Spell Points vs Vancian vs alternatives. What is the best?

[jt]

Frank's right; spell points incentivize you to spam the best spell for the situation, while prepared spells incentivize you to spam the best general-purpose spell for your expected day. The "best general-purpose spell" may in fact be multiple spells since you can expect the average day to include at least a couple situations where Solid Fog is better than Polymorph, even though Polymorph is better in a wider variety of situations. The key here is that preparing in advance means you're operating with lower information; prepared spell point casting wouldn't have this problem (though I'm not aware of anyone who's used it).

If the situation changes often enough that this is not a problem (more than once per encounter) then I'm not sure why you even need a resource management system at all.

[Username17]

Frank I've heard you make the "Power points makes you find the best power and spam it" argument many times and I don't think it has any legitimacy. Vancian magic has, if anything, more incentives to create a structure where you just select your best power as many times as possible. That a psionicist can look at their entire spell list and decide what the best power is right now makes them more likely to choose different spells than if they had to fill their slots with what was the most generically useful and powerful spells they know the day before adventuring. If you assume the existence of some kind of power that's always the best that psionicists would endlessly spam then Vancian would get an almost exactly identical problem with that power. The Vancian caster would fill every spell slot with that power they could and then do the same for the next best power in the slots they couldn't.

No. None of that is remotely true.

World of Warcraft characters do indeed need Cooldown timers to do things other than spam the same action over and over again until they win. Fungible power points encourages ability spam. That is just factually obviously a thing that they do.

Imagine a Sorcerer and a Psion. The Sorcerer knows a couple of 4th level, 3rd level, and 2nd level spells. And she has spell slots that can be used for 4th, 3rd, and 2nd level spells. And in any particular battle, the Sorcerer's player must ask what their most appropriate 2nd level spell is, what their most appropriate 3rd level spell is, and what their most appropriate 4th level spell is. And the Psion has the same. All of that is the same. Except the Psion also draws their 2nd level, 3rd level, and 4th level powers from the same fungible power point pool, which means that in addition to asking what the best from each column is, they are also asking which column is the most efficient expenditure of power points!

Fungibility narrows optimal solutions. It just does. Mathematically provably, provedly does. You claiming otherwise is simply wrong. Mathematically wrong.

-Username17

[Miniature Colossus]

What happens if we assume that the 'best' power is some form of non stackable buff/debuff/area control that offer decreasing marginal utility? So you only really want to use the best power once or twice before some other power offers more utility. Some of the Psionics problem seems to be that the best power is blasting which technically has increasing marginal utility as monsters get closer to being taken out of combat.

[Username17]

What happens if we assume that the 'best' power is some form of non stackable buff/debuff/area control that offer decreasing marginal utility? So you only really want to use the best power once or twice before some other power offers more utility. Some of the Psionics problem seems to be that the best power is blasting which technically has increasing marginal utility as monsters get closer to being taken out of combat.

There will be times when this sort of thing is true. And then, as jt noted, it won't really matter what your resource management system is or even if you have one.

The point is that "slots" includes spontaneous-slots like the Sorcerer or even spontaneous-from-list slots like the Dread Necromancer. Those classes are still slot casters because they have distinct numbers of uses of their abilities that are drawn from separate ability columns. The only thing that power points adds to that is the ability to trade in your uses of Column A and Column B abilities for more uses of a Column C ability. That's it. That's the only difference.

There might be times when you don't do that because for some other reason it's tactically superior to use a rotation of different abilities rather than spamming one of them over and over. But when evaluating power points versus slots, it's important to note that the only real difference is that power points allows for levels of monotonous repetition of ability spam that slots does not. And you have to ask yourself why you'd want to do that.

-Username17

[jt]

There are a lot of configurations of this that aren't covered under prepared vancian vs spell points, so I'm splitting it up.

