D&D 3E/3.5 Jonathan Tweet: Third Edition and Per-Day Spells

On the Third Edition design team, we were tasked with rationalizing the game system, but there were some big elements of the system that we didn’t question. We inherited a system in which spellcasters get better in three ways at a time as they level up; they get more spells per day, higher-level spells, and more damage with spells of a given level. In retrospect, that problem is easy to see, and we didn’t fix it. We also inherited a system that balanced powerful class features, notably spells, by making them usable once per day. The problems with that system are less obvious, and we didn’t fix this system, either. But the 3E system laid bare its own inner workings, and so soon enough designers saw that there were issues with this system, and over the years several of us designers have tried to address it one way or another.

In classic dungeon crawling, the default best strategy is to take each room one at a time and regain your hit points and spells after each one. That’s no fun, so people usually don’t play that way. For 3E, we spelled out that the game was balanced for four average battles between heal-ups, but actual practice varied. Whatever per-day powers are balanced at one rate of fights per day are necessarily unbalanced at faster or slower rates. Classes with lots of per-day power are too strong when there are one or two fights per day and too weak when there are five or more. Individual Dungeon Masters might be able to schedule the action in such a way that they maintain the sort of balance they’re looking for. If that works, it represents the DM’s efforts and not anything we on the design team could accomplish through system design. Many Dungeon Masters might find the per-day rules convenient precisely because they allow the DM to modulate the threat level up and down. DMs rule on how many encounters the party has in a day and whether they can suspend their mission long enough to reset their spells and other per-day powers. A dynamic I’ve seen over and over again, however, is that players with spellcasting characters are adept at talking the DM into letting the party rest. When the spellcaster is out of spells, they need a night’s rest a lot more than the other characters in the party need to press on. When a mission goes south and the encounters burn up more per-day resources than the DM figured they would, the party often simply camps out for the night and sets out the next day with spells reset to full.

Limiting spells by day also means that a spellcaster’s power level is different when they’re in a preliminary skirmish compared to when they’re in a climactic showdown. When it’s a high-priority battle or when the player knows that there’s a long rest afterwards, the spellcaster can use their best spells without worrying about holding back. This effect is something of a game-wrecker when the party arranges to jump the big bad guy after prepping up to full. With a well-placed teleport, the party’s spellcasters can unload all their best “per-day” spells for the one battle that matters that day (an “alpha strike”). Classes with at-will powers can’t “unload” the way spellcasters can.

The per-day system also changes up balance for NPCs. Generally, when a party attacks an NPC boss of some sort, that NPC is in a fight for their life, and they cut loose with every per-day power they can manage. Fighter NPCs aren’t particularly dangerous because they have no such resources to unleash. In my campaign, I found the psionicist NPC the most dangerous because they could use the point system to cast at full capacity every round. As player-characters, psionicists have all the balance problems of the wizard and then some.

Seeing the issues with per-day powers, the designers started experimenting with per-encounter powers in supplemental material. The psychic warrior, for example, had a “focus” that they could expend once in the battle in order to have a special effect. At that point, designers were still in simulation mode, and encounters that were “per-encounter” by fiat seemed too artificial. The psychic warrior had a believable, in-world reason for their “per-encounter” abilities. Tome of Battle: Book of the Nine Swords (2006) introduced special, limited-use powers for martial classes. By 4E, the designers fully embraced per-encounter powers.

Fourth edition established balance among the classes by giving all of them per-day and per-encounter powers. That’s one way to solve the balance issue. 4E is so well-balanced that it’s hard to make bad choices in character design. This approach had the unfortunate effect of making the classes all feel sort of the same.

With 13th Age, Rob Heinsoo and I took a different approach. We turned 3E’s four-fights guideline into a hard rule. You get your spells and hit points back not just by resting but only if you have engaged in a minimum amount of fighting. After your fourth fight (or after four fights’ worth of fighting), the party gets to reset to full. Alternatively, the party can admit defeat and get a heal-up without “earning” it, but admitting defeat entails a “campaign loss,” as determined by the GM. This system creates a lovely rhythm, with characters feeling flush and confident in the first fight, feeling hard pressed in the last fight, and then feeling good again when they heal up. I play a cleric in a 13th Age campaign, and the last fight before a heal-up is tough going. The last fight is so tough that we player all know that the decisions and rolls we made in the earlier fights all mattered in terms of what we have left for the last one.

