What is the best system for skills in a D&Desque game

[OgreBattle]

It's a common criticism of many games here, we know what the flaw are ("MTP! Off the Rng!")

So, what is a flawless (or as close to) system for skills?

You know, stuff like searching stuff, talking to folks to influence opinions, climbing stuff, researching stuff, sensing stuff.
And stealth, how's that shit suppose to work.

[echoVanguard]

There really isn't one. Any given system will have pros and cons with regard to how it handles at least one of those actions, and all of them will have at least some amount of MTP, however minimal, because they are not strictly-defined actions. Stealth in particular is an ugly target - making a stealth-based character almost guarantees extended arguments at the gaming table the second one person's idea of "what's realistic" clashes with another's. You will also have vociferous disagreements the very instant another character uses a social skill on a player's character.

For example, Shadowrun 4 seems at first glance like it handles these things well, but it really doesn't. All social skills operate at a penalty in virtually every situation, and you can pretty much forget about researching anything ever since the list of Knowledge skills is literally composed entirely of write-ins. No fooling, I have had an MC tell me I failed a knowledge roll because I didn't have "Snohomish Ganger Slang" on my character sheet. Stealth as usual is a disaster, but less so than in some other games because you have access to spells or cyberware that literally make you invisible - assuming you have access to those resources and your opponent's didn't conveniently have access to that technique's hard-counter equivalent.

In my experience, Rules-Light systems have fewer problems with these sorts of actions than crunchier systems, because people have fewer places to anchor their objections. If you're trying to find an entire ruleset to use, you have a lot of options - GURPS Lite handles most skills pretty well, especially interpersonal skills since they affect only reaction rather than force particular courses of action (although its combat is a complete disaster past a certain point), but it's hardly the only game in town. Nemesis Age's playtest document should be available Soon™ for an additional option.

echo

[hogarth]

3E D&D conflated three different types of things under the heading "Skills":

[*]Specialist stuff, e.g. Knowledge (Arcana). This is probably okay as a class feature that works along the lines of bardic knowledge. E.g. clerics know about religion, wizards know about magic, etc.

[*]Minor nonspecialist stuff, e.g. Jump or Climb. The 3E system is probably fine for these skills, since nobody really cares if you have broken the RNG for jumping or climbing. As long as you have some rules that work reasonably well at low levels (e.g. a level 1 average Joe NPC shouldn't need to worry about jumping over a 5' pit or climbing a ladder), you're good.

[*]Major nonspecialist stuff, e.g. Diplomacy, Stealth, Perception. This is a tough one. I still haven't seen a game that has done a good job at stuff like this, although 1E AD&D's "surprise 4 times in 6" and "surprised 2 times in 10" is starting to grow on me, for simplicity's sake.

[TOZ]

Depends on how much you use skills.

[Blicero]

For a D&D-style game, I think the SR4E/After Sundown method works fairly well.

You have skills like Jumping and Swimming and Lying and Stealth and shit, and then you have skills like Snohomish Ganger Slang or Famous Swedish Chefs, or whatever. You probably want the first category to scale with level, but the other category doesn't have to. And you can't trade proficiency in one category for proficiency in the other.

That way, every character will be mildly adept at things that are important for everybody to be good at for certain adventures to be viable, but every character can have a weird background that gives them esoteric knowledge.

The only problem is, as echo sez, that it is very important for all of your Passive Skills to be of roughly the same relevance. When one character walks in knowing about "Obscure Dishes from the Rendish Swamps," and another walks in with "Knowledge: Arcana," you have a problem.

But Skill systems in general are often kind of shite, and not just because no one has ever come up with a satisfactory Diplomacy or Stealth subsystem.

[tussock]

Searching stuff: I like MTP. I tell you about the searchable things, you tell me how you search them. Mostly like that line that turns up around clickable things on modern computer games.

Influence opinions: tricky. Mostly you just can, but there's some failure chance attached to whatever problems I can think up, which is more of a try harder signal. Actual diplomacy is just quest giving in disguise, yes you can, go fetch some foo.

Climbing stuff: yes you can. Unless you can't, but then the Thief can anyway. Climbing really big stuff might be exhausting, find a ledge and rest up.

Researching stuff: random time allocation.

Sensing stuff. If I want players to maybe miss something like a secret door, that'll just be 2-in-6 or whatever, give or take how quick they're moving past.