[*] One resource type or many. As Frank points out, under the standard spell levels scheme, every spell level is a different resource. You don't have to do that - you could just give wizards a flat dozen spells per day. You could also split your spell points by level (it'd be a bit weird). You could also split on something orthogonal to power, like Magic's colors. However you do it, giving someone multiple resource pools adds variety to what they do; you can run out of X and need to resort to Y. It also means more to track.
[*] Prepared or not. Preparation converts a homogeneous resource (level 3 spell slots) into heterogeneous resources (specific prepared spells) so you can run out of them separately, and because we've had it forever it's easy to forget just how damn clever that is. On the downside it does warp what sorts of abilities are useful. You could prepare spell points, but making a bag of things that adds up to X points gets tedious in a hurry.
[*] Chunky or continuous. If every power costs one point, that's the same thing as spell slots. If some spells take a double slot, that's still mostly slots. If spells take somewhere from three to seven slots, then I'm just describing spell points in an obtuse way. In between it gets vaguer. Chunkier resources are less flexible but usually quicker to reason about.
[*] Augmentability. D&D psionics traditionally let you tweak one ability by paying more points, but you could do this with slots by using the double slot from before. If you have level-based pools you can cast things from a higher pool to augment something (which is how metamagic works). If you have color pools you can do that thing Magic does with "if you spent red mana to cast this." Adding ways to augment spells makes your resources more continuous and has the same downsides as that. Done well, it also can mean you have fewer things to learn about a system, as one augmentable ability does the work of several others.

[MGuy]

I'm not truly convinced that Vanician casting is less boring than spell point casting. Not to say Vanician is in fact the more boring one after all but I don't think it is as 'mathematically true' as is being asserted here. (Yes I know that the mathematically true part was in reference to the narrowing of optimal choices and not which was more boring). The idea that people make perfectly rational choices as the norm, and then basing my design off of that just feels uncomfortable to me. I rarely see people think much about their choices and even make bad choices despite being given clear prompts about what choice they should make. That's just a minor gripe.

What I'm more concerned with is that I'm not sure of the validity of the 'mathematically true' claim about the narrowing of optimal solutions. I can see where Frank is going with the idea in the scenario he presented. To straighten it up a bit Frank is assuming that the Sorc at some point or another 'has' to use an ability off of another list because of attrition. That's the only way what he's saying can make sense. The idea is, and anyone can of course correct me if I'm wrong, at some point even if one of the Sorc's 3rd level spells 'could' be the most efficient use of a spell slot that in some situation he simply would not have access to it and therefore would HAVE to find another optimal solution from a spell slot at another spell level. This 'has' to be what he's talking about because all other things being even there would effectively be no difference between the psion's PP usage and the sorc's spell slots given the same 'spells' and 'spell levels' between the two of them. So what he's saying is that the Psion will not be in the same position as the sorc because the resource they use to determine spell slots allows them constant access to all of their spell slots. Therefore the psion can stick to a narrower number of optimal solutions because they will have consistent access to the 'most' optimal solution to a given problem while a sorc will eventually be forced to find another optimal solution given attrition. I don't know that I agree, in this case, that being forced to use less optimal abilities then equals a broader range of optimal solutions. It seems to me more like you're just losing access to optimal solutions and being forced to make choices among less optimal solutions. I get that the thinking here is that if you can't use the same optimal solution you were using then the next best thing becomes the most optimal solution and therefore broader range of optimal solutions. I don't think that's how anyone in real life views it though. I also don't think that you can just assume that the Sorc is necessarily using tactically distinct abilities from lower level or higher level slots. There is every possibility that once a sorcerer loses their top two spells slots that they run into a situation where their lower level abilities legitimately can't do anything in the situation they are in and they get stuck sitting on their hands. He 'could' be right in a specific scenario where the sorc is dynamically changing tactics as spells get fired off but there's no reason to assume this is true in any given combat.

When Frank claims that 3, 2, 1 is less interesting than 2, 2, 2 I can't say I agree because I don't know what these numbers mean, what the abilities are, that the options chosen from 3, 2, 1 are tactically distinct from each other while assuming the options chosen repeatedly from 3 and 2 aren't tactically distinct options within those columns. I don't know what these numbers actually mean in practice and simply claiming that one is more interesting than the other with finality (it being mathematically true) doesn't sit well with me. I'd need actual flushed out situations because I don't think the conclusion that's being pushed follows from the evidence being put forth. These things, under certain conditions could end up just as Frank is claiming but I'd have to look at these things in actual play because in my experience Sorcs aren't really any more or less interesting to me than Psions.

Also, JT's counter argument with solid fog and polymorph doesn't make sense to me either. The thing Frank seems to be against is spell spamming. I don't know that claiming to be able to spam generically useful spells is a 'good thing' considering what he's saying is boring. If you load yourself up with generically good spells that are good in most situations such that fight to fight you're not effectively changing your tactics then... what is it that is boring about the Psion doing literally that? Is it somehow worse for a Psion to be adequately prepared with more solutions for situations while a prepared caster instead uses the same tactics from combat to combat because they can't afford to be prepared for more strange situations?

