Power, Frequency, Versatility, Reliability -- four ability balance metrics

[Vahnavoi]

@PhoenixPhyre, KorvinStarmast: quite often, the most useful number for comparisons would simply be product of these measures.

Power, frequency and reliability are easiest to track, because often you can get the relevant number directly from the game system. Versatility is the hardest, for the simple reason that game systems rarely control for scenario. If an ability can be used for situation X but the question of how often X comes up is unanswered or, worse, unanswerable, you can't get a good measure.

@Quertus:

When you're mathematically balancing abilities, you do not account for player skill, or lack thereof, at that stage. At most, you can do that after several passes of playtesting, but trying to do so at an earlier or abstract stage makes the problem recursive in a really nasty way.

For example, if somebody's failing to use their 1/day abilities due to psychological loss aversion, the best answer to that is not to muck with the game system, it's to point out the player is following a bad strategy.

[Pauly]

The 4 variables assume a PvE approach. But some tables /games are more of a PvP approach. In which case a 5th variable “counter-ability” is needed.

“Counter-ability” is how easily a player can counter the effect. This can be considered part of reliability in a PvE approach where the NPCs/monsters usually aren’t actively seeking to avoid effects/exploit weaknesses. However in a PvP environment both players are actively seeking to counter the other’s effects it becomes an important consideration.

[RandomPeasant]

Power, frequency and reliability are easiest to track, because often you can get the relevant number directly from the game system.

Do they? Consider balancing the abilities Accurate Strike and Powerful Strike. Accurate Strike gets a to-hit boost. Powerful Strike gets a damage boost. Those numbers are "from the game system", but they depend on the opposition people are going to face in the same way that you note versatility does. Or how about the trade-off between an ability that is at-will and one that is X/encounter. That depends on how long encounters take, which is again something that is not an inherent property of the system. The properties of an ability are only meaningful based on the context in which that ability is used. As I said originally, this sort of ivory-tower number crunching is simply insufficient to produce balance. You have to define your balance point.

For example, if somebody's failing to use their 1/day abilities due to psychological loss aversion, the best answer to that is not to muck with the game system, it's to point out the player is following a bad strategy.

Experimentally, telling people they're behaving irrationally is a really bad way to get them to stop behaving irrationally. A design that is mathematically beautiful, but depends on players acting in ways they won't psychologically, is not a good design.

[Quertus]

@Quertus:

When you're mathematically balancing abilities, you do not account for player skill, or lack thereof, at that stage. At most, you can do that after several passes of playtesting, but trying to do so at an earlier or abstract stage makes the problem recursive in a really nasty way.

For example, if somebody's failing to use their 1/day abilities due to psychological loss aversion, the best answer to that is not to muck with the game system, it's to point out the player is following a bad strategy.

While I do not disagree, I'm more pointing out how one cannot simply *stop* at that step. Or, rather, one *should not* do so.

Most people I've known IRL are unwilling or unable to comprehend things at an appropriate depth (without liberal application of the (verbal) clue-by-four, and often not even then). So I'm just doing due diligence, and trying to set the expectation at "this isn't the only step - if you can't go through all the steps, don't bother starting".

Seemed very appropriate given that the OP asked, "is this complete?", and the answer is a resounding "no" - and not just in the dimensions one might expect.

I'm working to make sure anyone expecting 2-dimensional thinking realizes that the problem is more appropriately approached as 5d Wizard Chess.

And you are very wrong regarding what is the best strategy in that scenario.

[OldTrees1]

For example, if somebody's failing to use their 1/day abilities due to psychological loss aversion, the best answer to that is not to muck with the game system, it's to point out the player is following a bad strategy.

The best answer is to realize the 1/day ability is not valuable to the player. Ask them why. If it is due to loss aversion, then the best answer is replacing the ability with one that better fits the player preferences.

The best answer is not to try to convince your player to have a bad time and like it. Pointing out that the psychological loss aversion is causing a bad strategy will do nothing (they already knew that duh!) or will make them feel even worse (the new player now understands their aversion has a negative impact) or convinces them to pretend they don't have the aversion (aka suffering in silence because the GM told them to shut up).

[Cluedrew]

On Player Psychology: Ignoring player psychology is kind of like ignoring the fact you will have players. If some habit is required to play your game (or similarly if a habit breaks it) that is OK but you should be prepared to lead them to (away) from it. And if its something deeper than a habit you might not be able to.

