligtning Bolt are assumed to hit far more targets than realistic (should be 2.5 on average at most)

I can't even remember the last time I cast a Fireball on just two targets. :smallconfused:

[Rukelnikov]

I can't even remember the last time I cast a Fireball on just two targets. :smallconfused:

Depending on the level, and the prospects of further combats, I've ocassionally used it on single enemies. At lvl 5 its not only the best AoE damaging spell but also often the best single target damaging spell.

[LudicSavant]

Depending on the level, and the prospects of further combats, I've ocassionally used it on single enemies. At lvl 5 its not only the best AoE damaging spell but also the best single target damaging spell.

Fireball is frequently not the best option for dealing with single targets. Especially not on damage caster builds.

[Rukelnikov]

Fireball is frequently not the best option for dealing with single targets. Especially not on damage caster builds.

What would you suggest for damage? Scorching Ray? MM?

[LudicSavant]

What would you suggest for damage? Scorching Ray? MM?

Really depends on the party, build, and enemy matchups.

If you want to use Scorching Ray, something like a Bugbear Blaster can do nice things with that (since as of MPMM, they can append bonus damage to each of those shots).
If you want to use Magic Missile, Hexvokers are well known for their prowess there. Also I think it's just a generally efficient spell in a 1st or 2nd level slot.

But there are so, so many options. Examples include minionmancy, bouncing people around hazards, gishery, indirect damage, control, concentration spells that hit multiple times per slot, etc. I rarely approach two combats the same way, because my goal as a versatile caster is generally to utilize my versatility to exploit the weaknesses of whatever matchup spread I'm fighting that day. So I might spend one fight prevent a Beholder from using any of its eye rays because Fog Cloud is hilariously good against them, another I might be buzzsawing a magic resistant legendary creature by punting them back and forth through a Cloud of Daggers, another I might be fireballing a group, etc.

Don't get me wrong, there are situations where I'd hypothetically cast a Fireball on 1-2 targets. I just can't remember the last time I did so, which made me raise an eyebrow at the notion that 2.5 is an "at most" case.

[Rukelnikov]

Really depends on the party, build, and enemy matchups.

If you want to use Scorching Ray, something like a Bugbear Blaster can do nice things with that (since as of MPMM, they can append bonus damage to each of those shots).
If you want to use Magic Missile, Hexvokers are well known for their prowess there.

But there are so, so many options. Examples include minionmancy, bouncing people around hazards, gishery, indirect damage, control, concentration spells that hit multiple times per slot, etc. I rarely approach two combats the same way, because my goal as a versatile caster is generally to utilize my versatility to exploit the weaknesses of whatever matchup spread I'm fighting that day. So I might spend one fight prevent a Beholder from using any of its eye rays because Fog Cloud is hilariously good against them, another I might be buzzsawing a magic resistant legendary creature by punting them back and forth through a Cloud of Daggers, another I might be fireballing a group, etc.

I know the versatility of spellcasting, and while strictly dealing damage is usually not the most effective form of contribution a wizard (or caster in general) can bring to a party, if you want to deal damage, Fireball is often your best bet at 5th level, most wizards aren't hexvokers, or bugbears on the first round of combat, and without a source outside the Wizard class, Fireball is the best single target damage as long as the target isn't resistant/immune to fire and it doesn't make its save more than 75% of the time (to beat out 3rd lvl MM), or has a weird low AC decent Dex save.


Don't get me wrong, there are situations where I'd hypothetically cast a Fireball on 1-2 targets. I just can't remember the last time I did so, which made me raise an eyebrow at the notion that 2.5 is an "at most" case.

Oh, I don't think 2.5 should be considered an "at most" scenario at all, usually I see Fireballs if there's at least 3 enemies to be caught.

[LudicSavant]

I know the versatility of spellcasting, and while strictly dealing damage is usually not the most effective form of contribution a wizard (or caster in general) can bring to a party, if you want to deal damage, Fireball is often your best bet at 5th level, most wizards aren't hexvokers, or bugbears on the first round of combat, and without a source outside the Wizard class, Fireball is the best single target damage as long as the target isn't resistant/immune to fire and it doesn't make its save more than 75% of the time (to beat out 3rd lvl MM), or has a weird low AC decent Dex save.

You don't have to be a bugbear or a hexvoker, I only mentioned them because you specifically asked about Magic Missile and Scorching Ray.

It's not that special to be able to get more single target damage out of a 3rd level slot than Fireball. You can do it with Melf's Minute Meteors, Cloud of Daggers, Summon Whatever, lots of options. And there's always non-damage options, too.

