What makes a character competent?

[Telok]

It certainly can. But there are diminishing returns. Going from fighting one orc to fighting ten orcs feels like character growth. But going from fighting ten orcs to fighting twenty orcs doesn't, at least not to the same degree. Similarly, going from fighting generic orc raiders to fighting Grummsh's elite Bloodsworn feels like a step up, but if you reveal that there are even more elite Goresworn, that feels stupid. Things need to change in substantive ways for it to feel like character growth is occurring, and that definitely can happen in combat.

In addition there are systems that, correctly or not, have in effect or appearance a "level appropriate" design that keeps the things characters roll for at roughly the same success rates. This tends to make it feel that the characters are just the same, but with more adjectives or just another zero on the end of your numbers.

I find a good example in th Starfinder space combat stuff. The party spaceship levels up with the party, as do the space fights. The skill DCs are all things like 15+(1.5 * level). So you get DCs of 16 at level 1 and DCs of 45 at level 20. A character might start at +8 and end up at +39. That's a highly competent character by the system, but they really only "grew" from needing an 8+ to a 6+ on a d20. It's also the character's main attribute, boosted at every opportunity, maxxed skill, and some other bonuses probably from the character class and background. And if they stray from the approved stat-class-skill paradigm they'd get less competent as their levels increased (generally if it's an off-stat it starts about 2 less and ends about 3 or 4 less, if it's not a skill with a class bonus it loses 5 to 6 as the levels increase, and not a class skill is -3 all the time).

One benefit of that system would be that a chatacter is, by 10th level, significantly better at their chosen skills than a random CR 1/2 space goblin. But the CR & encounter system says that you don't see those any more at 10th level. You see CR 10 +/- 3 <fancy adjective> space goblins (or whatever monsters) that got the same +1.5 * level skill bonuses that the the other skill DCs got. Go all the way to 20 and you get super ultimate king elite space goblins with +39 skills like yours.

While such PCs will generally feel competent if they are the only person rolling, the fact that "being really good at something" is capped at the level of npcs who aren't bad at it makes you really feel the Red Queen's Race that you're in trying to keep ahead of slowly becoming incompetent. So your character is competent, if not significantly any better than any NPCs you meet, for as long as they keep upgrading themselves to the max at every opportunity. But you never really feel any progress in anything because you're locked into the "level appropriate" paradigm.

[NigelWalmsley]

The skill DCs are all things like 15+(1.5 * level). So you get DCs of 16 at level 1 and DCs of 45 at level 20. A character might start at +8 and end up at +39. That's a highly competent character by the system, but they really only "grew" from needing an 8+ to a 6+ on a d20.

The skill system is one area where I think Bounded Accuracy (or something close to it) might be good for the game. The DC to Bluff the super-elite Imperial Guard doesn't need to be any higher than the DC to Bluff the local town guard, and if it's not your Bluff expert can have real progression over the course of the game, rather than going from rolling +6 against DC 16 to +20 against DC 30.

But the CR & encounter system says that you don't see those any more at 10th level. You see CR 10 +/- 3 <fancy adjective> space goblins (or whatever monsters) that got the same +1.5 * level skill bonuses that the the other skill DCs got.

This is one of the reasons that I think having periodic fights where the PCs just get to cut loose against an army of chaff can be really good for providing a sense of progression. If your 3rd level fight is an Ogre, and your 10th level fight is a Fire Giant, that can feel like running in place. But if your 3rd level fight is an Ogre, and your 10th level fight is thirty Ogres, that provides a very visceral feeling of progression. This is one of the reasons I think 4e's Minions were a bad design choice, as they mean that the Ogre hoard monster is fundamentally different from the Ogre miniboss, preventing that sense of progression.

[Cluedrew]

I can lift some number of bricks. Let's pretend Superman can (only) lift the same number of tanks.

I described the cloud giants as scaled up orcs for a reason. There is a fundamental difference between bricks and tanks. Similar things can be said of the situations where lifting bricks is useful is very different from a situation where tanks is useful (tanks don't come up in construction very much). Honestly I don't know if the same can be said for orcs and cloud giants. What separates a cloud giant from a scaled up orc? See Telok's post on level appropriate encounters.

[OldTrees1]

The skill system is one area where I think Bounded Accuracy (or something close to it) might be good for the game. The DC to Bluff the super-elite Imperial Guard doesn't need to be any higher than the DC to Bluff the local town guard, and if it's not your Bluff expert can have real progression over the course of the game, rather than going from rolling +6 against DC 16 to +20 against DC 30.

For opposed checks, maybe. Although you will probably decrease progression rather than increase it. Bounded Accuracy (and similar) put a limit on how much you can progress. When you are at the level where you are bluffing the Imperial Guard, you might also occasionally bluff the local town guard. How much progress do you want leveling to have on lower level checks? Bounded Accuracy caps that progression.

For unopposed checks I do not see a reason to cap the progression. There is real progression from being able to do DC 10 tricks occasionally to being able to always succeed on DC 15 tricks.

[NigelWalmsley]

What separates a cloud giant from a scaled up orc?

The Cloud Giant has some level of elemental magic. In 3e it's kinda crappy, but in 5e he gets Control Weather and Telekinesis. The Cloud Giant is also Huge. So it's a giant monster you fight in the middle of a thunderstorm. Seems reasonably different from Orcs. And if you were writing a new game, you could easily tweak that some more to give him flight, or something like Call Lightning.