On Dimensions: I think "oh its actually a 5d problem" is the wrong way to think about adding detail. The problem is more like how describing a place's latitude and longitude (2 dimensions) doesn't actually tell you how you to get there.

For instance flexibility could further be sub-divided into two parts: flexibility across axis you can predict and flexibility across axis that are effectively random. This is a fairly important distinction, especially for anything you prepare ahead of time, but do I think that that this split should be described in a high level system like this? Not really. For one things don't always stay in one box or the other, they can shift back and forth due to a variety of situations and even other abilities. How long you have to find out and how long it takes to set-up/swap options are dimensions, but are hard to pin down at a high level.

OK how about fire damage? Does changing to fire damage increase or decrease an abilities power? Assuming any sort of elemental system in this game, well it depends on the enemies you are facing. Are they venerable to fire or resistant to it?

[RandomPeasant]

OK how about fire damage? Does changing to fire damage increase or decrease an abilities power? Assuming any sort of elemental system in this game, well it depends on the enemies you are facing. Are they venerable to fire or resistant to it?

And it depends on the abilities characters can have. Is there a feat or talent or something that adds extra damage to fire abilities? One that adds some kind of rider? Is there some kind of "use any fire power" ability that would gain versatility by making a new fire-based ability? The idea that you can do anything useful by breaking things down onto a four-axis (or N-axis for any reasonable N) scale is something I'm deeply skeptical of. Ultimately, the only way to achieve balance is to define a balance point, then test and iterate until you reach that balance point. That's a lot of work, but designing and manipulating complex systems so they do what you want is a lot of work. It's way doctors, lawyers, and engineers make so much money.

[Quertus]

I think people may have taken my "dimensions" comment a bit too seriously / literally.

Let's try this instead: this isn't a BDF problem, that can be solved by hitting it harder. It requires thought, adaptation, customization, and conversation: 5d Wizard chess, as played by a charismatic Bard.

You literally cannot make balance at the system level for something like, say, 3e D&D, when you've got tables that will nerf the Monk, and tables that only count damage as contribution, and tables like mine, and those one would expect of the Playground. Whichever balance metric you choose, everyone else will know that you were wrong, and that you built a horribly unbalanced system.

And they will be right. At least as right as we are claiming that 3e isn't balanced.

Thing is, 3e provided the tools for people to make their own balance. If you only count damage as contribution, you can say, "make a character with roughly X DPS", and everyone can build that. Even if I might build a totally OP Tainted Sorcerer Arcane Spellcaster BFC build, whose familiar / animal companion / summons actually "contributes".

Balance to the table. Because you cannot actually balance anywhere else. Attempting to do so will result in games that are less balanced, less capable of balance, and likely less capable of having meaningful conversations about balance, for any other paradigm of "Balance" other than the one you choose for your system.

Unless, of course, you want to try to create one balance fix for 3e that will satisfy me, tables that nerf the Monk, tables that only count damage as contribution, and the stereotypical Playgrounder table, all with one set of rules.

Never mind the many diverse effects of "player > character" (including psychological, playstyle, preference, and role-playing ones).

Don't get me wrong, I absolutely *love* statistics like these. Trying to create good rules for Power, Frequency, Versatility, and Reliability, and then looking at various characters through those lenses? Great fun, better than cats.

But it's just not a tool one can use to balance a system by more than one paradigm - and certainly not something one can use to balance a system blind for the unknown paradigm of a particular table.

But if you want to talk about dimensions? You've got the dimension of "psychology", to account for loss aversion, risk aversion, hoarding, etc. You've got the dimension of "spotlight sharing", measuring the distribution of how often and for how long your character gets to participate / shine / solo, vs sit there and twiddle your thumbs (and some individual variation on how individuals value different distributions). You've got the dimension of "playstyle", from "kick in the door" to "5d Wizard Chess", CaW vs CaS (vs…), D&D as dungeon crawl vs murder mystery vs horror vs comedy, etc. You've got the dimension of "values", from only counting damage towards contribution, to (over)valuing the Monk's diverse array of at-will abilities gained every level, to counting how many one-liners they got to deliver.

The way to make a game more balanced is not to say that people *must* play a heal-bot Cleric / blaster Mage / golf cart Fighter, *must* play a vampire filled with angst as their humanity slowly drains away, *must* stop at level 6, *must* have a kryptonite that makes them unable to play the game.