 

[The Crimson Binome]

My vote for the most relevant but obscure one, was that in 1e AD&D whenever the spell Haste was cast on you, that it aged you 2 years - a drawback that meant you saw it cast only a few times a campaign in groups that became aware of the drawback.

Did 1E also include the system shock roll for magical aging, that 2E had? Even if you were playing an elf, and you could withstand a few centuries without slowing down, it was still a random chance that you'd just die outright.

 

[Doug McCrae]

Most spells (and nearly all combat spells) resolved on the same initiative as they were cast, thus no casting time a la 1e-2e and thus much less opportunity to interrupt the casting. As a side effect this also meant spells on average resolved earlier in the round sequence.

Casting time as a modifier to initiative is an optional rule in 2e. If that rule is used, spells are generally faster than weapons. The casting time of a spell is usually equal to its level, though some useful combat spells such as dimension door and stoneskin have a casting time of 1, while weapon speed factors range from 2 (dagger) to 15 (two-handed sword). The player of a wizard actually asked me to use this rule - perhaps he felt it gave him an advantage.

 

[cbwjm]

Too bad cantrips in 3e weren't at-will instead of fire and forget (or just lose the slot in the case of sorcerers, etc). If they were at-will and scaled a bit with level, casters might have been slightly less inclined to rest all of the time since they would have still been able to contribute to the fight instead of ineffectively plinking away with their crossbow.

 

[Minigiant]

Actually, the 5e fighter's action surge is exactly this type of mechanic. I even expanded it and allow my players to double damage instead of double their actions.

it's more like somewhere between a daily and an encounter feature It's a step in the right direction. 4e, 13A, and even 5e are all steps in the right direction. I still kinda wished 4e, 5e, PF, or 13A would have fully delved into the idea of a truly heroic and amazing action that a warrior or expert could do once a day. Even barbarians in all these games are sorta tame. I wish designers would really dive into modern media and pick out the epic actions that match their settings.

But I guess that's what the psychic and wuxia classes and subclasses are for.

 

[Panda-s1]

There's more to the issue of 3e (and 4e and 5e, for that matter) casting unbalance that the article maybe intentionally skips over: in 3e successfully casting a spell was made much easier than in previous editions.

Here's how:

Most spells (and nearly all combat spells) resolved on the same initiative as they were cast, thus no casting time a la 1e-2e and thus much less opportunity to interrupt the casting. As a side effect this also meant spells on average resolved earlier in the round sequence.

The concept of 'combat casting' was brought in (an 'option' taken IME by every single caster ever played) making spells even harder to interrupt. In 1e-2e ANY interruption of any kind - even just a non-damaging jostle - would make you lose the spell.

Much or all of the real risk was nerfed or removed from some key game-altering spells. No more system-shock roll required when polymorphing someone not yourself, leading to rampant polymorph headaches. No more risk of instant death by teleporting into solid rock below your target point, thus causing 'alpha strike' to become the obvious go-to tactic for any party that could do it. Etc.

Magic in 1e-2e was rather high-risk high-reward. Take away the risk and no wonder casters got out of hand in 3e.

so many spells that make you age lol. I also went through the DMG and read about the stuff clerics are actually supposed to do to get high level spells. did anyone actually make clerics do ritual sacrifice in AD&D times?

Too bad cantrips in 3e weren't at-will instead of fire and forget (or just lose the slot in the case of sorcerers, etc). If they were at-will and scaled a bit with level, casters might have been slightly less inclined to rest all of the time since they would have still been able to contribute to the fight instead of ineffectively plinking away with their crossbow.

but if they did that a wizard could DESTORY A MOUNTAIN O: [/s]

 

[Alzrius]

Magic in 1e-2e was rather high-risk high-reward. Take away the risk and no wonder casters got out of hand in 3e.

Exactly this, and the thing to remember was that these were everywhere. You identified a good number of them, but there were others.

Wizards who prepared spells needed (if I recall correctly) ten minutes per spell level for each spell prepped. There wasn't any of this "you get them all in an hour, regardless of number" thing that 3E introduced. So if your wizard went "nova," if they were a high-level character they might need to spend the better part of a week re-upping. What we now think of as "concentration checks" were automatically failed if you were ever in a situation where you had to make one. And clerics didn't prepare their spells; they asked their gods for them (or at least, the ones above 2nd level), so you might not get that earthquake spell if your deity thought that heal would work better for you.

Oh, and wizards had a hard cap on how many spells of each spell level they could know, based on their Intelligence. That was if they made their roll to learn that spell at all; if they failed it, they couldn't try again until they'd gained a level.