For opposed checks, maybe. Although you will probably decrease progression rather than increase it. Bounded Accuracy (and similar) put a limit on how much you can progress. When you are at the level where you are bluffing the Imperial Guard, you might also occasionally bluff the local town guard. How much progress do you want leveling to have on lower level checks? Bounded Accuracy caps that progression.

Sure. But my point is that a lot of skill checks genuinely don't need progression. There's no reason that it should be harder to convince the king of the elves to help you because he is personally harder core.

For unopposed checks I do not see a reason to cap the progression. There is real progression from being able to do DC 10 tricks occasionally to being able to always succeed on DC 15 tricks.

Sure, but is there a need to have +40 checks and the associated DC 50 tricks? Bounded Accuracy doesn't have to mean no progression, it just means that progression stops at some point.

[OldTrees1]

Sure. But my point is that a lot of skill checks genuinely don't need progression. There's no reason that it should be harder to convince the king of the elves to help you because he is personally harder core.

It sounds like you are saying "Higher level characters (including NPCs) are not necessarily better in everything". The king of the elves is not necessarily harder to convince, because their DC is not necessarily higher. Belkar did not get better at convincing others despite gaining many levels. If that is what you are saying, then it seems tangential to bounded accuracy. Bounded accuracy changes nothing about the cases where neither DC nor ability increased. Bounded accuracy only changes the effect of increases.

Sure, but is there a need to have +40 checks and the associated DC 50 tricks? Bounded Accuracy doesn't have to mean no progression, it just means that progression stops at some point.

Is there a need for DC 50 tricks? Probably not. Depends on the size of your RNG (d10? d20? d100?) and on what your other progression systems (Monsters, Spells, etc) look like.

Bounded Accuracy, as I understand it, is having the change in modifier be dwarfed by the size of the RNG. For example I would not describe 3E skills as bounded accuracy because the change in the modifier was 1.5x to 2x the size of the RNG.

That expands the question, what does the ratio of "change in modifier" : "size of the RNG" say about progression in that system? AND how does that progression compare to other progressions you have across those same levels. Well that ratio tells you how long it takes to master a task. Now look at the other areas of progression or ask yourself, how long does it take to master a task in another area of progression? How long do I want it to take in this subsystem?

Oh! But you might have noticed a hidden variable. How much variance in ability do you imagine characters having? Is it "crit fail you baked a poisonous gas" to "crit success the god of baking fell in love with you" or is it "with a 1 you made only 11 loaves of bread. You will need to make another batch" to "with a 20 you made 13 loaves of bread. You enjoy the extra loaf.". That degree of variance can impact how big a ratio you want in "change in modifier" : "size of the RNG".

For the variance I use, I think a ratio of 1.5 (for non opposed checks) over 20 levels (aka 1d20+5 to 1d20+35) works well compared to monsters ranging from Goblins to Red Dragon fight literally inside a lake of lava. For opposed checks I want a smaller ratio.

------------------------

To phrase it all another way (2nd way is usually clearer):
0) You have a skill system you are designing and the RPG has other forms of progression too.
1) Take another progression system and make a continuum from first to last level.
2) On that continuum place some level appropriate examples from the skill system. Remember the purpose of mentioning the other progression system was to help you create level appropriate examples with that other system anchoring & calibrating your expectations.
3) Next draw on the continuum the minimum and maximum effectiveness you want a 5th level specialist in the skill to have. Can a nat 20 hit above a 10th level usage? Can a nat 1 hit below a 1st level usage? The longer this line segment the higher the variance your RNG will represent. You can make it narrow to have very consistent skill usage or long to have wild crazy luck.
4) Now take that line segment and compare it to the total length of your continuum. What is that ratio? For me that ratio falls around 2.5x so I have the modifier grow by 1.5x the size of the RNG so I can fit 2.5 line segments on that continuum.

[icefractal]

I don't find that Bounded Accuracy helps with this, personally. It's better for the opposite purpose really - making sure that low-level / unskilled people still have a chance in most cases.

Personally I would say that if you want to emphasize "Bob is really good at lockpicking", then you want both of:
1) Bob can nigh-effortlessly pick locks that novice lockpickers can handle at all.
2) Bob can, not guaranteed but with a reasonable chance of success, pick locks that other people have no chance to succeed at.

Bounded accuracy really doesn't do the latter - most things can be done by a lucky novice, and anything that can't is going to be very difficult for even a master.

What Bounded Accuracy does do is support both "DCs are what they are, they don't scale by the PCs level" and "PCs always have a chance to fail the rolls". But I'm here to contend that the second one isn't necessary or even desirable:



It's totally fine and not a problem with the mechanics if PCs can auto-pass many rolls related to their area of expertise.

I mean, this is what spells do. Fly auto-passes pits or sticky terrain. Teleportation auto-passes doors. And while people have different ideas about how easily this should be achieved, it's usually not to the extent of "at no level should a spell ever bypass an obstacle". If I wanted to play a low-power game where characters stay in normal human limits and magic is subtle, there are many, many options for that which aren't D&D.

Ditto in combat really - a high-level Cleric is just not threatened at all by a squad of ordinary zombies, and that's fine.

So if at a certain point, one of the PCs is "the master of unlocking" and just auto-picks 99% of locked doors the PCs come across, that's completely ok. It's not a problem that needs to be solved. There are still the many other skills and attributes where they aren't godlike to make them struggle, and if you're doing a spotlight arc you can introduce the Mercane League of Merchants interdimensional vault with truly godlike locks. Not every single skill check needs to be the source of suspense.