I also don't get how that leads to "If the situation changes often enough that this is not a problem (more than once per encounter) then I'm not sure why you even need a resource management system at all." I'm not sure I disagree with it because I am somehow unable to follow the logic that makes this statement true. I was under the assumption that the purpose of resource management is to provide incentives for a player minimizing their resource use on a given obstacle so that they can reach their goal with enough resources to overcome what presumably is going to be the biggest obstacle at the end and not to force players to not spam X thing or X set of things in a given combat encounter.

[Mask_De_H]

The idea of a resource management mini game is to give a player tactically interesting choices during a game, one of which is "do I blow my load now or do I conserve". Another is "can/should I use X right now? If not, how do I get in a situation to use X".

Another thing is, the argument against PP hinges on it being 2, 2, 2 instead of 3, 2, 1. A slot caster is forced to change what spell they use when they run out of one slot, so by default their most optimal spell choice changes. A PP system doesn't have this, so the PP caster can roll face on the most optimal spell choice they have until they're spent. The idea with the Sorcerer is that an optimal level 2 and 1 spell are viable life choices, because the Sorcerer player has to have fallbacks. A Psion does not. Since there's no segregation of powers you know besides cost in a PP system, finding the overall optimal Psion ability is a math problem. The hypothetical of a Sorcerer running dry with their top two/three spell slots is less important than the fact a Sorcerer has to use different abilities, creating a more complex answer to the resource management questions above.

[Emerald]

Also note that while we're talking about 3-2-1 as if the Vancian caster always wants to start with his Nth-level slots, run through all of them, go down to [N-1]th level slots, run through all of those, and so forth, that's a simplification for comparison purposes and things rarely if ever work out like that.

A wizard with 4/4/4/4 0th-/1st-/2nd-/3rd-level spells prepared might find that, of those, there are 0/1/3/2 spells that are relevant to a given encounter. He might start with the 2 3rds, go down to the 3 2nds, and end up with the 1 1st, sure...but chances are very good that the encounter isn't going to last 6 rounds and even if it does the circumstances may change midway, modifying the list of optimal spells.

For instance, if the encounter is against an invisible creature, a flyer, and a (Cold) creature, his 3 2nds might be see invisibility, earthbind, and scorching ray, so killing the first or second creature moves the first two spells to the not-useful-anymore column and killing the third changes the calculus because now a different spell might be more damaging without the fire vulnerability. Alternately, he uses those three spells perfectly optimally, and then all three creatures are dead because the party rogue shanked the now-visible creature and the party fighter shanked the now-grounded flyer, so while all those 6 spells were generally optimal for that encounter in a vacuum, he got more mileage out of his lower-level spells because they synergized better with the party.

And then there's pure adventuring day considerations. Use all your 3rds now, have no 3rds later and get screwed when you run into a boss monster. Use all your 1sts now, have no 1sts later and end up wasting 2nd- and 3rd-level spells on weak random encounters.

Those considerations are where the "tactically interesting" part comes in. Not just using 3-3-3-2-2-2-1-1 but sometimes using 2-3-3-2-1-1-3-1-2. Sometimes just using 2-2-2-1-1-1 because 3-3-3 would be overkill or not relevant. Wanting to pick those more niche spells instead of generally-applicable ones because you have more more-specialized spells instead of fewer less-specialized powers, and being able to because your 2nds and 3rds aren't competing for one another at all.

[jt]

Also, JT's counter argument with solid fog and polymorph doesn't make sense to me either. The thing Frank seems to be against is spell spamming. I don't know that claiming to be able to spam generically useful spells is a 'good thing' considering what he's saying is boring. If you load yourself up with generically good spells that are good in most situations such that fight to fight you're not effectively changing your tactics then... what is it that is boring about the Psion doing literally that? Is it somehow worse for a Psion to be adequately prepared with more solutions for situations while a prepared caster instead uses the same tactics from combat to combat because they can't afford to be prepared for more strange situations?

The point I was trying to make (perhaps not very well) is that the lower information you have when preparing spells vs casting spontaneously lets you sneak some variety in.

Let's say there's an optimal spell for a given encounter. As a spontaneous caster, you use it until the encounter is over. Let's also say that there's an optimal spell for an unknown situation (for 4th level spells its Polymorph). If you have to choose your spell before seeing the encounter, you'd pick this and use it until the encounter is over. This is even worse.

However, if you're bulk preparing spells in advance, you don't have to limit yourself to the best spell. If you know - or are willing to bet - that at some point during the day you'll see a specific situation which you have a better spell for (say combats), you can replace one generic spell with a more specific one for that (Solid Fog). If you're going to bet on seeing more of that, you can swap out more slots. So the optimal list contains a fair number of more niche spells, because you can make reasonable guesses. But if you guess wrong, you run out of generic spells and have to start making things work with the niche spells for the wrong situation, which is where a lot of the fun comes in.