3E added a lot of power to each level that characters gained, and also removed a lot of the limitations that casters had always faced. It's understandable why they did so, since those were things that a lot of people had been house-ruling away for years because "limits and restrictions aren't fun." But as Celebrim insightfully noted, the law of unintended consequences is a thing, and fixing one set of problems often means introducing another.

 

[The Crimson Binome]

Too bad cantrips in 3e weren't at-will instead of fire and forget (or just lose the slot in the case of sorcerers, etc). If they were at-will and scaled a bit with level, casters might have been slightly less inclined to rest all of the time since they would have still been able to contribute to the fight instead of ineffectively plinking away with their crossbow.

You describe reserve feats. They were one of the later additions in 3.5, and basically gave you an at-will magic attack that scaled with your highest-level unused spell slot. If you had the right feat, then as long as you had a fireball prepared in a level 3 spell slot, you could throw out a 3d6 mini-fireball. If you had meteor swarm prepared, then you also had a 9d6 mini-fireball, until such time as you ran out of meteor swarm for the day.

It did give strong incentive to hold onto your spell slots, although personally, I prefer a game where the wizard can contribute meaningfully with their staff or crossbow.

 

[RangerWickett]

A design conceit I don't see often in tabletop RPGs is the "super meter," a la fighting games.

In every edition of D&D except 4th, wizards open with their best spells. And in 5e, fighters will throw their superiority dice as quickly as possible, because the best way to win a fight is usually to pour as much damage into each enemy as necessary to remove them from the field as quickly as possible.

But in fighting games, you'll often start with a variety of attacks to probe your opponent's fighting style. You don't just use your strongest attack, because sometimes there isn't an opening to pull it off. And the whole time, you slowly charge up a super meter, and when the moment is right, you unleash an amazing attack that looks ultra cool.

Now obviously in a tabletop game, we don't want combats to take forever, so we have to balance complexity with speed of play. It's really hard to design a system that can make something other than "do damage" be a valid use of your turn without also making combat take forever. (I think I've cracked that nut, though; I just need to put my rules down on paper.)

Still, it'd be cool to see more super meter systems.

Pathfinder had something sorta like this with the Kineticist. You could spend an action to 'gather power,' and then your attack on the next turn would be stronger. In practice, though, it was almost never tactically viable to spend a whole turn powering up.

 

[Krachek]

Maybe use a mix system?
use rapid (5 minutes) short rest after 2 worthy fights, and a rapid long rest after 6 ( or 4) worthy fight.
otherwise player can use regular 1 hour or 8 hours rest.
it may encourage players to continue explore and take another fight since they will get a rest after the next one.

 

[Gorgon Zee]

What I got from the article was, "Here's some problems that arise from trying to make logical sense of the world, and our eventual solution for those problems was to stop trying to make logical sense of the world."

That's pretty disappointing.

I guess it's disappointing if you want a role-playing game to be a good simulation above and beyond all else, but for me, it's more disappointing to see rationalizations as to why you cannot rest overnight and regain all your resources.

Exploration of a dead city is a standard D&D trope. But any simulation-favoring GM running such a thing has to deal with the fact that the sensible approach is to alpha-strike a room or two, rest overnight and repeat. So they add "wandering monsters" to punish players for behaving normally. Or they need to add a plot device ("the pyramid will explode in 18 hours") to add a sense of urgency and stop resting. The simple approach of saying 'you don't get resource back until after three hard fights' immediately and simply solves the issue -- for GMs who don't require that simulation is the prime focus.

Given that is much of D&D is already designed to make the game simple and fun at the expense of realism (e.g. hit points, complete lack of wound rules, pass/fail successes) I'd say that since this has been an ever-present annoyance in D&D, I'd throw it into the camp of hit points: A system that is nothing like realistic, but makes the game easier and more fun.

Maybe I'm just more crotchety than most, but whenever a GM tells me "your cannot rest here because there are regular patrols" or "there are many wandering monsters here", or "the McMuffin will be unobtainable in X hours if you rest" I just here them saying:"You are not getting a full recovery here because it will make the game less fun and I don't like telling you that so I'll give you an in-game veneer of explanation which we all know really just means I don't want you to take a long rest as I've scaled the encounters assuming that"

In the real word archeologists take months or years to explore ancient ruins. Realistically, D&D explorers will do the same. But we're not really interested in realism, we're interested in having fun!