What if rolls were only used adversarially

[Skrum]

I don't think I need to spend a ton of time setting this up because we're all familiar with RNG and how it interacts with ability checsk, but in brief, a perennial problem with skills is they're inconsistent. Even a character with a +12 with their athletics check might roll a 2 and fail to climb a medium-difficulty rock wall. Some players will fully embrace the chaos of the dice, but I personally (and I get the sense that a lot of people agree with me) find this to be jarring and immersion-breaking - especially if at the same time I was rolling a 2, the party weakling with no prof and 10 str rolled a 19 and easily climbs up.

It's kind of infuriating, and interacts particularly poorly with spells that are often auto-succeed: yes the spell takes resources, but if there's an important roll on the line, it makes a lot more sense to let the spell be a sure thing than running the 15-20% chance the rogue or fighter or ranger fails their check.

So, what if skills were more like abilities. I'm imaging like several tiers of skill mastery, something like

Untrained
Novice
Adept
Expert
Master

DMs would then assign difficulty levels in terms of these tiers; an adept-level lock to pick, or expert-level wall to climb. If you've got that level of ability, you do the task. No roll required. If you one tier below, you can roll and you might succeed. If you have 2 or more tiers below, you can't attempt; it's too hard of a task for your level of mastery.

Rolling would mostly be used for adversarial rolls, like stealth vs perception or athletics for a grapple. Two (or more) creatures facing each other, often in combat.

I'm just spitballing and would want to really think over how this would shake out and set the tiers correctly, but the idea of skills being more ability-like - that is, reliable - appeals very highly. It would feel far more appropriate to have a skills-based class if skills were something to count on.

[LudicSavant]

I don't think I need to spend a ton of time setting this up because we're all familiar with RNC and how it interacts with ability checsk, but in brief, a perennial problem with skills is they're inconsistent. Even a character with a +12 with their athletics check might roll a 2 and fail to climb a medium-difficulty rock wall. Some players will fully embrace the chaos of the dice, but I personally (and I get the sense that a lot of people agree with me) find this to be jarring and immersion-breaking - especially if at the same time I was rolling a 2, the party weakling with no prof and 10 str rolled a 19 and easily climbs up.

It's kind of infuriating, and interacts particularly poorly with spells that are often auto-succeed: yes the spell takes resources, but if there's an important roll on the line, it makes a lot more sense to let the spell be a sure thing than running the 15-20% chance the rogue or fighter or ranger fails their check.

So, what if skills were more like abilities. I'm imaging like several tiers of skill mastery, something like

Untrained
Novice
Adept
Expert
Master

DMs would then assign difficulty levels in terms of these tiers; an adept-level lock to pick, or expert-level wall to climb. If you've got that level of ability, you do the task. No roll required. If you one tier below, you can roll and you might succeed. If you have 2 or more tiers below, you can't attempt; it's too hard of a task for your level of mastery.

Rolling would mostly be used for adversarial rolls, like stealth vs perception or athletics for a grapple. Two (or more) creatures facing each other, often in combat.

I'm just spitballing and would want to really think over how this would shake out and set the tiers correctly, but the idea of skills being more ability-like - that is, reliable - appeals very highly. It would feel far more appropriate to have a skills-based class if skills were something to count on.

This sort of reminds me of a concept in Kevin Crawford's OSR RPG about playing as up and coming divinities, Godbound (which draws heavily on the old school D&D ruleset).

It advises that if players are doing some skill check related to their domains, they just straight up win, no roll, unless they are being opposed by a divine force or the like.

[MoiMagnus]

D&D's d20 checks are high variance (outside of specific builds that eliminate it).
This is not much of an issue with attack rolls as you will do multiple ones, but this is indeed an issue for "one and done" rolls.
Without going homebrew, tools at your disposal are:

• Automatic success/failures or giving advantages/disadvantages. It is part of the power of the GM to say "you were specifically trained at this task so you get an advantage" or "you have a lot of time to do the task so you get an advantage". For the most part, an unmodified D&D skill check is a coin flip, a GM shouldn't ask for such a skill check if they're not fine with that level of randomness. In a lot of situations, the result shouldn't be random.
• If you're not satisfied with giving arbitrary bonuses to some specific characters for RP reasons, the other tool at your disposal is multiple checks. Like "for climbing this wall, make three athletic checks, DC 10, but you need at least 2 success out of 3" or "make athletics checks DC 15 until you succeed, and tell me how many tries it took you so that I know how much time you took and how many times you might have hurt yourself". Or even combining the two "make athletics checks DC 10 until you get two success in a row, but note that having a failure after a success means that you fall from almost the top so you will take some small damages" (with maybe some additional rules on how you can help another player, etc).