Additionally, in practice most people don't prepare the optimal list, and instead fall into a trap of trying to prepare for every contingency, so they run out of specific situations early and end up with weird mixes of stuff more easily than they should. You rarely see someone prepare more than double of any given spell, even though optimally in 3.5 your top 4-5 slots should be the highest level mass buff or battlefield control spells you can carry.

However, as my more recent post clarifies, that's not actually a property of slots vs points, but rather a property of prepared vs spontaneous. I think there are different issues with prepared spell points (mainly: it's fiddly) but if you used them you would run into the same neat thing where you get interesting decisions because you prepared the wrong stuff.

[deaddmwalking]

This seems relevant...

In our heartbreaker, some spells require a full action and others require a standard action. Spamming the same full action spell over and over is potentially undesirable because usually people want to move in combat (for example to see an opponent that is behind cover). It is not totally uncommon to spam a spell that modifies your weapon attacks (ie, channel a flaming burst through your sword). The type of action required can be a cost that modifies a spell.

Additionally, sometimes a spell has a good chance of 'winning' an encounter. Using that spell in Round 1 makes sense, but it is less important to use it if the encounter is almost over (most enemies are severely damaged). You might use charm monster in Round 1, but you'd be less likely to use it in round 6.

It is still true that repeating the same action round after round after round is usually not interesting or satisfying - trying to ensure that there are multiple ways to contribute meaningfully when you have an action is an important part of the game - if there is one best action in the same combat and it is the same action EVERY TIME it gets old fast. There are certainly times when an opponent, due to their vulnerabilities or resistances, encourages spamming the same attack repeatedly, but as long as the opponents change frequently enough that isn't NECESSARILY a problem.

[jt]

Oh, speaking of solutions in people's homebrews...

I've broken combat into a sequence of named phases: ambush, ranged, melee, close. By default, everyone gets a turn then you advance to the next phase. Attacks/spells/etc have specific phase(s) they work at, so you can't spam the same thing over and over again. This solves the problem quite bluntly.

Additionally, many spells skip your turn and take effect on your turn in the next phase, which lets them do things twice as impressive as other abilities, so that's what makes magic feel special. (And I'm still experimenting with cross-combat fatigue mechanics.)

[Daztur]

Frank I've heard you make the "Power points makes you find the best power and spam it" argument many times and I don't think it has any legitimacy. Vancian magic has, if anything, more incentives to create a structure where you just select your best power as many times as possible. That a psionicist can look at their entire spell list and decide what the best power is right now makes them more likely to choose different spells than if they had to fill their slots with what was the most generically useful and powerful spells they know the day before adventuring. If you assume the existence of some kind of power that's always the best that psionicists would endlessly spam then Vancian would get an almost exactly identical problem with that power. The Vancian caster would fill every spell slot with that power they could and then do the same for the next best power in the slots they couldn't.

That's why Vancian works better if you don't let people memorize multiple copies of the same spell. The inflexibility of Vancian casting encourages information gathering (to know what to prepare) and creativity in figuring out how to use your crappy leftover spells at the end of the day to do something useful.

You could also keep wizards from going quadratic by emphasizing them getting more spell slots instead of higher level spell slots so wizards wouldn't gain much raw power or novaing ability but would gain a lot of flexibility.

In fiction terms memorizing spells is kinda dumb, what works better is to flavor it as demon binding. Wizards do rituals in downtime to bind various demons and then in return for being let go the demons do a task for the wizard which in game terms is having a spell cast. Also provides a reason for complicated wizard labs, that shit is necessary to bind a lot of the demons.

[Chamomile]

One thing to note is that if you can only use a spell once, then spells which are identical or nearly identical in performance but have different names become more valuable. You get this in Magic: the Gathering and the original Guild Wars, both of which limit the number of copies of a power you can have access to at once to four and one, respectively, where a pair of near-identical but technically different cards/abilities will sometimes be equipped so that you can have eight copies of what is essentially the same card or effectively halve the cooldown on your ability or, in the case of Vancian casting, cast the same spell twice. That's something you can design around no problem, but it's not something that 3.X actually did design around, so if you backport this as a house rule into a 3.X game, be aware that there are probably lots and lots of redundant spells that someone can dumpster dive for in order to fill up on spells that do [thing] even if they can only have one copy of each.