IMO, the best way to balance a game (or, at least, one as robust as 3e) is to make even *more* wildly unbalanced options that the players can mix and match, allowing them to both create their concept, and balance to the table, for whatever definition of balance their table uses.

And then these fun metrics can be used when players need to communicate: "I built a DPS UMD skill monkey Rogue, and even set aside some of my starting WBL for adaptive mid-game purchases, expecting I was set for power and versatility, but the samey nature of the encounters, and lack of downtime, has left me unable to use more than a single repetitive, suboptimal tool from my toolkit, participating less than the übercharger's Legacy Weapon's permanent summons".
.

[Cluedrew]

Actually to cut through a lot of the previous discussion (also, to be quick), I thought I would mention why I like of liked this system: I'm not sure if there is a good way to measure each axis or combine them, I do think there is a value in having the major ways something can be good defined. First because its a good communication tool to have important terms shared between people. Second as a kind of sanity check, if something is coming in unusually high or low on a bunch of these scales than you should have some special note explaining what is not captured in these systems that makes it balanced. So its a good outlier check. Or if something radically changes how other things rate then it should also be examined very carefully.

[Quertus]

its a good communication tool to have important terms shared between people.

Strongly agree.

Second as a kind of sanity check, if something is coming in unusually high or low on a bunch of these scales than you should have some special note explaining what is not captured in these systems that makes it balanced. So its a good outlier check. Or if something radically changes how other things rate then it should also be examined very carefully.

This feels… overloaded. So I'm gonna babble for a bit.

Spoiler: babble

Show

One thing that's been on my mind for a bit is, what kind of characters would be balanced / work well together.

It's tricky, conceptually, because many characters live in quite a range. D&D characters go from zero to slaying gods. Jedi go from farmboy Luke to TK flattening armies or Force storming entire fleets. Whereas Conan… is Conan. And a CoC character will never get over not being the center of the universe.

A D&D Wizard with "all the spells" (or, you know, a D&D Cleric) can solve any problem… tomorrow. Whereas a WoD Mage can invent and cast new spells on the fly. Even if they're a novice, and are only capable of Forces/Prime effects.

A D&D Wizard can learn a new spell from a scroll in hours, or research an entirely new spell from scratch in weeks or months. A WoD Mage invents spells in… as long as it takes them to think of it. A Matrix character can learn new skills (like Kung Fu or helicopter pilot) in moments.

A Star Trek character gets communication, sensory, blasting, and Teleportation powers, generally at-will. Oh, and a star ship. And matter replication. And probably more.

A superhero character… can live almost anywhere from zero to deity… but doesn't usually progress the way a D&D character can.

A Battletech mech eats peasants and soldiers for breakfast, but is pretty useless outside combat, and struggles to fight stereotypical Dragons.

A Time Lord can, as the name suggests, travel through time. And fly and teleport, and other nifty tricks, like self-resurrect. But they're pretty much just "a guy".

Naruto ninjas, I'm told, have quite a huge variance in their power, and most are not just 1-trick ponies. How much of their power is "level" vs "bloodline", and how quickly they can gain power, though, I don't know enough to answer. (Although the phrase, "younger than you, stronger than me", or something like that, seems to have been spoken early on…)

IIRC, early edition D&D Wizards can only cast 1 spell per minute (as rounds were 1 minute long), putting them at a severe disadvantage vs most beings. IIRC, my first Exalted and Scion characters each acted once per second, one high-level Heroes superhero I built acted twice per second, and some Battletech mechs can make an even higher rate of attacks than that.

And then they're things in D&D (including the Hecatoncheires) that can make 100+ attacks per round. Or infinite.

Harry Potter Wizards… seem to run on "at will, effect vs counter" logic, making them crazy powerful vs things that don't have counters.

And I'd still expect a sniper to one-shot almost every single character I've just described.

I don't know if there's anything useful in there. But I think I care about Growth Range, Growth Rate, Action Rate, and Adaptation Rate as valid ratings, and as balance concerns.

[MoiMagnus]

Reliability -- when you use this power, does it <do thing> all the time? Or just some of the time. A power that requires some form of attack or defensive or casting check is less reliable than one that just happens; a power that bypasses immunities or "just works" is more reliable than one that has an element of random chance.

Slight subtlety, reliability should be about "being actually useful", not just doing the thing it is supposed to do.
Let me explain myself:
A power saying "+2 to your next attack roll" is significantly less reliable than a power saying "when you would fail an attack by 2 or less, succeed instead". Because literally 90% of the times the first one is useless [either you failed despite the bonus, or you would have succeeded even without the bonus] while the second one is always useful when used.