If you're going for simple homebrew, I've seen quite frequently peoples claiming they use 2d10 or 3d6 instead of a d20. Either for every check, or only for the checks that are deemed "low randomness" by the GM.

[Theodoxus]

There's an OSR (I'm away from home, and I don't recall the name off the top of my head) that uses tiered skills that I'm taking and modifying for my own homebrew.

The premise is, you start as a Novice in a skill, and each success (or natural 20, depending on how quickly you want to scale up), increases that skill proficiency by one step in the word 'NOVICE'. So, by the time you have six successes, you've spelled NOVICE and now become an EXPERT. Same deal, until you're a MASTER. That particular game stops there. I added ADEPT and GRANDmaster, to mirror the 5 tiers of Proficiency Bonus in 5E. So, naturally, if copying the concept to standard 5E, a Novice is +2 PB, a Master +4 PB, etc. This allows for both tiered skills as a concept, and different rates of proficiency for an individual. Just because you're trained in Animal Handling and Stealth shouldn't mean you're equally good if your Wis and Dex are the same.

Now, to marry it to your proposal, if you're willing - is I would allow the tiered auto success as you state, but still let the player roll to see if they 'learn' something new. On a success, they move up a pip. On a fail, they don't - but it has no consequence on the actual action taken. This way, not everyone is advancing their skills at the same instant at level up; skills that are used more often will scale up faster than those that don't (allowing players to either emphasize specific skills or be more 'renaissance men' and spread them out.

The other advantage, is you can have 'untrained' skills (everything else) and allows characters to organically grow as they attempt things (and hopefully succeed so they learn) that they otherwise never get proficiency in.

For completeness, I'm going in a slightly different direction; I'm using the Ability Check alternate skill system from the DMG, and then allowing PCs to specialize in specific skills. If you're trying to jump a chasm and don't have Jump, you're rolling a Strength check. However, if you have proficiency in Jump, you auto succeed up to 1.5x your strength score. If the distance is further than that, you'd need to roll (total = distance jumped - another ruling, yes).

Just different options to consider.

[stoutstien]

Isn't that exactly what the rules already says but just worded differently?

[Skrum]

Isn't that exactly what the rules already says but just worded differently?

*Shrug*

Sorta? There's a sentence at the beginning of the chapter that says if there's "chance of failure," than you roll. That's it. While that can certainly imply what I've said, it's hardly rules, is it?

[clash]

Ah I had a similar concept in my one rules light system. Basically you automatically pass any check lower than your score. If you score isn't high enough you roll but the roll goes up or down and has consequences for failure. Basically it's always better to play it safe if you can.

[stoutstien]

*Shrug*

Sorta? There's a sentence at the beginning of the chapter that says if there's "chance of failure," than you roll. That's it. While that can certainly imply what I've said, it's hardly rules, is it?

I was thinking about the multiple ability check section of the DMG. If the task is possible and given enough time they're going to succeed. the only time you would roll in mostly situations is if they are choosing to rush it.

[greenstone]

If I remember right, the CORPS game system had skills and difficulty ranging from 0 to 5, with the rule that you only rolled dice if the difficulty was higher than your skill. So, if a character was skill 2 with opening locks and they came across a level 1 or 2 lock, they opened it without rolling dice. Level 3 and higher locks required a roll.

The idea is that a skilled woodworker shouldn't be roling for success in making ordinary furniture. There's no chance of failure, they have made plenty of chairs before. A skilled alchemist shouldn't be rolling for success for making ordinary medicines. Save the rolls for the consequential things - a masterwork piece of furniture, an ancient elixir to cure the king's curse, the lock to the palace treasury.

Maybe for 5E it could be something like "if your bonus is more than the difficulty - 10 then you succeed". So if a character with DEX+8 attempts to open a difficlty 10 or 15 lock, they succeed without any rolling.

I could see this reducing the number of rolls, and the rolls it removes aren't the ones that add tension and excitement at the table.

It would make combat more deadly.

If you are worried about the swinginess of dice, though, consider using 2d10 or 3d6 in place of 1d20.

[Skrum]

If I remember right, the CORPS game system had skills and difficulty ranging from 0 to 5, with the rule that you only rolled dice if the difficulty was higher than your skill. So, if a character was skill 2 with opening locks and they came across a level 1 or 2 lock, they opened it without rolling dice. Level 3 and higher locks required a roll.

