Miscast risks as a balancing system for magic

[Harshax]

WHFR has combat fumble tables where you can injure yourself. At least, 1st edition did. Not sure about the new versions.

So, if you can maul yourself with your own attack. Spell miscasts could be thematically appropriate. WFRP has lots of other thematically appropriate reasons why magic can backfire dramatically.

[jt]

But the question is whether they're a good balance lever. Math-wise, they are the very definition of adding imbalance.

It sounds like you're conflating balance and consistency, which I don't think is the most helpful definition for balance. I'd consider balance to be more akin to expected value.

That is, a weapon that does 1d4+2 damage has the same expected value (4.5) as one that does 1d8 damage or a spell that does 1d12-2 damage (if you roll a 1 it "miscasts" and heals 1 damage). Each is less consistent than the last, but I'd consider all of these balanced.

(I'd also consider all of these boring, but that is generally true of toy examples.)

[Foxwarrior]

Well, teeeechnicallly, if you're fighting a guy with 7 HP, the 1d4+2 damage weapon is pretty bad compared to the 1d8 damage which gets a one-shot 25% of the time, and the 1d12-2 which gets a one-shot 33% of the time, while against a 3 HP enemy the 1d4+2 is clearly the best.

[jt]

And if you're fighting a load of guys with 9 HP, the spell is the only one that ever one-shots anybody.

But does it have any bearing on the underlying point?

[Zinegata]

It sounds like you're conflating balance and consistency, which I don't think is the most helpful definition for balance. I'd consider balance to be more akin to expected value.

That is, a weapon that does 1d4+2 damage has the same expected value (4.5) as one that does 1d8 damage or a spell that does 1d12-2 damage (if you roll a 1 it "miscasts" and heals 1 damage). Each is less consistent than the last, but I'd consider all of these balanced.

(I'd also consider all of these boring, but that is generally true of toy examples.)

Except baking in consistency is what real design balance is actually about. Probability in a vacuum is really easy. Balance by contrast is not just about getting a reasonable average, but about actually hitting those average numbers consistently.

For example: If you roll a D10 you expect an average roll result of 5.5. If you want to kill a goblin with 15 HP, you need to roll an average result around 3 times.

The thing is the variance of your actual results ranges from 1 to 10. So if you keep rolling natural 1s, you will need as many as 15 rolls to kill a goblin. By contrast even if you keep rolling natural 10s you still need at least two hits to kill the goblin.

The damage rolls are thus actually imbalanced in this context, because poor rolling disproportionately punishes the players. If they keep rolling poorly, then the goblin lives for 15 rounds instead of the expected 3, whereas a perfect roll only shortens their lifespan to 2 rounds from 3.

(Caveat: And yes, I know the goblin can die in two rounds even if you don’t roll two natural 10s. But it doesn’t “balance” out the fact that a goblin with a lot of luck can end up living 5x longer than the “expected” result of 3 rounds. A goblin can do a lot of damage in those 15 rounds.

What it instead means is that while in most cases the goblin lives 2-3 rounds, you will also get cases where the goblin lives far longer than just 3 rounds and that can totally ruin all your assumptions. ).

By contrast if the game has a base damage of 5, and you only roll a D3 to get either a -1, +0, or +3 modifier , then the number of rounds to kill a 15 HP goblin ranges from 4 to 2. You need 4 if you rolled three -1s, 2 if you rolled two +3s, and 3 in all other cases which is the majority.

And note: It doesn’t look balanced at first glance because the positive modifier is +3 while the negative one is -1. But in terms of possible outcomes the chance of needing 2 hits to kill a goblin is the same as the chance to kill it in 4, and in most cases you get the average result which is 3 hits to kill a goblin.

In short, game design balance is not about probability numbers in a vacuum. Its about balancing the outcomes when all variables are in play.

[Zinegata]

WHFR has combat fumble tables where you can injure yourself. At least, 1st edition did. Not sure about the new versions.

Warhammer Critical Fumbles:

https://wfrp1e.fandom.com/wiki/Critical_Fumbles

Injuring yourself is about the worse it can get with the fumble tables, as opposed to magic miscasts where getting the character killed outright isn't even the worst possible outcome.

In any case, no one's saying it's wrong to add fumble or miscast tables based on theme.

The objection is to the idea you can use it to balance things. Its about the equivalent of trying to control a raging fire by pouring more gasoline on it.

[OgreBattle]

bonking yourself with a hammer is goofy, but I think being imbalanced, stunned, having a penalty to act or giving a bonus for badguys to act upon you is a good use of a 'fumble' table.

On 'fumbles', I want to move away from the D&D "ha ha 1 in 20 chance to crit or humilating fumble!" and have more a more steady gradient of little to big misses and hits. Those "success with complications" systems are kinda there but instead of leaving it up to social contract as to what happens its a steady effect like being flat footed.