[PhoenixPhyre]

Slight subtlety, reliability should be about "being actually useful", not just doing the thing it is supposed to do.
Let me explain myself:
A power saying "+2 to your next attack roll" is significantly less reliable than a power saying "when you would fail an attack by 2 or less, succeed instead". Because literally 90% of the times the first one is useless [either you failed despite the bonus, or you would have succeeded even without the bonus] while the second one is always useful when used.

I think that's a valid point.

There's subtlety around timing--for example the 5e Paladin's Divine Smite ability (choose to trigger it on hit) is quite reliable--every time you use it, it adds damage[1]. Compare that to the paladin's smite spells (e.g. Searing Smite), which are bonus actions with concentration that take effect on your next hit. Which means that if you use it and miss both times that turn (assuming level 5+), you could get hit and lose concentration before you get to actually benefit from the spell. In return, the smite spells generally do more than just damage. Is it enough to make them worth it? Meh, opinions vary.

Spoiler: Applying the framework to DS vs Searing Smite

Show

This is a case of a clear tradeoff at the pair of abilities level (using searing smite because that way I can just do 1st level comparisons):
1) Power: 9 (2d8) radiant damage vs 3.5 (1d6) fire + CON save each turn or 3.5 (1d6) fire damage. Target can also take an action to end it. To make Searing Smite stronger, it needs to burn for 2 rounds (totaling 10.5 damage). Generally, however, prompt damage is better than delayed damage, even if the delayed damage is guaranteed (which this is not).

2) Frequency: At level 2[1], they have the same frequency. You can use them each once per turn, both cost the same resource. At level 5+, Divine Smite is more frequent (you can use it on each hit, including hits not on your turn such as OAs) at the cost of more resources.

3) Versatility: Push. Basically both do damage and that's it. Same targeting restrictions, no other conditions. One tiny benefit to Searing Smite is that by doing recurring fire damage, you can (in theory) keep a troll from regenerating for multiple turns with one ability. One tiny benefit to Divine Smite is that you can blast through a zombie's annoying Undead Fortitude ability.

4) Reliability: Divine Smite wins hands down. You'll rarely, if ever, use Divine Smite and do nothing[2]. Searing Smite costs concentration, which is a big cost, plus can be utterly wasted. Divine Smite also has the more reliable damage type--many more things resist or are immune to fire than to radiant. Not only that, but there's an additional save (and the generally-best save around for most monsters) that can completely negate the additional damage. Without that, Searing Smite stinks.

Verdict: DS >> SS. In each category, divine smite wins or ties with searing smite, and is significantly better in reliability and (at higher levels) frequency.

[1] paladins don't get spell slots or Divine Smite at level 1.
[2] The one exception is if you misjudged the enemy's HP and the base weapon damage would have killed it anyway. But this exists for Searing Smite as well, so it's not a significant difference.

Another example from 5e about timing is the Protection fighting style (when an enemy attacks an ally within 5 feet of you, use your reaction to impose disadvantage on the attack). The default interpretation is that you have to do this before the attack is rolled. And that's not very reliable--if the target has high AC anyway, it's likely the attack would miss, so disadvantage doesn't do much. If they have low AC, then disadvantage isn't a huge help. There's another interpretation (which is the one I play with as a house rule) that says you can do this after the roll, but before damage is applied. That's way stronger--you can save it for times where the first roll hit. Or even turn a crit into...probably not a crit. That second interpretation turns a really lack-luster ability (one generally not worth taking) into one that's quite good (as good defensively as the archery style is offensively). Which is why I use it at my tables.

Both interpretations have the same Power -- impose disadvantage on one attack roll. They have the same Frequency -- once per turn (costing a reaction) and the same restrictions on when you can activate it (when an ally within 5 feet of you is attacked and you have a shield). They have the same Versatility -- both can only be used for that one thing (reducing the chance an ally takes damage). But they have radically different Reliability, and that makes a huge difference in their overall usefulness.

[LibraryOgre]

As an aside, in the old Bard's Tale Construction Set, I got tired of playing at level 1. So, I created a monster with several overwhelmingly powerful attacks... 100d100 and things like that... but then gave it a mediocre AC and HP, and set its AI so it never attacked. Roll in, kill it quickly, and gain a bunch of XP right off the bat.