The idea is that a skilled woodworker shouldn't be roling for success in making ordinary furniture. There's no chance of failure, they have made plenty of chairs before. A skilled alchemist shouldn't be rolling for success for making ordinary medicines. Save the rolls for the consequential things - a masterwork piece of furniture, an ancient elixir to cure the king's curse, the lock to the palace treasury.

Maybe for 5E it could be something like "if your bonus is more than the difficulty - 10 then you succeed". So if a character with DEX+8 attempts to open a difficlty 10 or 15 lock, they succeed without any rolling.

I could see this reducing the number of rolls, and the rolls it removes aren't the ones that add tension and excitement at the table.

It would make combat more deadly.

If you are worried about the swinginess of dice, though, consider using 2d10 or 3d6 in place of 1d20.

I definitely would not apply this to attack rolls or the like; this would be only for non-adversarial skill checks.

And yes, 3d6 would normalize rolls and reduce the chance of feel-bad failures. That might be the simpler solution.

I like the idea too of rolling multiple dice - like, roll 3 and if you're untrained, all must succeed. If proficient, 2 must succeed. If you have expertise, only 1 needs to succeed. That has a lot of appeal as well.

[JellyPooga]

Part of (IMO the Primary) the issue(s) with D&D's system in general is that it functions on a high value, linear RNG and has binary results; pass or fail, degree of success is irrelevant. This is fine for quick, fire-and-forget style play, particularly when combined with the simplified Dis/Advantage system that means you don't have to track the minutiae of every dice roll. When it comes to representing character proficiency, on the other hand, it leaves something to be desired; it's too swingy to accurately portray competence in a task without GM fiat betraying player agency and there's little reward in stacking the odds in your favour because there's no incentive to go any further that hitting the DC on the nose. So you end up with characters that are "unreliably ok" or often "laughably bad" at things they should be "heroically good" at.

The first fix is to ditch the single d20+mod basic roll. It really is a horrible die type with few benefits;
- d10 or d12 gives much tighter bounds and thus more predictable results
- 2d10, 3d6 or (my personal favourite) 5d4 gives you a bell curve (and some far more interesting possibilities on critical hit mechanics than "I rolled one specific number that has just as much chance as any other! Yay!"...how about crits for 2-, 3-, 4- and 5-of-a-kind offering increasing grades of critical success? Or perhaps offering "poker rules" for pair, two-pair, full house or straight?)
- d% gives finer granularity and more intuitive eyeballing

Implementing any of these will give much more reliable results and therefore better verisimilitude with regard to character proficiency, without having to adjust the rules too much. Introducing Dice Pools in place of d20+mod is another option, but is much more drastic and would require significant change.

Second fix is to implement degrees of success and failure. Rolling significantly higher or lower than your target should have a tangible benefit. Plenty of other systems do it and it's not hard. Heck, D&D does it (badly) with critical hits. Just expand those rules to be actually interesting and based on character skill (i.e. actual result) rather than just random chance (i.e. just the die roll).

Third fix is to take the die roll out entirely. Give characters aspects, traits or background features that offer automatic success beyond their stat line. That Half-Orc has the "Brawny" trait? Yeah, he doesn't have to roll to kick down doors and lift portcullises; that's part of who his character is. In opposed contests, he rolls as normal, but gets an extra degree of success for free. The puny Wizard arm-wrestling our Brawny Orc will have to roll two successes to the Orcs none in order to win that contest; possible, but much harder.

[Luccan]

How will you set DCs under this system? Like, if something is an expert check, will it be so high a novice actually couldn't roll that high (say, 25+?) or is the cut off more arbitrary, so you could assign any number after deciding who is allowed to roll (so even a Master check could end up with only a DC 15, something you could roll even without bonuses)? Because I don't think either of those scenarios are actually desirable

[Skrum]

How will you set DCs under this system? Like, if something is an expert check, will it be so high a novice actually couldn't roll that high (say, 25+?) or is the cut off more arbitrary, so you could assign any number after deciding who is allowed to roll (so even a Master check could end up with only a DC 15, something you could roll even without bonuses)? Because I don't think either of those scenarios are actually desirable

Probably closer to the latter. I've often thought that DC 15 or even 10 should the limit for untrained checks of certain types - like the knowledge skills, picking a lock, medicine checks; these skills represent special knowledge and training and good luck can't propel a novice to success at even moderately difficult tasks.