[Quertus]

Bounded Accuracy, Starfinder, Storm Giants, oh my.

So, let's say you're trying to be a Simulationist, and model Reality.

You would need to understand how many "steps" / "tiers" there are for any given skill, and just how much a member of each tier should outclass the preceding tier.

So, let's suppose that, for chess, we look at everyone from noobs to the best grandmasters. Let's pretend that we find 5 people who sit exactly at the line of "can utterly crush the person beneath them 99/100". In 3e, with the "taking 10" mechanic, that's +10 to the roll for each tier, for a +0 to +40 expected range of "chess skill" to simulate IRL. If the skill is "roll once" / if we are measuring at what we want the "atomic" level of a single skill check to represent.

Or, you might find it's better modeled as 7 tiers with a 75% success rate, do math to find that's a +X skill difference, and expect a 6X range in skills. If the skill is "roll once" / if we are measuring at what we want the "atomic" level of a single skill check to represent.

Repeat this for MtG, and Fort Night, and hide & seek, and "take this stone from my hand", and poker, and writing instructions that people can follow, and obstacle course driving, and skiing, and thermonuclear astrophysics, and anything else where you can objectively test "better". See what patterns develop in measurable win rates, and determine how to model that with your skill system.

Personally, I doubt that 5e Bounded Accuracy will model much IRL, if we actually did the math.

Or you might say, forget realism, let's make this a good game.

OK, but… what *does* make for a good game?

I'll argue that a "level treadmill" of reskinned foes does not, and that "hire a horde of Hobos for pennies, because they give smarter answers than the genius expert" does not, either, for most genres / feels of games.

So, what does?

I am of the school of thought that the default desire should be "get better people" rather than "get more people", that the PCs should be better at solving most problems in their area of expertise than the more numerous townsfolk / hirelings / redshirts. To facilitate PCs accomplishing things themselves being the default, experts should be better at solving problems than even large numbers of less-skilled individuals.

I like the progression of one Orc to 10 Orcs or an Ogre to 10 Ogres or a Troll and so on. However, while the King's Army may be a boon against the army of Orcs, those soldiers - who cannot anticipate lightning and dodge raindrops, who cannot walk on clouds or outrun thunder - should be useless even in functionally infinite numbers when facing a Storm Giant. Yes, the Giant has bigger numbers, but it should also be playing a fundamentally different game. Just like the high-level PCs.

Now, you may disagree with my assertion that no number of fighters can take out a fully operational battle station the size of a small moon, but remember: no ship that small has a cloaking device. That battle station ought to be able to play completely different games, where only the tide (or the bandwidth requirements of its occupants' Twitter posts) will tell of its presence.

[kyoryu]

So, let's suppose that, for chess, we look at everyone from noobs to the best grandmasters. Let's pretend that we find 5 people who sit exactly at the line of "can utterly crush the person beneath them 99/100". In 3e, with the "taking 10" mechanic, that's +10 to the roll for each tier, for a +0 to +40 expected range of "chess skill" to simulate IRL. If the skill is "roll once" / if we are measuring at what we want the "atomic" level of a single skill check to represent.

I believe you cannot Take 10 in situations where there is stress... aka any actual conflict with opposition.

Or, you might find it's better modeled as 7 tiers with a 75% success rate, do math to find that's a +X skill difference, and expect a 6X range in skills. If the skill is "roll once" / if we are measuring at what we want the "atomic" level of a single skill check to represent.

Math aside, this is actually how most rating systems work, with a bit more extra math to adjust skills.

Personally, I doubt that 5e Bounded Accuracy will model much IRL, if we actually did the math.

I'm not sure it needs to. D&D, at the end of the day, does not promise to simulate everything. It's about heroic adventurers doing heroic things. And while 3.x edged into the "generic system" territory, it frankly did it pretty poorly.

Or you might say, forget realism, let's make this a good game.

Indeed.

OK, but… what *does* make for a good game?

I'll argue that a "level treadmill" of reskinned foes does not, and that "hire a horde of Hobos for pennies, because they give smarter answers than the genius expert" does not, either, for most genres / feels of games.

The outer loop (level treadmill) is much less important than the inner loop (combat and encounters). Also, since the game is aimed at "heroic adventurers", hiring a horde of hobos is kind of not the issue, since the game is player focused.

IOW, you're still applying simulationist logic even when attempting to talk about the game from a more gamist POV.

So, what does?

I am of the school of thought that the default desire should be "get better people" rather than "get more people", that the PCs should be better at solving most problems in their area of expertise than the more numerous townsfolk / hirelings / redshirts. To facilitate PCs accomplishing things themselves being the default, experts should be better at solving problems than even large numbers of less-skilled individuals.

And they are, essentially. Now, it's pretty true in 5e that an army of X size composed of Y ability individuals will be equivalent to a party of 4-6 adventurers of N level and a particular class makeup. That's been true all along - the only thing that has changed is the values of X, Y, and N.

But even if said army exists, then 4-6 adventurers are a better choice - they're more mobile, less obvious, and require less concern about logistics.

