The RNG And What It Means

[Hiram McDaniels]

I've been lurking on this forum for a while, and I keep seeing the term "falling off the RNG" as being damning evidence against the quality of a games design.

What I'm wondering is what this actually means. Does falling off the RNG refer to the sum modifiers outpacing the randomizer range, like adding +38 to a d20 roll? Or does it refer to skill bonuses making it impossible for a novice character to succeed in climbing a waterfall, and impossible for an epic character to fail in kicking down a wooden door?

Either way, what is the actual practical effect this has on table play?

[erik]

Falling off the RNG is when the RNG is not a factor in success/failure. It is not always a bad thing. In d&d it is desired that at a certain point a high level character not care about the combat threat of lessers and having those lessers unable to generate a big enough number to matter is one way to accomplish that.

It is a failing of design when that is not the desired outcome but it happens anyway.

[Hiram McDaniels]

Ahh...I see.

What about capping target numbers, so that the probability of hitting them increases as the characters get bigger numbers, but is theoretically possible for a novice character to achieve on a lucky roll? Like in a D&D like game where the highest DC possible is like 25, so that mud farmer tier characters hit it 5% of the time and Cthulhu punching tier players hit it 80% of the time? Is this the sweet spot, or just a bad mushroom trip?

[Ice9]

One thing that people sometimes confuse it with is "having bonuses bigger than the size of the RNG", which actually is completely different. For example:

Attack Bonus: 1d20 + 1000. AC Range: 1002-1020. On the RNG.
Attack Bonus: 1d20 + 14. AC Range: 5-15. Off the RNG.

That aside, being off the RNG is relative - you are off the RNG relative to tasks you can't fail at or can't succeed at. Or relative to another character if you auto-succeed at tasks they auto-fail (or vice versa).


It's not always a bad thing:
* Off the RNG relative to lower-level characters. IMO fine, as long as long as foes don't go obsolete so fast your options get sparse. Opinions vary on this - 5E is (theoretically) based on the idea that baseline Orcs should still be a potential threat at 20th level.
* Off the RNG relative to same level characters. IMO, a bad thing if it's for core adventuring competency stuff. For things that are a side niche not everyone's expected to have (and don't take up lots of table time), it's fine - Bob the Ranger can easily track shit you have no chance of following, for example. Some people take the stance that this should be handled as an explicit switch (Legendary Tracking, Y/N) instead of de-facto via divergent bonuses.

[Hiram McDaniels]

It's not always a bad thing:
* Off the RNG relative to lower-level characters. IMO fine, as long as long as foes don't go obsolete so fast your options get sparse. Opinions vary on this - 5E is (theoretically) based on the idea that baseline Orcs should still be a potential threat at 20th level.

I'm of two minds on this.

One the one hand, it makes sense for Kings to shit their pants at the thought of a marauding goblin horde and explains why dragons don't just rule everything.

On the other hand, adventurers are supposed to be special. That's why they get to go kick devils in the dick while everyone else stays home at the mud farm and watches their children die of cholera. But if a company of archers can just shoot the lich king to death and be home before dinner, what the fuck do they need adventurers for?

[OgreBattle]

On the other hand, adventurers are supposed to be special. That's why they get to go kick devils in the dick while everyone else stays home at the mud farm and watches their children die of cholera. But if a company of archers can just shoot the lich king to death and be home before dinner, what the fuck do they need adventurers for?

In Warhammer Fantasy battle that works because the Lich has skeleton knights that generate a fear aura terrifying your archers away... and it's also a tabletop game where your knife wielding goblin still contributes to fighting greater daemons of chaos. I don't like it for D&D though.

In a 20 level game of D&D something that's 6-8 levels above you should be highly resilient or outright immune to your attacks. Then you have roughly 4 tiers in the game for "realistic"/heroic/super heroic/Batshit gameplay.

[Username17]

War hammer is also well known for being absolutely terrible at depicting the difference between weak and powerful things. In the fluff, it takes a hundred or a thousand guardsmen to take down a space marine and even more skaven clan rats to take down a lich king. But of course the table top game gives the rate of space marines murdering guardsmen to guardsmen murdering space marines in close combat at only 6:1. The Lich king vs. clan rats situation is harder to math hammer because it depends on the magic rules (which are highly edition specific and often rather vague), but the numbers involved are obviously orders of magnitude lower than what is described in the fluff. Of course, the numbers of physical models required to match Warhammer's fluff and art would be impossible to collect and paint, so we can see a solid practical reason for the game to reduce scale.

One thing an RPG can do is to treat squads of soldiers differently from individuals. A squad or swarm doesn't have to be made up of it's parts, it can be its own thing. So a single ant can be off the RNG with respect to a starting character while a giant swarm of ants is not. And you can do the same thing with Orcs and high level characters: individual Orcs can be off the RNG while when you put them into a squad they are in aggregate still on the RNG. That way you can essentially use logarithms to extend the RNG - the Orcs can trade damage potential for accuracy and can thus achieve hit rates per Orc that are very much lower than 1 in 20 or whatever the limits of your random number generator are.