At least in my own playstyle, if I'm a Wizard that is actually trying to do single target damage, they'll have something better than Fireball for that. And if they're not trying to be a single target damage caster, they'll probably have some way other than damage to deal with single targets efficiently. In either case, I don't generally end up casting Fireball at single targets.

[Rukelnikov]

You don't have to be a bugbear or a hexvoker, I only mentioned them because you specifically asked about Magic Missile and Scorching Ray.

It's not that special to be able to get more single target damage out of a 3rd level slot than Fireball. You can do it with Melf's Minute Meteors, Cloud of Daggers, Summon Whatever, lots of options. And there's always non-damage options, too.

At least in my own playstyle, if I'm a Wizard that is actually trying to do single target damage, they'll have something better than Fireball for that. And if they're not trying to be a single target damage caster, they'll probably have some way other than damage to deal with single targets efficiently. In either case, I don't generally end up casting Fireball at single targets.

I understand that, and I don't think its a particularly good strategy either, but more than once I've glanced at my memorized spells, and ended up Fireballing a single powerful enemy.

[LudicSavant]

Oh, I don't think 2.5 should be considered an "at most" scenario at all, usually I see Fireballs if there's at least 3 enemies to be caught.

Seems like we're pretty much on the same page then.

[Waazraath]

Oh, I don't think 2.5 should be considered an "at most" scenario at all, usually I see Fireballs if there's at least 3 enemies to be caught.

Just checking, you and LudicSavant are aware that this in the context of an average amount of targets with each and every fireball casted, and where most spells are assumed to be fireballl? Cause yes, of course you will hit 4 targets sometimes, and sometimes even more (though this is quite rare ime, but ymmv based on the type of games you play) - but you will also hit 1 target sometimes (sometimes an entire encounter is one big critter).

An average of 2.5 seems more accurate to me, but again, curious what other folks averages would be in a model like this.

[Chronos]

If we can decide on a fair average for number of targets for Fireball, I'll edit the spreadsheet. Or anyone else can make their own copy and edit it, if you'd prefer.

Quoth Waazraath:

- The cleric's spiritual weapon has the same problem as Sunbeam
I don't understand this criticism at all. The problems claimed for Sunbeam were that I was giving it too many targets and it's concentration. Spiritual Weapon is single-target and not concentration.



- additionally, the elephant in the analysis which has not been mentioned yet is that this only looks at offense. Defense is at least as important (staying alive, staying alive), and where the Fighter, (to a lesser extent) the Rogue have those resource-less build into their class (hp, armor, and class features like uncanny dodge, evasion, second wind), especially wizards are depending on exactly the same long rest resource they are using for their offense. Resources MUST be spent on these to stay alive, and a % of slots should be allocated for this for a realistic assessment. The cleric could arguably do without (though assuming some slots being spent on Aid or the like to keep up would not be a stretch).
Depends on subclass. An abjurer, for instance, has more HP than a fighter, and an AC at least comparable, plus they can keep their distance, which results in at least some decrease in incoming damage.

[Witty Username]

Just checking, you and LudicSavant are aware that this in the context of an average amount of targets with each and every fireball casted, and where most spells are assumed to be fireballl? Cause yes, of course you will hit 4 targets sometimes, and sometimes even more (though this is quite rare ime, but ymmv based on the type of games you play) - but you will also hit 1 target sometimes (sometimes an entire encounter is one big critter).

An average of 2.5 seems more accurate to me, but again, curious what other folks averages would be in a model like this.

So, for me I would expect 4, because 1 and 2 aren't situations where fireball gets cast. Usually, for my table fireball gets used to fight groups specifically, it's single target damage isn't all that great.

[gloryblaze]

So, for me I would expect 4, because 1 and 2 aren't situations where fireball gets cast. Usually, for my table fireball gets used to fight groups specifically, it's single target damage isn't all that great.

If I understand correctly, Waazraath isn't saying that fireball as used by a normal wizard in a normal adventuring party would have an average of 2.5 targets. Instead, they're addressing the specific thought experiment posed in the OP:


*Second-level spell slots are used on Scorching Ray. Third through fifth level spell slots are used on Fireballs and/or Lightning Bolts, each of which is assumed to hit four enemies.


So in a scenario where a wizard a) spends all their spell slots over the course of an adventuring day consisting only of combat encounters and b) spends every single one of their level 3-5 spell slots on fireball, what's the expected number of targets?

This really depends on the assumptions of the thought experiment. If the wizard faces a series of random Medium encounters until they're out of resources, at least some of those encounters might be solo encounters or duo encounters that drag the average number of targets below three. Conversely, if the wizard is in an infinitely large white room full of an infinite number of Tiny spherical cows perfectly packed 4 per 5-foot-cube, the average number of targets will be vastly higher.