I like the progression of one Orc to 10 Orcs or an Ogre to 10 Ogres or a Troll and so on. However, while the King's Army may be a boon against the army of Orcs, those soldiers - who cannot anticipate lightning and dodge raindrops, who cannot walk on clouds or outrun thunder - should be useless even in functionally infinite numbers when facing a Storm Giant. Yes, the Giant has bigger numbers, but it should also be playing a fundamentally different game. Just like the high-level PCs.

And here we get to the real issue, since "anticipating lightning and doging raindrops" is not something that is in any way "simulationist". You have a preference for, essentially, superheroic level play, and the power level that is associated with it (which is, realistically, almost orthogonal to being "competent"). Bounded accuracy ruins that for you since it means that rather than the pure power fantasy of 3.x, you end up with a more grounded, but still frankly kinda extreme, power fantasy.

The issue isn't really that you want an infinite number of hordes to die against a storm giant. You want a PC to be worth an infinite number of "normal" people.

And that's okay. But it's not universal. To some people, that's a feature, not a bug.

Now, you may disagree with my assertion that no number of fighters can take out a fully operational battle station the size of a small moon, but remember: no ship that small has a cloaking device. That battle station ought to be able to play completely different games, where only the tide (or the bandwidth requirements of its occupants' Twitter posts) will tell of its presence.

I do disagree with your assertion - or, more accurately, your presumtpion that it is the ideal mode of play.

You seem to be falling into the trap of trying to prove your preferences. You don't need to, and you can't (since they're just preferences).

[Telok]

So I've been thinking about this. The best way I can come up with to quantify 'compentence' is to phrase it as a question: "Would you trust the character to <activity>." Where the <activity> is something like "be a lifeguard at the local pool for your children", "represent you (legit innocent) in a court of law", "not harm you during a circus knife throwing act", "tame & train a lion for you", or "hunt to provide food for your family". I chose this format because things like "competent warrior" or "competent negotiator" are too general to quantify and "competent to swim the English Channel" or "competent to shoot a 50 yard target with a 30# pull compound bow" feel too specific to be useful. The best I came up with is an activity that, as presented, involves you and/or something you care about and has a potentially serious penalty for failure but where most people have a reasonable agreement of the expected level of success (pool lifeguards are expected to keep drownings to a negligible number, knife throwing acts almost never end with a knife in the eyeball).

Most of the common systems we all know don't have any problems creating characters that are competent at basic combat. They might not always produce fun characters for all combats, or characters that are more competent in anything but the most basic combats, but in general the characters in modern systems all pretty much manage the basic level of combat that the system presents. Granted this may look like "the party full of fighters has to be locked in a small room with the dragon and just spam basic attacks because nothing they do really affects critters that much bigger and stronger than them" but if that's the basic combat format that the system presents and the fighters manage to get the expected outcome (some systems do expect them to lose, there have been plutonium dragons) then that's working as intended.

Therefore the questions about competence are pretty much always about non-combat stuff and I think, most often, it ends up depending more on how well the system/books communicate their expectations and methods than it does on the actual dice mechanics. The result of the mechanics of 1d20+3 vs 10, 1d20 under 18, 6d10 keep 3 vs 20, 13d6 want 3 or more at 5+, and 1d6 to roll 2+, are all about the same. They're all roughly a 15% +/-2% failure rate. Which, frankly, is pretty terrible if 1/6th of the time the result of the lifeguard getting out of the chair because someone is horsing around results in dead bodies floating in the kiddy pool (Paranoia games excepted of course, that's what clones are for).

So our first step is what are you rolling for. Is the character rolling to be able to climb, or for how fast they can climb? For being able to tame a lion, or for how well it's trained? Some things we expect a novice to be able to do, children can often climb a fairly large range of things even before they master language. Others we expect to require training in order to attempt anything at all, like flying jet airplanes, or advanced mathematics. Then comes the number of rolls and the success rates. Are you rolling once to hunt for food for the entire month at a 15% fail rate, or is it rolling every couple of hours at a 50% fail rate to find a week worth of food? Structures like X successes before Y failures, amount per success towards a goal, and margins of success all modify this. Last step is to check the variance in results between 'competent' and 'incompetent' characters, because that's part of our perception of competence. We expect a competent lion tamer to produce a rather different set of results than an incompetent lion tamer, prehaps most notable there should probably be rather less blood loss by the competent lion tamer.

Prehaps the last thing I would consider is the difference between community discovered 'best practices' and what people who aren't aware of those best practices are getting from the system/books. This, unfortunately, probably can't be quantified in any meaningful way. It would require a method of surveying the people who are not plugged into the communities of people (like GitP, Enworld, RPG.net, etc.) who search for and talk about better ways of doing things or dissect in detail what the intent behind the published system. The best we can get is anecdotes about what we've seen happen with other non-networked GMs and players. We here know that three 30% chances are not equal to a 90% chance of success, that having a thief roll stealth 5 times is pretty much assuring failure, that rolling for everyone driving to work every day at 99% success gives you insane accident rates. We understand that most game systems don't really call for swimming checks in calm lakes by people who know how to swim and that hobbit vs. ogre arm wrestling shouldn't be a single 1d10 vs. 1d10+2 "who rolled higher?" check. But not all the game books come out and say those things, or they don't always say it explicitly, or it's stuck in a single sentence half way through the book in a section of vague general advice. So while the people here understand to avoid the "roll for pants" syndrome that's not everyone playing for running the games. Again, unfortunately, there's no quantifying that.