It also gives more time for friends and foes to act upon the result instead of the result being immediate. Like if your 'miscast' 'fumble' is "The wizard gets a painful headache, they cannot take Reactions and are now easy to hit until next turn" it means the wizard's allies should cover them or maybe pick em up and haul them away while the wizard's foes make a beeline for power attacking them.

Say with warhams 40k warp phenomena maybe your allies have a chance to inject you with the serum, hold their hand and use your social bonds to keep em with you, etc. I also think this leads to interactions that people want with interrupts, but not doing actual interrupts.

[Foxwarrior]

Yeah jt, I do think Zinegata is conflating balance and consistency. I guess players often do too though, they look back at a session and notice that one character accomplished nothing (because of poor dice rolls possibly) or stole the spotlight by rolling well, and infer that the game is unbalanced when really it's just designed to have a certain rate of lucky and unlucky flukes.

Also,

It doesn’t look balanced at first glance because the positive modifier is +3 while the negative one is -1.

Oh my.

[jt]

I guess players often [conflate balance and consistency] though, they look back at a session and notice that one character accomplished nothing (because of poor dice rolls possibly) or stole the spotlight by rolling well, and infer that the game is unbalanced when really it's just designed to have a certain rate of lucky and unlucky flukes.

That's a fair point. And high variance abilities are more likely to have a string of either. So in the hypothetical fandom of that toy example, we can also expect lots of forum threads where people say spells are overpowered or underpowered.

Balancing the actual mechanics is hard enough. Doing that while also balancing perception of balance... I'd rather close my eyes, cover my ears, and pretend that's not a problem. Though if we're being adults about it and designing around that too, there are levers for doing that somewhat independently. Perceived balance seems to start with special effects budget vs number of drawbacks and conditions.

[owlassociate]

Here's an idea for miscasts: You make a spellcasting check when you cast a spell and if you fail the check, the spell is still cast but you're unable to cast it again until a condition is met. Maybe there's an additional miscast effect like taking some damage or a status effect (or even something kind of wacky like an uncontrolled teleport or a demon pops out), but it doesn't ruin the spell it just adds some danger, drama or chaos.

[MGuy]

The objective shouldn't be to force in a mechanic just because. The first question that should be asked is 'what do I want the game to look like?'. Outside of trying to fit some theme there is nothing that I can imagine, or have seen mentioned so far, that tells me why anyone would want miscasting to be a thing outside of fitting the setting. Once it's been decided that you are trying to fit something in for setting purposes then you have to ask yourself how to best get that theme across. So before trying to push miscasting ask yourself why you want that instead of having spells/abilities/whatever simply challenging a defense that the target has. What benefit are you getting from randomizing the ability of the player to even be able to use their abilities? If you are trying to fit a theme then you want to ask yourself what known insetting occurrences you're trying emulate and how best to represent them through whatever implementation scheme you have decided to use.

[owlassociate]

Well, in my example my goal was for magic to be dangerous, dramatic and chaotic but as a side effect it also turns miscasting into a resource mechanic.

[Foxwarrior]

Actually, maybe let me tell you about my D&D character, Mozilla Sparklemaster. He got like two or three of those "when you roll for a random effect on a table, roll an extra time" abilities, and then learned almost exclusively prismatic and other random effect table spells (and lesser restoration so he could use Blood Sorcerer on Reincarnate often enough to reincarnate all the enemies) so turns in combat would be like "let me designate a point and see what the AoE of my fire blast will be" or "let me see what effect my magic slot machine gives me this time" or "I blast the enemies with prismatic something or other"

Miscasting is about bungling something predictable, rolling on a table is about being barely in control to begin with.

[czernebog]

Another important aspect of balance is narrative agency. If a psyker character's action outcomes are "enemies all explode into a fine mist" or "nothing happens," then you can set probabilities so that a psyker who attempts to do their thing once a round has the same expected time to bring about a win condition as a guardsman with a lasgun who plugs an enemy once a round.

As already noted in this thread, merely matching expectations is insufficient for mechanical balance: a low-variance damage roll is worse at ending a fight in one hit. Just looking at expectations does not capture narrative agency, however. You need to look at the distribution of outcomes in the appropriate event space. When the psyker can occasionally clear the battlefield, it feels like they have a lot more agency than the schmoe with the rifle. Some of this may come from fluff (psychic powers or magic being cooler than "mundane" lasrifles or chainswords), but I'm willing to bet that players will treat someone with a "hail mary" power as having greater ability to control narrative outcome, even if the expectations show that, if the guardsman and the psyker were each to fight the same battle 1000 times, the guardsman would tend to win more often.