[Waazraath]

If I understand correctly, Waazraath isn't saying that fireball as used by a normal wizard in a normal adventuring party would have an average of 2.5 targets. Instead, they're addressing the specific thought experiment posed in the OP:



So in a scenario where a wizard a) spends all their spell slots over the course of an adventuring day consisting only of combat encounters and b) spends every single one of their level 3-5 spell slots on fireball, what's the expected number of targets?

This really depends on the assumptions of the thought experiment. If the wizard faces a series of random Medium encounters until they're out of resources, at least some of those encounters might be solo encounters or duo encounters that drag the average number of targets below three. Conversely, if the wizard is in an infinitely large white room full of an infinite number of Tiny spherical cows perfectly packed 4 per 5-foot-cube, the average number of targets will be vastly higher.

This is absolutely understood correctly.

As for the assumptions behind the thought experiment: I interpreted it as a way to make statements about the game as it is played in reallity (so no infinitely large white rooms). Of course experiences will differ, but when I go through my published adventures (as a more or less 'objective' measure on how many creatures/encounter are to be expected), there are plenty of encounters with 1, 2 or 3 opponents, at every level, making 4 targets hit averagely with FB or LB extremely unlikely.

[Chronos]

Hm, I see the point: In actual gameplay, if a wizard is facing a whole bunch of weak enemies, Fireball will be a go-to option, and in those cases, it'll hit a lot of enemies, driving up the average number hit in actual gameplay... but that might not apply in this somewhat-artificial scenario where the wizard is spending all slots of levels 3, 4, and 5 on Fireball, because there might not be that many target-rich encounters in the whole day. To account for this properly, I think we'd need to know a typical distribution of number of enemies per encounter.

[PhoenixPhyre]

Hm, I see the point: In actual gameplay, if a wizard is facing a whole bunch of weak enemies, Fireball will be a go-to option, and in those cases, it'll hit a lot of enemies, driving up the average number hit in actual gameplay... but that might not apply in this somewhat-artificial scenario where the wizard is spending all slots of levels 3, 4, and 5 on Fireball, because there might not be that many target-rich encounters in the whole day. To account for this properly, I think we'd need to know a typical distribution of number of enemies per encounter.

Or, if the point is learning and exploring the possibilities, do the work for a range of values. Say 1 monster to 8 monsters. Present brackets, not single values.

[LudicSavant]

Hm, I see the point: In actual gameplay, if a wizard is facing a whole bunch of weak enemies, Fireball will be a go-to option, and in those cases, it'll hit a lot of enemies, driving up the average number hit in actual gameplay... but that might not apply in this somewhat-artificial scenario where the wizard is spending all slots of levels 3, 4, and 5 on Fireball, because there might not be that many target-rich encounters in the whole day. To account for this properly, I think we'd need to know a typical distribution of number of enemies per encounter.

Trying to find a 'typical distribution' is likely just going to be people arguing over playstyle preferences. A table could be anything from 'my DM never wants to run more than 1 creature at a time' to being packed to the brim like this (https://cdn.discordapp.com/attachments/821299993787760640/1009246031403302983/unknown.png) (screenshot is from the 5th encounter of the day of one of Pack Tactics's games, apparently).

I recommend just measuring for different cases and making it clear what those cases are. There's no reason you have to take only a single measurement.

[PhoenixPhyre]

Trying to find a 'typical distribution' is likely just going to be people arguing over playstyle preferences. A table could be anything from 'my DM never wants to run more than 1 creature at a time' to being packed to the brim like this (https://cdn.discordapp.com/attachments/821299993787760640/1009246031403302983/unknown.png) (screenshot is from the 5th encounter of the day of one of Pack Tactics's games, apparently).

I recommend just measuring for different cases and making it clear what those cases are. There's no reason you have to take only a single measurement.

I see we agree on this. Marked bundles of cases make for better data anyway over point assumptions that inevitably end up skewing things.

[Chronos]

It's easy enough to change the number of targets of a spell in the spreadsheet. What's harder is that, for different numbers, different spells become optimal. With four targets, Fireball is definitely the best choice, but what's best at 1 or 2 targets?

[Frogreaver]

IMO. I actually would do this analysis the opposite. I would start with the Fighter and determine his damage over a given adventuring day + chance to hit.

I then would look at fireball and set a few different chances to hit and compute how many enemies on average fireball would need to target to achieve break even. *(Need to account for spells like shatter and magic missile/chromatic orb as well - though for this exercise maybe just turn all level 1 and 2 slots into MM for ease - or alternatively assume those slots go toward defense/utility).