Stealth: you can't see a hidden person, because they already made that roll. Well, unless you've got some funky power that gives you an X-in-6 of "feeling" something is wrong, like Paladins and Elves and dogs do.

[OgreBattle]

So what works for stealth then

or rather, why does stealth fail in existing systems?

[ishy]

Stealth often fails, because it is 1 person doing the stealthing. That means if the stealth girl wants to do something, everyone else needs to take a coffee break.

[Username17]

Stealth has a number of failure points:

• "Reasonableness". If you can't stealth down an empty corridor with a guard in it, you basically can't stealth. Many games simply hand out automagic failure for some incredibly bullshit scenarios like "sneaking across open areas". That kind of crap makes it so that you can't sneakily approach a fortress at all, which makes you wonder how people did that in stories and human history all the fucking time.
• "Law of Large Numbers". If you roll dice a lot of times, you'll get some black swan events. Since there are normally several player characters and several NPC guards, that is a lot of die rolls. So it's a lot of opportunities for PCs to roll 1s and NPCs to roll 20s.
• "An army moves at the rate of its slowest soldier"
  -Napoleon
  When Team PC sneaks around, they are spotted if the guard with the best perception test spots the PC with the worst stealth check. So if you have even one clumsy dwarf in full plate, your sneaking opportunities are few and far between.

You could make a stealth system that didn't succumb to those problems, but I don't think I've ever seen one.

-Username17

[Lago PARANOIA]

Re: Reasonableness
If any portion of the game needs to be overexplained by the rules, it's this. The game should flat-out state that the DM should err on the side of player empowerment and that DMs who do things like not let people sneak down an empty corridor with a guard in it are Gygaxian assholes out to screw over players and that's extremely non-standard.

Re: Clumsy Dwarf Problem
Personally, I can't and don't really see a reason for skills and stats for that matter not to be capped from level advancement and/or point-by-point investment (though the latter has some additional problems anyway) after a certain point in the game. If you want to go past that point, you have to invest in certain class abilities or magical items. This prevents the clumsy dwarf from falling off the RNG and makes it feasible to do things like make him leave his armor at home and give him a stone which had Silence enchanted on it.

The thinking is that when something is a solo activity like wailing on a monster, if offends the other players at the table if someone is able to automatically succeed while barely touching the dice. If we're talking about a group activity, though, it's okay for one or two people to blow away the challenge while the success rests on the hands of everyone else. So while stealth and whatever for that matter should be designed around non-stealth specialists, the game should also not go out of its way to make stealth challenging for non-stealth specialists.

[OgreBattle]

Stealth has a number of failure points:

• "Reasonableness". If you can't stealth down an empty corridor with a guard in it, you basically can't stealth. Many games simply hand out automagic failure for some incredibly bullshit scenarios like "sneaking across open areas". That kind of crap makes it so that you can't sneakily approach a fortress at all, which makes you wonder how people did that in stories and human history all the fucking time.
• "Law of Large Numbers". If you roll dice a lot of times, you'll get some black swan events. Since there are normally several player characters and several NPC guards, that is a lot of die rolls. So it's a lot of opportunities for PCs to roll 1s and NPCs to roll 20s.
• "An army moves at the rate of its slowest soldier"
  -Napoleon
  When Team PC sneaks around, they are spotted if the guard with the best perception test spots the PC with the worst stealth check. So if you have even one clumsy dwarf in full plate, your sneaking opportunities are few and far between.

You could make a stealth system that didn't succumb to those problems, but I don't think I've ever seen one.

-Username17

I've played games of warhammer where it seems to work out well enough. 40k's infiltrate (and some special character) pretty much have super stealthiness represented as "place guys down wherever you want to because you weren't detected up to this point that the fight breaks out"

It clicks for me as a tabletop miniatures wargame. I'm thinking hypothetically it could work in tabletop but I haven't actually done so.

The stealth guys in a wargame, their function is to bypass strong enemies to take out vital targets or seize objectives. It's about the same in tabletop story RPGs. In warhammer they work well enough.

So yeah, I generally think of the stealth guy not being in the same group as the heavy dorf when he does his shadow crawling thing.

[Lago PARANOIA]

So yeah, I generally think of the stealth guy not being in the same group as the heavy dorf when he does his shadow crawling thing.