As far as miscasts go, giving the psyker an additional outcome along the lines of "all my allies explode into a fine mist" makes the disparity in narrative agency worse. A "make the party lose" entry in the miscasts table could be used to make the math work out better so that both the guardsman and the psyker have equal probabilities of a win or a loss, but the psyker now has two ways to end the fight, and one of them decisively decides the fate of the rest of the party. Miscast outcomes like "psyker is now a hostile Daemonhost" worsen the disparity between how a psyker can determine a battle's outcome and how other characters can.

[Foxwarrior]

A side note to czernebog's post: I think a game that uses a hail mary psyler as its narrative agency balance point is fundamentally more fun than one that uses a lasgun pinger as its balancr point. I'd hope that any lasgun-wielding mundane PCs would have the sorts of situational skills and gadgets that let them sometimes cause dramatic changes in the course of battle.

[Zinegata]

It doesn’t look balanced at first glance because the positive modifier is +3 while the negative one is -1.

Oh my.

Funny, because Gloomhaven has those kinds of “imbalanced” bonuses and yet the result is one pf the best-balanced games ever that keeps working despite having no DM to fudge results.

And thats because it doesn’t care about having balanced bonuses - it aimed to have balanced outcomes. In particular, Gloomhaven didn’t particularly care about having balanced HP or damage bonuses. It instead balanced itself around “how many attacks are needed to kill a monster”; because the game in turn was premised on the player characters having a finite number of attacks they can fire off in an encounter. If Gloomhaven wasn’t so finely balanced, the players would literally always lose because they’d run out of attacks before the AI ran out of monsters; or it would be so easy that they always kill the monsters long before they run out of attacks.

Instead, the consistent feedback is that the encounters are always at the fine line between victory and defeat, which makes the game interesting and never just a pushover for the PCs or a hopeless battle against an overwhelming foe.

Again, the issue here is you have a very limited view of balance. Balance is about having consistent outcomes, and outcomes are the result of multiple variables. Its not about consistent bonuses for any single variable.

If you balance based on outcomes, there is no need to play any “perception of balance” deception to begin with. Even a string of extremely good or bad rolls won’t wreck the game’s balance at all. You will instead always be at the sweet spot of what you want the players to experience.

[Zinegata]

Another important aspect of balance is narrative agency. If a psyker character's action outcomes are "enemies all explode into a fine mist" or "nothing happens," then you can set probabilities so that a psyker who attempts to do their thing once a round has the same expected time to bring about a win condition as a guardsman with a lasgun who plugs an enemy once a round.

As already noted in this thread, merely matching expectations is insufficient for mechanical balance: a low-variance damage roll is worse at ending a fight in one hit. Just looking at expectations does not capture narrative agency, however. You need to look at the distribution of outcomes in the appropriate event space. When the psyker can occasionally clear the battlefield, it feels like they have a lot more agency than the schmoe with the rifle. Some of this may come from fluff (psychic powers or magic being cooler than "mundane" lasrifles or chainswords), but I'm willing to bet that players will treat someone with a "hail mary" power as having greater ability to control narrative outcome, even if the expectations show that, if the guardsman and the psyker were each to fight the same battle 1000 times, the guardsman would tend to win more often.

Here’s the thing though:

The players are not wrong that the hail mary powers are in fact OP. Having a smaller chance of them triggering doesn’t “balance” it out. This is the single most common misconception about balance in game design as a whole; albeit it particularly afflicts tabletop rpgs.

And the reason for that is this: This “balancing” is premised on a truly random distribution of results. If you keep rolling a D20, this model presumes you will roll an equal number of natural 1s as you do natural 20s. If you roll 100 times, you expect to roll five 20s, and five 1s.

Modern game design no longer accepts this premise, because in practice the dice aren’t that consistent. Just because you rolled five 1s already doesn’t mean you will never roll another 1 in your next 95 rolls. Instead your next roll has the exact same chance as before - which is 1 out of 20.

So if you balance your game in such a way that it breaks if you roll that sixth 1, then your design isn’t balanced. Its not even consistent. Its just broken because you put your design in a position where it now has a 1 in 20 chance of breaking.

In the case of your example, you can make the math favor a Guardsman with a gun in the long run, but all you need is the first session to break in favor of the psyker to make everyone at the table confirm that your game is in fact broken.

So why even let that possibility occur in the first place?

Modern design is again about balancing and controlling outcomes. If you never want one player to hog all the spotlight in one session, then that balancing needs to be baked into the rules. Sure, the randomness might mean they might have more of the spotlight - but it should only be a modest and well-defined increase. Say, instead of 25% of the spotlight for 1 of 4 players, a string of good luck means at most they get 40% which still leaves everyone else a modest reduction from 25% to 20%.