The other skills, I think adding automatic disadvantage on checks above that DC is defensible. It would be for similar reasons to using 2d10 or 3d6 - preserve the power of having proficiency and don't let the die roll be the most important factor.

[rel]

My approach when I run 5e is to call for rolls and assign DC's based on who's doing the activity as much as the activity itself.

The party come up to a wall that needs climbing; The acrobat climbs without needing to make a check, the portly wizard makes a DC 15 strength ability check with the athletics proficiency, the druid who saw fit to become a wolf fails to climb automatically.

Each character is trying to climb the same wall, but they don't all find it equally easy.

[KyleG]

Im in the same boat. Im not a fan of how skills-attributes and even tools interact in 5e, nor as indicated above how the rogue can out religion the cleric potentially or the wizard out climb the barbarian, at least as far as the dice roll. If im building a religious rogue sure but by pure dice rolls it sucks.
Im also not a fan of how skill dcs work either. I shouldnt be able to fluke a hard lock at L1 i might have to revist ot at level 9. Not to mention that even at level 20 with max modifier i still might get a 12 just by fate or the rogue might get 37. That's quite the discrepancy.

[Rynjin]

I am once again asking 5e players with this complaint to re-implement the Take 10/20 rules from 3.5

[Justin Sane]

My "do I ask for a roll?" checklist:

1st, is it downright impossible? If so, auto-fail, no roll.
2nd, can it fail at all? If not, auto-success, no roll.
3rd, are there consequences for failure? If not, auto-success, no roll.

Three points to check, usable with nearly every system.

For DnD specifically, point 2 removes the need for the Take 20 rule, and point 3 removes the need for the Take 10 rule.

[stoutstien]

My "do I ask for a roll?" checklist:

1st, is it downright impossible? If so, auto-fail, no roll.
2nd, can it fail at all? If not, auto-success, no roll.
3rd, are there consequences for failure? If not, auto-success, no roll.

Three points to check, usable with nearly every system.

For DnD specifically, point 2 removes the need for the Take 20 rule, and point 3 removes the need for the Take 10 rule.

The 10 and 20 in 5e isn't the die result. it's the amount of time that might be needed to complete the task in this manner.

[Justin Sane]

The 10 and 20 in 5e isn't the die result. it's the amount of time that might be needed to complete the task in this manner.

I know.
For Take 20, if the PCs are searching the ancient library top to bottom, for as long as it takes, until they find the information they need? I'm not even going to bother adding up their modifiers.
For Take 10, if the Rogue is trying to open a particularly stubborn lock, but we've established there's no guards patrolling nearby, so he can take his time? I'll just narrate it takes him a couple of tries, but he gets it open eventually.

[Blatant Beast]

*Shrug*

Sorta? There's a sentence at the beginning of the chapter that says if there's "chance of failure," than you roll. That's it. While that can certainly imply what I've said, it's hardly rules, is it?

The DMG does give guidance that the game can be handled with only rolling rarely, or rolling for everything, with most games falling in the middle between those two poles.

My personal belief is the "roll dice for everything" model is not a good fit for 5e, in much the same way the "only roll dice when it matters" model was not a great fit for 3e.

[Yakk]

The non-combat attribute checks of 5e lack something combat has. Magnitude.

What if we added it?

What if a success generated some kind of success points equal to your proficiency bonus? It could be (prof+attribute, min 1) or just (prof, min 1).

So someone trained in a skill produces 2 points of success instead of 1 on the same check at level 1. A level 20 character with expertise generates 12 points of success.

If we had a kind of skill challenge system this could be turned into "HP of damage on the Challenge" type thing.

...

Challenges

Challenge are complex problems. A Challenge has a HP track and Complications.

When you engage a Challenge, it starts fighting back. This can be because things happen over time, or because as you go deeper into the castle the patrols get thicker and more dangerous.

The players describe how they are handling the challenge. You can ask for more steps if you like. Once the stakes are large enough, you go an ask for checks. PCs who are passive and don't help can act as dead weight, but generally ask for a check from each PC.

For each PC that succeeds, grant success points equal to the proficiency bonus of any skill or tool used (min 1). Failures can sometimes just be a lack of progress, and in other cases make Complications worse.

Then issue a Complication. Complications can sometimes be reacted to, or dealt with later by spending resources or actions.

The more rounds the players take to overcome the Challenge, the more Complications hit them. So 'dead weight' players will want to actively help.