That works on a battlefield. It's considerably lacking when we start asking ourselves about the party using Stealth on Shadow Moses Island or the Mad Doctor's Laboratory.

[hogarth]

Re: Reasonableness
If any portion of the game needs to be overexplained by the rules, it's this. The game should flat-out state that the DM should err on the side of player empowerment and that DMs who do things like not let people sneak down an empty corridor with a guard in it are Gygaxian assholes out to screw over players and that's extremely non-standard.

But if you "err on the side of player empowerment" then you tend to run into asymmetrical laws of physics which depend on whether you have the "PC flag" or not. Some players* hate that.

E.g. Player #1 wants to make it moderately difficult for his PC to sneak up on an alert NPC, but player #2 wants to make it very difficult for NPCs to sneak up on his alert PC.

*I'd say "simulationists", but I'm afraid someone might poop his pants if I use that word.

[echoVanguard]

Stealth has a number of failure points:

• "Reasonableness". If you can't stealth down an empty corridor with a guard in it, you basically can't stealth. Many games simply hand out automagic failure for some incredibly bullshit scenarios like "sneaking across open areas". That kind of crap makes it so that you can't sneakily approach a fortress at all, which makes you wonder how people did that in stories and human history all the fucking time.

In what scenarios would this actually be doable? Typically when people perform this sort of action in the real world, it takes one of three requirements:

1. Practical invisibility. This usually takes the form of camouflage or darkness.

2. Distraction. If all the guards are looking at the building that just caught on fire, they're not looking at you. This also covers things like waiting until a guard turns his back to cross a hallway, but this wouldn't help you pass through an occupied hallway with no doors. This also covers actions like "knock a dude unconscious instantly and quietly" which we tend not to like in TTRPGs because it makes stealth too powerful.

3. Subterfuge. If you look like a peasant wandering around, guards are going to ignore you if the area you're in isn't a no-go area for peasants. Similarly, modern spies focus on looking like innocent bystanders or other "authorized or innocuous" actors when attempting to move through areas of actual attention.

In other words, when in doubt, think back to Thief: The Dark Project.

[*] "Law of Large Numbers". If you roll dice a lot of times, you'll get some black swan events. Since there are normally several player characters and several NPC guards, that is a lot of die rolls. So it's a lot of opportunities for PCs to roll 1s and NPCs to roll 20s.

I'm not sure this is an actual problem. Assuming all participants are on the same RNG, you'll fail about as often as you'll fail any other cooperative task. The fact that the stealthing group uses their lowest member's stealth and the spotting group uses the most perceptive member's spot is not so much a problem as a reflection of the fact that stealth is powerful and thus difficult to accomplish.

[*] "An army moves at the rate of its slowest soldier"
-Napoleon
When Team PC sneaks around, they are spotted if the guard with the best perception test spots the PC with the worst stealth check. So if you have even one clumsy dwarf in full plate, your sneaking opportunities are few and far between.[/list]

"Hide! They only saw me!" -- Jack Burton, Big Trouble In Little China
Assuming that both the stealth expert and the stealth un-expert are both on the RNG, I don't believe this is a real problem. In the event the stealth expert needs to scout ahead, they can do that without Clanky the Dwarf - but they can't be off the RNG in the positive direction, or they just say "wait here a half-hour, guys, I'll go scout out the entire dungeon and come back with a map that has all the loot marked on it".

You could make a stealth system that didn't succumb to those problems, but I don't think I've ever seen one.

So, if I'm understanding you correctly and my arguments are sound, then a stealth system that doesn't succumb to those problems would require only:

1. A way for characters to stealth appropriately and effectively.
2 and 3. The requirement that stealth and perception both be on the RNG for all participants and not require a tremendous amount of rolls.

Correct?

echo

[OgreBattle]

It's considerably lacking when we start asking ourselves about the party using Stealth on Shadow Moses Island or the Mad Doctor's Laboratory.

Well, why is heavy dorf being dragged into a sneaking mission in the be the first place?

I wonder if the problem is in archetypes and strict specialization. In the first Conan movie, Conan is in formal duels, fighting in heavy armor, sneaking shirtless, and building traps.

If the heavy dorf takes off his armor, should he be a competent enough sneak? Subotai the Thief will still be better at it.