-Username17

[Maxus]

It's damning when it happens in unexpected ways to people (read: PCs and PC-like characters) who are meant to be roughly as powerful.

To go to the obligatory D&D example, I'd expect a level 15 Fighter to beat up a level 15 Wizard in a boxing match.

A level 15 wizard out-grappling the Fighter is sort of weird.

A level 7 wizard being able to break most of the enemies in the game over his knee is so weird that it messes with the game and causes table arguments about What D&D Is. The same applies to other old builds like the Cleric Archer. Some classes can incidentally pick up stuff for one level and turn other party members and all their gear into...blockers. Something an enemy has to go around.

Unrelated: Digging the name/pic/sig. Night Vale ftw.

[spongeknight]

Falling off the RNG is generally worst when you're talking about inter-party discrepancies. If you have a buffed cleric swinging at enemies for a +30 to hit and a grandmaster monk swinging at +10 to hit, there is no range of AC that the d20 can account for that lets those two people fight against the same enemies. Either the cleric hits with every attack or the monk misses with every attack. For characters of the same level who should be contributing against the same enemies, that sucks.

That's the real problem with falling off the RNG, and many game systems have that problem. It's pretty hard to fix unless you're willing to put every player on rails and dictate their numbers so that everyone is equal.

[OgreBattle]

One thing an RPG can do is to treat squads of soldiers differently from individuals. A squad or swarm doesn't have to be made up of it's parts, it can be its own thing. So a single ant can be off the RNG with respect to a starting character while a giant swarm of ants is not. And you can do the same thing with Orcs and high level characters: individual Orcs can be off the RNG while when you put them into a squad they are in aggregate still on the RNG. That way you can essentially use logarithms to extend the RNG - the Orcs can trade damage potential for accuracy and can thus achieve hit rates per Orc that are very much lower than 1 in 20 or whatever the limits of your random number generator are.

Are there already rules in D&D for doing that? I could see Aid Another being used though that involves a lot of die rolling.

[spongeknight]

Are there already rules in D&D for doing that? I could see Aid Another being used though that involves a lot of die rolling.

There's a Mob template, yes, though it's not very good.

[Hiram McDaniels]

It's damning when it happens in unexpected ways to people (read: PCs and PC-like characters) who are meant to be roughly as powerful.

To go to the obligatory D&D example, I'd expect a level 15 Fighter to beat up a level 15 Wizard in a boxing match.

A level 15 wizard out-grappling the Fighter is sort of weird.

A level 7 wizard being able to break most of the enemies in the game over his knee is so weird that it messes with the game and causes table arguments about What D&D Is. The same applies to other old builds like the Cleric Archer. Some classes can incidentally pick up stuff for one level and turn other party members and all their gear into...blockers. Something an enemy has to go around.

Unrelated: Digging the name/pic/sig. Night Vale ftw.

I suspect that in WotC's zeal for making orcs a credible threat for epic level characters, they forgot that epic level monsters are supposed to be a credible threat for commoners.

I do believe that some amount of bounded accuracy is good, but not to he point where characters only see a 20% gain in efficacy over their entire careers.

It is problematic when there is that much disparity between specialized characters and non-specialized characters in terms of skill rolls. That's why I advocate capping DC's and essentially counting down from the top: so that average adventurers have a shot at passing the adventure, while specialists merely do so with greater frequency.

(Love Nightvale. I'm looking forward to the recording of their crossover with the Thrilling Adventure Hour).

[Lago PARANOIA]

It is problematic when there is that much disparity between specialized characters and non-specialized characters in terms of skill rolls. That's why I advocate capping DC's and essentially counting down from the top: so that average adventurers have a shot at passing the adventure, while specialists merely do so with greater frequency.

Please elaborate. Right now, it sounds like you're describing 4E D&D's DC-by-level nonsense. Which is bad.

[DSMatticus]

Actually, it sounds more like that describes nothing at all. Or everything at all. Or maybe it describes 5e, in which commoners can sometimes succeed at stealthy tasks grandmaster ninjas have failed to.

Saying that you advocate capping DC's is exactly like saying that you advocate for a finite set of tasks containing an element which is most difficult. Or a finite set of integers containing an element which is largest. It's... not exactly a bold assertion, and says nothing about the distribution of the contents of that set. It certainly doesn't suggest anything in relation to the RNG.

Saying that average adventurers should "have a shot" at making a skill check while specialists should be able to do so with greater frequency means... well, if you're talking about characters of the same level, that's true for some skills. Of course specialists at level 20 should not be on the same RNG as the average adventurer at level 1. But whether or not a specialist at level 20 should be on the same RNG as the average adventurer at level 20 depends on what the skill in question is. Something like stealth? Probably. Everyone needs to be making stealth checks, so a specialist shouldn't be so far above and beyond his peers that they aren't playing the same stealth game. But it would be weird if a specialist and Joe Average were rolling the same, say, forgery checks.