So, let me try to apply this to the system I'm currently working with. It's a roll & keep of exploding d10s with the characters being slightly super-powered, starting a good notch above the common folk in their areas of expertise, and ending up in full on super hero type territory. Both skills and attributes go from 1 to 5, roll both and keep the attribute. They get a few built in rerolls each session and the ability regularly to boost their rolls, but those are more the slightly super-powered starting point. Mostly they just have overall higher starting stats and skill options compared to regular people & basic soldier types. In theory the "average" target number for anything is 15 which roll 3 and keep 2 hits about 47%, but with work a character can start at the near max of 10 keep 5 which doesn't drop to a 50% fail rate until somewhere between target numbers 40 & 45. Since 3 keep 2 is where even the lowest end non-combatants have their 'day-to-day living' type checks and the basic regular soldiers have 4k2 fighting skills that's probably baseline "competent" for normal, daily, un-heroic activity that you shouldn't be rolling for anyways. Let's see, straight ability checks contain not being afraid of hordes of zombies at 15, nerving yourself up to charging into machine gun fire at 20, putting yourself out when you're on fire at 15. Not bad chances for getting multiple tries on say, roll 3 and keep 3, which gives you about 65% chance of making the DC 15 each time. Check with skills include stopping someone from bleeding out with 6 seconds of work at 20, 30 if you're both trying to run away while doing so, unjamming a gun while being stabbed is 15, giving an effective motivational speech to the crew of a starship in combat is a 20, various space navigation stuff is 20, casting a basic spell is 15 but you want some wiggle room because of how the magic system works, pulling an airplane out of a stall is a 15, driving about 40 mph on a sheet of ice is about 20 after modifiers. 6 keep 3 puts out at about 82% for the 20s and 96% for the 15s, since an assistant gives a +5 on the skill checks (not all checks are appropriate for it though) and you get multiple checks one several of those that's probably pretty good for "competent" at almost all of of the system. However, it's a hideous mutant system thrown together by a random guy on the internet. There is literally no explanation of what DC 15 arcana, lore, tech-use, pilot, medic, or perception check should look like, let along what a DC 30 means beyond a descriptor of "heroic". It is missing anything on when to roll stuff and different types of checks (things like 5 success before 3 failure, or 100 points of rolls over DC 20). So, absolutely and completely not new-GM friendly since it totally lacks useful advice and sets of examples (I had to go through four sections of two book to get that spread of DCs to look at). You definitely run the risk of getting "roll for pants" syndrome from DMs who don't understand probability. On the plus side since the dice generate a bell curve probability set your reasonably confident of beating someone with both fewer rolled and kept dice that yourself, even if you only get one roll and almost certainly if you get 2 or more opposed rolls.

[Quertus]

Prehaps the last thing I would consider is the difference between community discovered 'best practices' and what people who aren't aware of those best practices are getting from the system/books.

We here know… that having a thief roll stealth 5 times is pretty much assuring failure,

Although this is a much-maligned example, I, personally, always loved this mechanic.

It wasn't a question of *whether* the thief would be found, it was a question of *when*.

This had major impact on what "competence" looked like for a Thief. They had to have good risk management skills, and good ability to adapt to unforseen circumstances to create an exit strategy. And, of course, on those rare occasions where the Thief in question actually *did* make all their rolls, they looked substantially more competent in comparison.

I believe you cannot Take 10 in situations where there is stress... aka any actual conflict with opposition.

Hmmm… we may be interpreting that differently. I could be wrong.

The outer loop (level treadmill) is much less important than the inner loop (combat and encounters).

I feel I'm missing context to understand this sentence.

Are you suggesting that a noob experiencing an x% success rate against non challenges, leveling through a master experiencing x% success rate against top-end challenges will result in feelings of competence? That, no matter my current climb skill, any narratively-appropriate climb checks always having an x% success rate somehow invokes a feeling of competence?

If so, I'm not buying it.

Also, since the game is aimed at "heroic adventurers", hiring a horde of hobos is kind of not the issue, since the game is player focused.

And smart players hire a horde of Hobos, get a better result than the party experts could have, pocket the extra cash, and feel really competent in their decision-making skills.

That's Determinator play right there.

It doesn't feel like "heroic adventurers" to me. It doesn't feel like the system encourages "heroic adventurers" to me.

IOW, you're still applying simulationist logic even when attempting to talk about the game from a more gamist POV.

That sounds like me

And they are, essentially. Now, it's pretty true in 5e that an army of X size composed of Y ability individuals will be equivalent to a party of 4-6 adventurers of N level and a particular class makeup. That's been true all along - the only thing that has changed is the values of X, Y, and N.

"Immune to nonmagical weapons" gives voice to the untruth of that statement.

"DR 50/+3" comes pretty close, too.

I'm less certain for systems that aren't D&D though.

But even if said army exists, then 4-6 adventurers are a better choice - they're more mobile, less obvious, and require less concern about logistics.

The army costs less. The adventurers hire the army, and pocket the leftover funds.

And here we get to the real issue, since "anticipating lightning and doging raindrops" is not something that is in any way "simulationist".

Strongly disagree. I never said that i was stimulating *this* world!

Still, I'm not actually seeing the relevance for whether or not our rules *successfully* model "dodging raindrops" with acceptable versimilitude… to that piece of the conversation.

You have a preference for, essentially, superheroic level play, and the power level that is associated with it

Guilty as charged. Sort of. I'm intentionally exaggerating for effect, to try to ensure that people know which *direction* I'm headed.