[virgil]

I've seen more than one system attempt to fix that by having the 'leader' make a roll for the entire group; with varying limitations. Sometimes it requires all of the followers to spend a FATE point, other times it's the leader making a check modified by how much worse the group is to him.

[tussock]

So what works for stealth then

or rather, why does stealth fail in existing systems?

A high level Thief in AD&D can easily make his Hide in Shadows roll and have an entire army walk by without seeing him standing there.

If anyone comes close, he gets way improved surprise chances (arbitrarily so) when he acts, or can just let them keep walking by. There's no large radius around a torch where you can't hide (well, 5' or something), and in fact you need some light nearby or infravision will see you.

Move Silently is similar, if you make the roll, you're silent, like the spell, and you know it. Moving Silently is usually incompatible with Hiding in Shadows, so you need proper cover to sneak past someone. It also improves surprise chances (less so, but still arbitrarily large). You roll once every 60', which means once.

That shit works. You roll once, success chances approach 99%.


3e just fucked up the math, or maybe they stealth-nerfed stealth. The probability of making one check is about right, but the game forces you to make dozens of them to do anything useful, not to mention that all lights have a huge anti-stealth radius. Cumulative probability causes constant failure and the rapid evisceration of anyone trying it.

The PCs basically cannot earn surprise by the rules proper, let alone scout or cross open ground.

[Stubbazubba]

You could just look at the highest roll and the lowest roll, and look at degrees of success/failure; if the lowest roll fails by a number of degrees at least one less than the number of degrees of success of the highest roll, then the party is successful. For instance; the Thief totals a 25 on a DC 18 Stealth check, a difference of 7, we'll say that's in the third degree of success (1-3, 4-6, 7-10, etc.). So long as the lowest roll only fails by up to 6 (the highest in the 2nd degree range), the party is successful. So, as long as Clunky McDwarf can muster up a 12, then the stealth expertise of TAT, The Awesome Thief, signaling when to move and when to stop, finding the best shadows for the party, etc., etc., actually helps mitigate the losses that Clunky would have had himself.

Alternatively, you could look at the total number of everyone's degrees of success and failure, and find the difference there. This would mean that you need to have most of the party on the 'fairly competent' side of things. Either of these assumes that the party is all on the same RNG.

[hogarth]

A high level Thief in AD&D can easily make his Hide in Shadows roll and have an entire army walk by without seeing him standing there.

How many shadows do you need in order to hide?

Is it impossible to hide if there aren't any shadows?

Who the fuck knows?

[Lago PARANOIA]

One failure I've seen in stealthing rules -- and this is most exemplified by the first 5E D&D playtest but it certainly isn't the only offender -- is that it's too visually oriented. A creature without eyes should not, all other things being equal, have a much harder time spotting critters in a cave of dim light than those that have it, but that's the way things roll.

Personally, I think that Stealth (and perception) should be universal to all senses. Things like Move Silently and Scent should just be specializations within those skills. If you have a Good Eyesight specialization for Perception it very often applies except in condition of total darkness or whatever.

Well, why is heavy dorf being dragged into a sneaking mission in the be the first place?

Because the alternative is not to go on sneaking missions ever again, because the player of the clumsy dwarf can't roll up (or may not even want to) a level-equivalent character in the 15 or so advance minutes he gets for the stealth mission.

I'm not sure this is an actual problem. Assuming all participants are on the same RNG, you'll fail about as often as you'll fail any other cooperative task. The fact that the stealthing group uses their lowest member's stealth and the spotting group uses the most perceptive member's spot is not so much a problem as a reflection of the fact that stealth is powerful and thus difficult to accomplish.

Stealth is a special case because adding people above and beyond the lowest members' competency not only rarely helps but it often hurts. The clumsy dwarf has an easier time sneaking into the castle than the clumsy dwarf and stealthy elf, because the elf also has a non-zero chance of blowing their stealth roll and making things harder for the both of them.

But if you "err on the side of player empowerment" then you tend to run into asymmetrical laws of physics which depend on whether you have the "PC flag" or not.

Obviously, the expanded stealth rules help NPCs as much as PCs and PCs should also be made to accept that a ninja can sneak into their camp (which they set up at daylight in the desert) despite the fact that two elves were on watch.