[deaddmwalking]

One of the things that 3.x did right was allow clerics to drop any prepared spell for a cure spell of the equivalent level. In 2nd edition, you pretty much had to prepare all the healing spells; in 3rd edition you could prepare cool stuff and sometimes even be allowed to use it.

Making all wizards work like sorcerers where you can prepare spells as usual but cast them in any combination offers a middle ground. There are encounters where you might cast the same 3rd level spell three times in a row, but you're also more likely to prepare a spell that is highly specific/rarely used, knowing that if the opportunity doesn't come up, you can use that slot to cast a second copy of a more versatile spell.

In spell points, you can convert some number of 1st level 'slots' into a 3rd level spell; spell slots with substitution gives you more flexibility that straight prepared spells but it doesn't allow you to convert any higher level spells into lower level spells or vice versa.

[John Magnum]

IIRC 5e actually does have Wizards work that way. I think Pathfinder released a class that worked a lot like that, but I don't remember what it was. I thought it was the Magus but it looks like that just has regular 3e Wizard style preparation.

[Whiysper]

Yep, all 5e prep casters prepare x spells, and then cast y spells/day chosen from list x.

As you say, it's sort of a sweet-spot middle-ground, and proved very popular - at least at my table.

[angelfromanotherpin]

Does anyone know where that style of casting originated? The first time I saw it published was in Cook's Arcana Unearthed.

[MGuy]

IIRC 5e actually does have Wizards work that way. I think Pathfinder released a class that worked a lot like that, but I don't remember what it was. I thought it was the Magus but it looks like that just has regular 3e Wizard style preparation.

Arcanist in PF. One of my players is using it now.

[OgreBattle]

So how does Shadowrun handle the "spam your best spell" problem that comes with being able to cast everything you know until you faint?

[Mask_De_H]

So how does Shadowrun handle the "spam your best spell" problem that comes with being able to cast everything you know until you faint?

Drain codes and Force, I think. The amount of damage you've already taken, plus the general unpredictability of the Drain roll disincentivizes rolling face on the biggest spell you know because if you flub the roll, you're out of the fight or seriously hurt.

[deaddmwalking]

So how does Shadowrun handle the "spam your best spell" problem that comes with being able to cast everything you know until you faint?

Drain codes and Force, I think. The amount of damage you've already taken, plus the general unpredictability of the Drain roll disincentivizes rolling face on the biggest spell you know because if you flub the roll, you're out of the fight or seriously hurt.

But you're still incentivized to spam the best spell you have that has no risk of drain, right?

[pragma]

My experience with Shadowrun 3 and 4 was that you could optimize your drain stat at char gen and rarely experience drain at all even when throwing around fairly high force spells. I think the drain pricing was in general a little low, and especially low on stun bolt and a variety of powerful illusions.

SR5 was silly for a lot of reasons, but the idea of increasing drain codes wasn't a bad one even if the execution was.

[Whipstitch]

So how does Shadowrun handle the "spam your best spell" problem that comes with being able to cast everything you know until you faint?

It doesn't. That system's approach to the problem is to hope that the solutions to various problems are different enough that you cast Detect Life while searching for people and Heal when you want to heal. As long as the drain codes are the same Shadowrun doesn't differentiate between Power Bolt spam or the Five Moves of Doom. It's a fungible system and since Frank already explained what that means I don't feel bad about saying you're all just fucking around.

[souran]

So, people who are arguing that spell points don't cause people to just trade away their lower level spells for a few extra castings of their best spells know that they are not just theoretically wrong but emperically wrong as well right?

This exact thing has been tried twice, first it was a bad idea and didn't work with players option: spells and magic in 1996 and it was still a bad idea that didn't work again 8 years later with Unearthed arcana and each time the end result is that players dump as many spells as they can below their current max level and take a few extra max level castings.

It's also not hard to figure out why. Assuming your encounters are level appropriate, then "current max level" spells generally are "I win" buttons. Lower level spells usually are not. every time I have seen spell points used it has amplified the 10 minute workday issue.

[hogarth]

So, people who are arguing that spell points don't cause people to just trade away their lower level spells for a few extra castings of their best spells know that they are not just theoretically wrong but emperically wrong as well right?

It's trivial to come up with a spell point system where that doesn't happen. For instance, you could come up with a system where level 1 spells cost 1 point, level 2 spells cost 2 points, level 3 spells cost 4 points and level 4 spells cost 8 points and a 7th level wizard has 15 spell points. Or you could have a spell point system where each spell level has a cooldown (along the lines of the recharge magic variant in 3.5 Unearthed Arcana).