The thing is, baking in these safeguards is a lot of actual work, which most designers don’t want to do. Its basically the “bug fixing” phase of a design, rather then the fun high ideas phase.

[Foxwarrior]

It doesn’t look balanced at first glance because the positive modifier is +3 while the negative one is -1.

Oh my.

Funny, because Gloomhaven has those kinds of “imbalanced” bonuses and yet the result is one pf the best-balanced games ever that keeps working despite having no DM to fudge results.

Sorry, I should clarify: I didn't mean "oh my, you correctly predicted what I thought 'looks balanced' looks like, but I'm unconvinced by your argument." I meant "oh my, it's fascinating and bizarre to see a statement drip with so much condescension while at the same time being so far off the mark. I wouldn't take the stance that symmetrical distribution of results on the probability curve of a single action was terribly relevant to the balance of entire characters relative to each other... but I guess you would (and still do, you've just taken the 'results' step one step further), so somehow you've gotten our two sides of this conversation confused somehow."

But maybe I'm seeing condescension where there isn't any, just because if I said it I'd feel the condescension dripping out of my mouth.

[Ice9]

I've played Gloomhaven, and while it's fun it's not the model I'd want for a TTRPG. It is tightly balanced, but that balance depends on:
* Actions are strictly constrained (as usual for a board game).
* Constant time pressure, like taking a turn to pick up some gold could cause you to lose.
* Balanced tightly enough that one person playing badly can make the whole party lose. So if you have a non-tactical-minded player, all your options are bad and are going to make them and/or other players feel bad. Same problem as other co-op board games like Pandemic.

None of which are desirable for a TTRPG. Also now that physically playing at the same table isn't a thing, those style of games don't really do anything that a co-op video game with voice chat wouldn't.

[Zinegata]

But maybe I'm seeing condescension where there isn't any, just because if I said it I'd feel the condescension dripping out of my mouth.

The only person being condescending here is you, but you already knew that since you're too busy imagining my condescension out of your own pretty obvious insecurity.

Why else would you assume the simple statement of facts to be condescending?

[Zinegata]

I've played Gloomhaven, and while it's fun it's not the model I'd want for a TTRPG. It is tightly balanced, but that balance depends on:
* Actions are strictly constrained (as usual for a board game).
* Constant time pressure, like taking a turn to pick up some gold could cause you to lose.
* Balanced tightly enough that one person playing badly can make the whole party lose. So if you have a non-tactical-minded player, all your options are bad and are going to make them and/or other players feel bad. Same problem as other co-op board games like Pandemic.

None of which are desirable for a TTRPG. Also now that physically playing at the same table isn't a thing, those style of games don't really do anything that a co-op video game with voice chat wouldn't.

I agree on the third point (though I would note that most DMs actually tend to throw weaker encounters at a party to begin with), but not on the first.

Because if you think about it, Gloomhaven's basically just Vancian casting taken to extremes.

Instead of spells being the only attacks restricted to a per/day basis, they make all attacks restricted on a per-day basis.

So Gloomhaven is basically an all-Wizard Vancian casting party.

And in any case, action economy is something that is very rarely actually well-considered in TTRPGs when they should be. A lot of problems in D&D and other similar TTRPGs actually stem from how badly the "core engine" is designed in terms of balance, and action economy is one of the main elements of this core engine.

[Previn]

Here's the thing, people who play spellcasters want to play characters who cast spells. Shitting in their Cheerios by having their spells not work or force them to roll on table 420: "hilarious" random mishaps would need to actually balance the game in order to justify itself and it never does, because it adds another set of moving parts to the equation.

Except not only does this exact thing appeal to noticeable group of players, it's why we have the pants on head stupid wild mage/magic every single DnD edition.

[MGuy]

There's an audience for everything no matter how dumb. I've never met anyone who can give me a good reason to roll people's attributes but I've also met people who have changed their mind about me GMing for them because I refuse to have people roll for their stats/HP.

[Omegonthesane]

Here's the thing, people who play spellcasters want to play characters who cast spells. Shitting in their Cheerios by having their spells not work or force them to roll on table 420: "hilarious" random mishaps would need to actually balance the game in order to justify itself and it never does, because it adds another set of moving parts to the equation.

Except not only does this exact thing appeal to noticeable group of players, it's why we have the pants on head stupid wild mage/magic every single DnD edition.

Generally the wild mage is an optional archetype or PrC rather than a rule that applies to all players and can't be opted out of. Which is to say, there's a kind of player that likes it when things happen outside of the mandate of anyone on the table, rather than the threat os uch being an effective balancing point.

[OgreBattle]

There an implementation of Wild Magic that's not dumb then?