It's PC-focused, however, because DMs traditionally have access to and have used plot devices of phlebtonium to unilaterally circumvent the 'Reasonableness' clause. For example, these particular ninjas have desert camo and are trained in the art of hiding in mirages and desert hazes. Or that they're using a special oil that makes their footsteps sound like the window. Or that they wait for (or causes) a rare cloud to pass by and sneak into that.

[deaddmwalking]

Personally, I think that Stealth (and perception) should be universal to all senses. Things like Move Silently and Scent should just be specializations within those skills. If you have a Good Eyesight specialization for Perception it very often applies except in condition of total darkness or whatever.

This way lies the Paizo madness, where Gnomes gain a +2 on scent-based Perception checks, and Elves gain a +2 on sound-based Perception checks. Unless they've changed it again...

Then you get to argue whether your specialized bonus applies or does not apply to this particular case.

[hogarth]

It's PC-focused, however, because DMs traditionally have access to and have used plot devices of phlebtonium to unilaterally circumvent the 'Reasonableness' clause.

I've probably seen just as many examples of players saying shit like: "Our PCs are walking back-to-back, shining powerful flashlights everywhere, staying away from anything that looks remotely like cover, etc."

I used to play in one 2E campaign where almost every encounter was an (unavoidable) ambush, so maybe I'm biased towards "GMs love stealth" rather than "players love stealth".

[Lago PARANOIA]

deaddmwalking: Well, if you don't want to do that kind of specialization then don't. It's understandable. However, you will still need to do something about Stealth/Perception's vision-bias; if you're not okay with blindness or darkness or deafness not affecting the success or failure of stealth then you will need to either split the skills or have specializations.

I've probably seen just as many examples of players saying shit like: "Our PCs are walking back-to-back, shining powerful flashlights everywhere, staying away from anything that looks remotely like cover, etc."

That's not unilaterally circumventing the Reasonableness clause. In fact that's exactly the kind of convoluted action declaration people would expect in systems that don't err on the side of the stealth-user. If you don't want people salami-slicing their actions that much then you need to give them the benefit of the doubt on whether their stealth 'auto-fails'.

[TarkisFlux]

Treating stealth as "not being noticed" instead of "not being seen" goes a long way towards fixing the reasonableness concern. People look at shit all the time that they don't register, and there's no reason stealth can't be modeled on that instead. Especially in a game with no facing, where it's not even obvious that the guard would be looking down the hallway while you're trying to sneak across it. Distraction and subterfuge can be used to supplement that, by letting them notice you or tricking them into focusing on something else, but it's not a replacement for letting people get to the walls in the first place without cover.

The law of iterative probability screwing the stealther is trickier to get around. You can start by fucking removing the "make a check whenever you might notice someone doing something" bullshit from the game. If you want people paying attention to mean anything, it needs to have an action cost. So you let people take 0 to notice things for free, and make a check or take 10 as a standard action. People still notice things that are sufficiently obvious, like their hands and other people if you set the DCs properly, but they don't notice stealthy things unless they're keeping watch or really perceptive naturally.

As for the slowest soldier thing, you write an ability that lets a stealther bring their party along for the ride. They make a check or take 10, take a hit to their result, and the party stealths along at that result. In conjunction with not making watchers roll every 6 seconds (or whatever), you get a group that can get farther unnoticed than they could previously, even if they can't actually go as far as the stealther could solo.

Or that's what I did with my skills reworking. If any of those are terribad, it would be nice to know.

[Edit] I should add that there are triggers that would cause someone who had failed to spot something (either with their take 0, 10, or a real check) to want to try a real check. Specifically, failing to notice something by 5 or less. In practice it seems to work out so that people get close and then start changing their behavior (moving more slowly, setting up distractions, looking more closely, etc.) and rolling checks against each other. So stealthing in still fails on occasion, but less often and at a much less obnoxious distance.

[hogarth]

I've probably seen just as many examples of players saying shit like: "Our PCs are walking back-to-back, shining powerful flashlights everywhere, staying away from anything that looks remotely like cover, etc."

That's not unilaterally circumventing the Reasonableness clause.

I'm not sure what you're talking about. What I meant was that I've seen PCs play the "reasonableness" card (e.g. "No one can reasonably sneak up on us, so it automatically fails") as often as the GM, if not more so.