The use of non-skill checks is allowed. If you use an ability, feel free to grant (proficiency bonus) success points or (spell slot level).

Challenges are measured in Tier (1-4 is tier 1, 5-10, 11-16, 17-20, 21+ is tier 5) and Complexity.

Tier	DCs	Success Points
T1 (1-4)	5-15	8*Complexity
T2 (5-10)	10-20	12*Complexity
T3 (11-16)	15-25	16*Complexity
T4 (17-20)	20-30	20*Complexity
T5 (21+)	25-35	24*Complexity

A T4 Complexity 4 problem has 80 Success points required.

PCs have 6 proficiency bonus; that is 13 successes. With 4 PCs, that is just over 3 successes each.

Expertise doubles your success rate. DCs are in the 20-30 range, so failure happens a lot if you aren't optimized for it.

Now, if we used check modifier instead of proficiency, things flatten out a bit. The expert still hits +17ish, but a talented untrained character can pull off +3.

The value for a typical skilled PC is ~4 at level 1 up to ~10 at level 20. At ~2-3 successes per PC, 8 successes per complexity in T1 up to 30 successes per complexity at T4. The 22 attribute expert has a modifier of +18, significantly better than the typical skilled character.

You'll note I'm trying to avoid being overly negative with failures. The point is that Complications happen, the goal is to defeat the problem before they happen. This encourages lower skilled PCs to try to contribute instead of sitting it out. That also makes me want to use total skill modifier, and not proficiency bonus.

[Theodoxus]

Funnily enough, at one point, I was creating a casting system that worked very similarly. Basically each spell had a Target Number (TN) that mages would need to exceed in order to cast. Most TNs were 10x spell level, though some were higher, and you could ritually cast any spell, but the TN was 10x the normal one (so typically 100x spell level.

Your rolled Arcane, Religion, or Nature checks, with each point adding to the total. Rituals could have multiple casters, allowing them to be cast much faster than a typical 5E ritual, though at the opportunity cost of multiple people doing the same thing.

I never got around the polishing it, but I felt like it would at least brought casters down a step. Sure, you could finish off a combat with a Fireball... but it might take 3 rounds to ultimately cast it.

[Sorinth]

If you just map the skill modifier to your Untrained, Novice, Adept, Expert, Master tiers then I think it should work fine. So for example Untrained is 0 or less, Novice is +1 to +3, Adept is +4 to +6, Expert is +7 to +9, Master is 10+. Though I would maybe consider allowing a roll when 2 tiers below but have that roll be at disadvantage.

[Abracadangit]

Pull the trigger on this, in games you run. Seriously.

I ran some version of this a while back, operating under similar logic. I think my version was, if the character's bonus (ability + prof if applicable) + 10 was equal to or greater than the DC, and they were proficient in the skill, they auto-succeeded. People loved it; monks could parkour over railings without any problem, barbs got to feel mighty even while not raging, etc, etc.

It's a better way to live.

[Skrum]

Pull the trigger on this, in games you run. Seriously.

I ran some version of this a while back, operating under similar logic. I think my version was, if the character's bonus (ability + prof if applicable) + 10 was equal to or greater than the DC, and they were proficient in the skill, they auto-succeeded. People loved it; monks could parkour over railings without any problem, barbs got to feel mighty even while not raging, etc, etc.

It's a better way to live.

Heh, mechanically that's giving everyone Reliable Talent. I totally see the logic of it, especially if everyone really wants to lean in to the "hero" part of heroic fantasy.

I'm definitely going to do something like this; I'm just so continually frustrated and disappointed by the existing skill system.

[Abracadangit]

Heh, mechanically that's giving everyone Reliable Talent. I totally see the logic of it, especially if everyone really wants to lean in to the "hero" part of heroic fantasy.

I'm definitely going to do something like this; I'm just so continually frustrated and disappointed by the existing skill system.

I should have specified; much like your suggestion, it was only if there wasn't an intelligent force/NPC opposing the check. Plus social checks required a little more selling on the player's part. But yeah, for the other skills, it was rad. I also ran knowledge skills as totally passive (unless you were trying to ID a monster), because they're ALREADY situational by definition. Failing a knowledge check and forever being sealed off from information your character is supposed to know is obnoxious.

At some point, the Mad Libs-ness of the rogue failing their Expertise-d Acrobatics, with a banana peel sound effect, or the wizard failing Arcana just becomes stupid. I understand that it's "not the character failing, it's circumstances conspiring against them" or whatever, but it's still stupid.