Similarly, different skills should improve with level at different rates. Levelling up should improve speech skillers slower than athletic skills. A low-level party should be able to talk to the King of Giants and have that be a meaningful encounter, even if arm-wrestling him is right out.

[OgreBattle]

Are there already rules in D&D for doing that? I could see Aid Another being used though that involves a lot of die rolling.

There's a Mob template, yes, though it's not very good.

Any threads on tgd where folks try to make mob templates or equivalents not suck?

[ishy]

Ahh...I see.

What about capping target numbers, so that the probability of hitting them increases as the characters get bigger numbers, but is theoretically possible for a novice character to achieve on a lucky roll? Like in a D&D like game where the highest DC possible is like 25, so that mud farmer tier characters hit it 5% of the time and Cthulhu punching tier players hit it 80% of the time? Is this the sweet spot, or just a bad mushroom trip?

Lets talk about a Pit Fiend instead of Cthulhu.
This one, from 3.5:

I believe the Pit Fiend should not consider Steve, the crap covered farmer, a threat at all. Nor should it care about 500 Steves.

There are a couple of ways to do that, for example:
a) you can make sure Steve is off the RNG
b) the Pit Fiend has abilities which allow it to shrug off Steve attacks

If you go for solution b), then it is okay for the Pit Fiend to be on the RNG.
You should not decide what your RNG looks like, before you know what it is supposed to do.

[RobbyPants]

Any threads on tgd where folks try to make mob templates or equivalents not suck?

I made one a couple years back, but I'm not going to say it doesn't suck.

It was a quick attempt to throw something together for 3E for a mid-level adventure. It did what I needed it to, but it's going to suffer the usual problems any 3E template suffers, in addition to trying to get it to scale well with level.


Granted, I still like this overall approach, but I think it's something that needs to be written into the system earlier.

[Stubbazubba]

I do believe that some amount of bounded accuracy is good, but not to he point where characters only see a 20% gain in efficacy over their entire careers.

It is problematic when there is that much disparity between specialized characters and non-specialized characters in terms of skill rolls. That's why I advocate capping DC's and essentially counting down from the top: so that average adventurers have a shot at passing the adventure, while specialists merely do so with greater frequency.

So...you want a skill system? They all do that: literally every game has a highest DC. What does counting down from the top mean?

It sounds like you want to have a set range of DCs and then have set bonus maximums that put specialists close to, but still below, that DC, and set bonus minimums that put "average adventurers" below that DC, but not by the entire length of the RNG. Or in other words, you advocate not going off the RNG. Great.

This hasn't really answered anything: is this only true for characters of the same level, or are level 1-20 characters all facing the same DCs? Those are two very different games. If it's the former, you have to decide how quickly average/specialist characters outpace themselves between levels; how many levels does it take before you're "off the RNG" from where you started? If it's the latter, you have to figure out how to make a meaningful level progression that doesn't rely on much vertical advancement. Good luck with that.

Once you've got that figured out, you just pick a % chance of success for the low end and the high end, figure out where that is on your chosen RNG, and then write your bonuses such that they stay within those ranges. For example, for a d20 game, let's say I want the high chance to be about 85%, and the low to be about 35%: that's a difference of 50%, or 10 points on the d20. So I cap bonuses at +10, which should keep characters on the RNG.

Is that what you mean by count backwards from the top?

[Hiram McDaniels]

I do believe that some amount of bounded accuracy is good, but not to he point where characters only see a 20% gain in efficacy over their entire careers.

It is problematic when there is that much disparity between specialized characters and non-specialized characters in terms of skill rolls. That's why I advocate capping DC's and essentially counting down from the top: so that average adventurers have a shot at passing the adventure, while specialists merely do so with greater frequency.

So...you want a skill system? They all do that: literally every game has a highest DC. What does counting down from the top mean?

It sounds like you want to have a set range of DCs and then have set bonus maximums that put specialists close to, but still below, that DC, and set bonus minimums that put "average adventurers" below that DC, but not by the entire length of the RNG. Or in other words, you advocate not going off the RNG. Great.

This hasn't really answered anything: is this only true for characters of the same level, or are level 1-20 characters all facing the same DCs? Those are two very different games. If it's the former, you have to decide how quickly average/specialist characters outpace themselves between levels; how many levels does it take before you're "off the RNG" from where you started? If it's the latter, you have to figure out how to make a meaningful level progression that doesn't rely on much vertical advancement. Good luck with that.

Once you've got that figured out, you just pick a % chance of success for the low end and the high end, figure out where that is on your chosen RNG, and then write your bonuses such that they stay within those ranges. For example, for a d20 game, let's say I want the high chance to be about 85%, and the low to be about 35%: that's a difference of 50%, or 10 points on the d20. So I cap bonuses at +10, which should keep characters on the RNG.

Is that what you mean by count backwards from the top?

Yeah...that's exactly what I was talking about.

I'd rather see less vertical advancement and more horizontal advancement, personally.