(which is, realistically, almost orthogonal to being "competent").

I mean… I've got a manager who has shown approximately zero growth since he started with the department. I tend to consider that "inability to grow" to be a sign of a certain lack of competence.

So, yes, I associate "growth" and "competence" - perhaps more than it is reasonable to, but I refuse to believe that they are unrelated, that "I *cannot* grow significantly better than the clueless noob" could possibly produce something worthy to be called "competence" - at least, for anything even remotely relatable given my experiences in this world.

Bounded accuracy ruins that for you since it means that rather than the pure power fantasy of 3.x, you end up with a more grounded, but still frankly kinda extreme, power fantasy.

Eh, pardon this genius for saying, but the 3.x skill system felt kinda a "weakness fantasy" in terms of how much growth one could expect from level 1-20 (let alone on this supposed "e7" world) in terms of representing "tiers of skill" for "odds of winning against someone of lower tier" (for single-roll contests).

I await anyone posting any real-world numbers on any actual studies that have been done in any fields, but… my instincts say 3.x didn't go far enough in the *math* (and, worse, simultaneously went, at times, "too far" in the effects. So that dissonance was jarring.).

The issue isn't really that you want an infinite number of hordes to die against a storm giant. You want a PC to be worth an infinite number of "normal" people.

Close. I want Shakespeare to be better than a thousand monkeys - whether the PCs are Shakespeare or the monkeys in this example.

If the PCs are all dance noobs, they cannot expect to fumble their way to dancing better than the dance master just by virtue of getting more attempts.

If the PCs are all programming noobs, they cannot expect to out program the programming expert just by virtue of getting more attempts.

If the PCs are all writing noobs, they cannot expect to write a better novel than the top novelist just by virtue of getting more attempts.

If one of the PCs is an expert on thermonuclear astrophysics, the rest of the party cannot expect to out theorize their one expert, just by virtue of getting more attempts.

I don't even want to see the world-building that would result from such a baseline, where numbers are King/Trump.

I then tried to make the example more obvious and over the top, about dodging rain drops and moving faster than thunder. Although "can only be harmed by magical weapons" might have served me better as an example.

And that's okay. But it's not universal. To some people, that's a feature, not a bug.

Eh, we're probably too far afield for me to pick apart which pieces of that are true, and how.

I do disagree with your assertion - or, more accurately, your presumtpion that it is the ideal mode of play.

You seem to be falling into the trap of trying to prove your preferences. You don't need to, and you can't (since they're just preferences).

There certainly are issues where I could be more clear which pieces are "my preference" and which are "I consider this universally best" (and which are "I'll call this 'universally best' even though I should know better, just because… I haven't *seen* anything better / I cannot easily describe a more accurate depiction of the actual state".

But I think that there are (or should be) "you must be this tall to ride" barriers to entry on various challenges. Someone who had neither grown teeth not developed language skills is ill equipped to be a professional taste tester judging fine cousine on the Food Network, for example. If the Death Star was a functional "scry and die" stealth platform, a rag-tag fleet of fighters should not be expected to be an adequate defense against it. If a Storm Giant could sense anything the storm touches and hurl lightning unerringly at any foe who so much as touched a rain drop within miles of its position, how many random peasants would you consider to be a match for it?

I know that I'm still struggling to make my point, but I'm talking about the notion that higher levels of skill involve playing a completely different game, that the toddler who struggles to stand or put on shoes and Hamilton choreography are worlds apart. That the crayon that they struggle to hold and press to paper and the works of Van Gogh are worlds apart. Even if the child is prolific, and makes 100x more scribbles than Van Gogh made paintings.

[Florian]

I'm not really sure how that applies. Competence is mostly around what you're doing at the time, not necessarily growth. The point with him gaining power is less about "learning = competence" and more about the fact that he always faces "appropriate" opposition, and that many of the challenges he faces in later books would not have him be "competent" if he faced them earlier in his career.

Ok, let´s differentiale between "learning" and "experience".

To make this easy, let's use coding as an example.
You might know a system and programming language inside and out (learning), but unless you have worked on multiple projects (experience), you don't actually know what is expected, how the actual procedure is and so on.

Someone with with low learning but high experience will be highly functional within his field of knowledge, but mostly fail when outside of it. Someone with high learning and low experience will be hit and miss.


Harry is a case of low/high. As a character, he is highly experienced, but mostly acts outside of his weight class. Each "loss" increases his learning, bringing him up to the point that he can bring his experience to bear. He's an "old dog learning new tricks", so to speak.

In D&D terms, high/low is rolling on a skill and trusting on luck, while low/high is take10/take20.

Hence, that is why I mentioned D100. You can try you luck and either succeed or fail, but your character has the chance to learn from failing, by doing so, overall increase "power".

That also brings me back to the original topic. A competent character has a certain routine covered. The example low/high coder will reliably do his work, unless something exceptional is asked for.

[Democratus]

I consider a character competent if they successfully fill their story role.

A bumbling, but funny, character is competent if their role is to bring comic relief.

[kyoryu]

Although this is a much-maligned example, I, personally, always loved this mechanic.

It wasn't a question of *whether* the thief would be found, it was a question of *when*.

It's fine, if intended. The problem is that most of the time it's not intended.

I also don't buy that the checks are actually completely independent... checks cover a lot of things, and a number of those things are going to be the same across those checks. I think handling it as a bonus is realistic, but either way can work. Just understand the math.

Hmmm… we may be interpreting that differently. I could be wrong.

https://www.d20srd.org/srd/skills/usingSkills.htm

When your character is not being threatened or distracted, you may choose to take 10. Instead of rolling 1d20 for the skill check, calculate your result as if you had rolled a 10. For many routine tasks, taking 10 makes them automatically successful. Distractions or threats (such as combat) make it impossible for a character to take 10. In most cases, taking 10 is purely a safety measure —you know (or expect) that an average roll will succeed but fear that a poor roll might fail, so you elect to settle for the average roll (a 10). Taking 10 is especially useful in situations where a particularly high roll wouldn’t help.

Emphasis mine.

I feel I'm missing context to understand this sentence.

Are you suggesting that a noob experiencing an x% success rate against non challenges, leveling through a master experiencing x% success rate against top-end challenges will result in feelings of competence? That, no matter my current climb skill, any narratively-appropriate climb checks always having an x% success rate somehow invokes a feeling of competence?

If so, I'm not buying it.

I'm suggesting that being able to effectively tackle the things that you're dealing with will result in "competence", and that the in-the-moment feel is more important to "competence" than progression.

And smart players hire a horde of Hobos, get a better result than the party experts could have, pocket the extra cash, and feel really competent in their decision-making skills.

That's Determinator play right there.

It doesn't feel like "heroic adventurers" to me. It doesn't feel like the system encourages "heroic adventurers" to me.

I see that as a failure to engage in the basic idea of the game. IOW, yes the game about heroic adventurers breaks down when you don't use it to play, you know, heroic adventurers. Go figure.

"Immune to nonmagical weapons" gives voice to the untruth of that statement.

"DR 50/+3" comes pretty close, too.

I'm less certain for systems that aren't D&D though.

D&D is a bit of an outlier, yes.

But even so, armies don't attack giants with swords. They attack them with trebuchets and ballistae.

The army costs less. The adventurers hire the army, and pocket the leftover funds.

"Game designed to be fun when played as a small group of heroic adventurers doesn't work well when used as an economic army management simulator. News at eleven."

Strongly disagree. I never said that i was stimulating *this* world!

Still, I'm not actually seeing the relevance for whether or not our rules *successfully* model "dodging raindrops" with acceptable versimilitude… to that piece of the conversation.

You're assuming it's a goal.

I mean… I've got a manager who has shown approximately zero growth since he started with the department. I tend to consider that "inability to grow" to be a sign of a certain lack of competence.

So, yes, I associate "growth" and "competence" - perhaps more than it is reasonable to, but I refuse to believe that they are unrelated, that "I *cannot* grow significantly better than the clueless noob" could possibly produce something worthy to be called "competence" - at least, for anything even remotely relatable given my experiences in this world.

Meh. They're likely related (at least in the real world), but there's a difference between "competent at doing the thing I need done right now" and "is an employee that will grow". They're not totally orthogonal, but they're not the same.

Eh, pardon this genius for saying, but the 3.x skill system felt kinda a "weakness fantasy" in terms of how much growth one could expect from level 1-20 (let alone on this supposed "e7" world) in terms of representing "tiers of skill" for "odds of winning against someone of lower tier" (for single-roll contests).

Maybe? Though I don't think that, realistically, "grandmaster chess player beating noobs" is really within the range of things they're trying to model.

Any model simplistic enough to run at a table is going to fail in places. The best we can realistically hope for is a model that behaves reasonably well in the defined range of "interesting" things.

And, usually, that means that people do get better over time, and ideally that characters don't have to be specialists to have any chance of doing things.

If you really wanted to do "Chess Wars: The RPG" you wouldn't resolve a chess game in one match. Like combat, there'd be enough rolls and enough stats behind it that you'd get the results that combat gives you (where, barring TO, a level 20 will smushify a level 1).

I await anyone posting any real-world numbers on any actual studies that have been done in any fields, but… my instincts say 3.x didn't go far enough in the *math* (and, worse, simultaneously went, at times, "too far" in the effects. So that dissonance was jarring.).

I think you're giving it an impossible task. I think the question is "under what scenarios does the skill system need to give interesting results, and what do we want those properties to be."

You're saying everything and I firmly believe that it's unachievable to actually model that, especially in a situation where you want non-specialists to be able to do anything at all.

And I don't think that's actually relevant for actual, at-the-table play, to feel competent. Apparently, you do. You seem to want to know that, mathematically, these things will be handled by the system, even if they'd never come up at the table.

Close. I want Shakespeare to be better than a thousand monkeys - whether the PCs are Shakespeare or the monkeys in this example.

If the PCs are all dance noobs, they cannot expect to fumble their way to dancing better than the dance master just by virtue of getting more attempts.

If the PCs are all programming noobs, they cannot expect to out program the programming expert just by virtue of getting more attempts.

If the PCs are all writing noobs, they cannot expect to write a better novel than the top novelist just by virtue of getting more attempts.

If one of the PCs is an expert on thermonuclear astrophysics, the rest of the party cannot expect to out theorize their one expert, just by virtue of getting more attempts.

I don't even want to see the world-building that would result from such a baseline, where numbers are King/Trump.

I then tried to make the example more obvious and over the top, about dodging rain drops and moving faster than thunder. Although "can only be harmed by magical weapons" might have served me better as an example.

Though the "magical weapons" thing isn't really about the characters.... an army outfitted with magical weapons would be effective...

Again, these are white room examples. Your issue seems to be that in white room cases, you might be outperformed. Does this actually come up?

In systems I tend to run, this is mostly handled by a few things (and I'll agree, the D&D skill system in most cases is not my favorite).

1. Narrative permission - "what you can do" and "how likely you are to accomplish it" are handled by separate widgets. So, no, a newb cannot fix a nuclear reactor.
2. Focusing on "do you get what you want". No, a newb dancer will not "out-dance" an expert in any meaningful way. However, they could impress the judges more with their "raw, fiery display", even as the judges. If the result is framed as "who does better", then the experts should win all the time. "Do you impress the judges" is a different situation.
3. Not rolling at all when the result is obvious - yes, the grandmaster will beat the newbie. Why even roll?
4. Not handling actually dramatic things with a single roll, but instead expanding them to a larger sequence of rolls.
5. Bell curve results.

Eh, we're probably too far afield for me to pick apart which pieces of that are true, and how.

There's no need to pick anything apart. It's simple - you want a game experience where a character can defeat literal armies, and that's important to you. To me, not only is it not important, but the idea of that system actually kind of turns me off of it a bit (presuming that we're talking roughly the same type of things, which is the case in D&D. In an RPG about playing gods, obviously it would be different).

There certainly are issues where I could be more clear which pieces are "my preference" and which are "I consider this universally best" (and which are "I'll call this 'universally best' even though I should know better, just because… I haven't *seen* anything better / I cannot easily describe a more accurate depiction of the actual state".

Subjective: The skill system should give these results under these systems
Objective: The skill system is effective at creating the results given by the subjective criteria

But I think that there are (or should be) "you must be this tall to ride" barriers to entry on various challenges. Someone who had neither grown teeth not developed language skills is ill equipped to be a professional taste tester judging fine cousine on the Food Network, for example. If the Death Star was a functional "scry and die" stealth platform, a rag-tag fleet of fighters should not be expected to be an adequate defense against it. If a Storm Giant could sense anything the storm touches and hurl lightning unerringly at any foe who so much as touched a rain drop within miles of its position, how many random peasants would you consider to be a match for it?

The critic thing? I mean, like, they do that in a lot of cases, right? Bring on somebody that's not an expert to give color or whatever, though not to the extreme in your result.

And to get to the point, you again seem to be conflating "power" and "competence". For you to feel competent, you have to be tackling things that nobody else can even touch. And that's fine, but it's clearly not universal, given the sheer number of settings and games that don't do that.

I know that I'm still struggling to make my point, but I'm talking about the notion that higher levels of skill involve playing a completely different game, that the toddler who struggles to stand or put on shoes and Hamilton choreography are worlds apart. That the crayon that they struggle to hold and press to paper and the works of Van Gogh are worlds apart. Even if the child is prolific, and makes 100x more scribbles than Van Gogh made paintings.

I play a lot of Fate, and that's mostly handled through scale and sheer narrative denial (which is something that doesn't exist in D&D). I think the range of scales in the game world is also fairly orthogonal to the idea of competence, unless your literal ideal of competence is superheroism (which would preclude things like the Kids on Bikes genre from ever having competence, which I think is an error).

That also brings me back to the original topic. A competent character has a certain routine covered. The example low/high coder will reliably do his work, unless something exceptional is asked for.

Which indicates actual areas of competence - if doing the areas they have things covered, they're competent. If they're pushed outside those areas, they're not.

Which is why I'm describing competent in a relative way - "can the character realistically impact the obstacles in their way, and the challenges they face?"

[Florian]

@kyoryu:

You avoid how that should be handled in a TTRPG. Dark Heresy has a simple rule: If the combination of skill and factors is at or above 50%, don't roll, the character is competent enough to succeed.

[kyoryu]

@kyoryu:

You avoid how that should be handled in a TTRPG. Dark Heresy has a simple rule: If the combination of skill and factors is at or above 50%, don't roll, the character is competent enough to succeed.

I don't think "auto success" is what defines competent.

And I think that defining that is up to each game, even for that scenario. D&D effectively does it with Take 10. Fate does it with narrative permissions and only calling for rolls in situations where there's "interesting" consequences for failure.

I think a good way of getting the feeling of competence is to avoid incompetence comedies - if there is a failure of a character in something that they "should" be able to do, blame it on environmental situations rather than abject failure. The enemy ducked at the last minute, the lock is jammed, etc.

[Florian]

I don't think "auto success" is what defines competent.

And I think that defining that is up to each game, even for that scenario. D&D effectively does it with Take 10. Fate does it with narrative permissions and only calling for rolls in situations where there's "interesting" consequences for failure.

I think a good way of getting the feeling of competence is to avoid incompetence comedies - if there is a failure of a character in something that they "should" be able to do, blame it on environmental situations rather than abject failure. The enemy ducked at the last minute, the lock is jammed, etc.

You managed to entirely miss the point I was trying to make.

D100 is more or less the only system I know that actually manages to model "learning". You don't have the abstraction of EXP, but rather using skills will lead to the eventual advancement of said skills, with a heavy focus on failure over success.

For example, you have the task difficulty of +/- 80 between hellish and routine, time modifier of +/- 80 for immediately and all the time in the world and so on, you can model competency just fine by allowing Auto Access.

Dresden is just an interesting by-note, because "losing" is not normaly part go gaming, at least not in the sense of D&D.