D&D 5E Skills in 5e

[Lanefan]

Chalk one up for the pink flowers on this one, in light of I'd prefer a stat-based very loosely defined roll-under skill system (except for Thieves) rather than anything I've seen since.

Lanefan

 

[AbdulAlhazred]

Maybe untrained skills could have a DC limit on checks that can be even attempted, a limit like e.g. 15 (hence lower than the max roll result)?

This was suggested WAY back in one of the early columns by MM, but the problem is it creates a 'two tier' system where either you're trained or why bother. What happens when nobody happens to be trained in some skill? Even in 4e with its short list of skills chances are one or two of them won't be represented by any PC with training.

4e DOES have 'trained only' for some very specific uses of a few of the skills (Arcana and Thievery mostly), but that's about it. The DM certainly can always invoke that in other situations, but a lot of the skills are a rather hard to nail down as 'training' and more represent knack and talent (IE Perception, its hard to imagine a reason for why my high WIS guy cannot hear something that your trained guy can). I'd rather have some rules to treat things like Perception as group checks. This really shouldn't be too hard, the best guy makes a check and the others effectively Aid Another in 4e parlance, or you need more than N successes (better for stuff like Stealth checks for instance).

 

[Blackwarder]

I also voted pink flowers because I also prefer a roll under mechanic for ability checks with 3d6 for a standard check and adding/subtracting dices for harder or easier tasks respectively plus subtracting dices for having proficiency.

In this system if your dice total is lower than your ability score than its an automatic success so most easy tasks will be an automatic success for ability scores of 12 and up.

I think it's more elegant and straight forward than the DC system and its easier to rule on the fly around the table.

You can also add interesting things for the mechanic, for example for every six that you roll you get a problem/hazard that you need to try and solve, it might be as simple as losing HP and it might be another roll you need to do.

Warder

 

[Li Shenron]

This was suggested WAY back in one of the early columns by MM, but the problem is it creates a 'two tier' system where either you're trained or why bother. What happens when nobody happens to be trained in some skill? Even in 4e with its short list of skills chances are one or two of them won't be represented by any PC with training.

This is something I would totally like.

Those who aren't trained just don't bother, unless the check is basically a save (so they better try even if the chance is low, such as in the case of a Balance check) OR unless the check is easy like low-level challenges.

We have bounded accuracy for combat, why shouldn't we have also bounded accuracy for checks? Why should we always have to assume that low level PCs struggle to jump a 10ft pit, and high level PCs never have an issue with that? Personally I totally hate systems where all PCs automatically get better at doing everything.

So going back to your questions?

Q1: Why bother if you're untrained?
A1: Don't bother. This is exactly what I want: that those who aren't trained don't bother, because if they bother it means soon they'll figure out that it's always worth that all PCs try their luck at every Knowledge check or Profession check, Listen/Spot, Survival, Search, Track etc and every other skill the results of which affect the whole party and failure doesn't carry consequences.

Q2: What happens if nobody has skill X?
A2: A wonderful thing happens: that the party has to find another way to solve the problem! Once again, I hate games where the party covers everything, and I hate adventures where the DM requires one and only one skill or ability in order to continue. That's bad adventure design, thus if nobody has skill X and there is no other way to continue the adventure, this immediately shows the DM is a bad DM.

How do you go if you feel the opposite about these two?
1) You can always house rule that everybody is trained at everything, so everyone can always try the check. This is much easier than having a default where everyone can try everything and then having to figure out how to house rule which skills should allow this and which not, and with what limits.
2) You can always grant everybody bonus skills so that the party covers all the skill list and more. This is always easier than taking skills away.

edit: Also notice that at the moment the 5e system in fact IIRC already has some embedded limit of DC25. That's the highest DC you can beat with a skill check if you're untrained, because (without magical aid) you're bound to max 20 in any ability score thus +5 in bonus, therefore you're never going to roll more than 20+5 unless you're trained in the skill. Perhaps this is going to be good enough for my purposes, but at the moment it still feels quite high to me.

 

[Stormonu]

I voted for stat + skill + roll, but I also strongly like stat + roll, gain advantage if skilled.

Problem is, with the latter you can get some weird results where the talented (high stat) noob outperforms the skilled artisan, and I don't quite like that.

In the end, I think I'd actually prefer stat + skill + roll, gain advantage with the roll if skilled. Max the skill bonus to +5 and have it mean more than be "just a number" - go with the old guild rankings; initiate (+1), novice (+2), journeyman (+3), master (+4), grand master (+5). If you aren't of the right "rank" when making a skill test, you suffer disadvantage - negating the advantage you normally get for being trained in the skill. Adhocing the appropriate rank (i.e., where the DM doesn't deliberately set the level) might involve, say, subtracting 12 from the DC (so a DC 15 skill challenge would be a "Journeyman (+3)" rank test).

Rogues would still get their "skill dice", allowing them to add, say +1d6 to a given skill roll.

 

[Li Shenron]

I voted for stat + skill + roll, but I also strongly like stat + roll, gain advantage if skilled.

Problem is, with the latter you can get some weird results where the talented (high stat) noob outperforms the skilled artisan, and I don't quite like that.

It's not the only problem... there is also the problem that advantage doesn't allow to do anything that you weren't able to do when unskilled (because the max roll result is the same), the problem that advantage doesn't scale (while flat skill bonuses or skill dice do) unless you change the limit of 2 dice rolled, the problem that you can cancel advantage with disadvantage (e.g. in the rain, everybody becomes unskilled at several skills?) and also in general the source of advantage/disadvantage in the game should be kept as spare as possible otherwise there's a high chance of always ending up with both, cancelling each other.

 

[Sadrik]

Two issues that are valid in any skill system.

I have however a long-time issue with characters using untrained skills...

Sometimes I just don't like that ALL PCs get to roll. Just think about it: if you call for a Listen check to notice something and TWO PCs can roll the check, it's a little bit like one PC rolling with advantage, because one successful roll is enough. Now just try to figure out to what kind of advantage is equivalent the case of FOUR or more PCs being allowed to roll Listen for the same instance.

Group rolls and trained only skills.

Group rolls need to be a standard rule and not some sort of poorly thought addition. They need to be right up front and easy to do. I'd like to see something where a group of 8 orcs can make a single roll to spot the sneaky halfling. There are a few issues at work here. As construed currently, there would be a lot of contested rolls, hide/spot and move silently/listen. This is very unwieldy and would be nice for both sides to make one roll and resolve the action. The other and most dramatic problem is the contested roll system. Opposed d20s are very swingy (I wont go into detail here but see my arguments in this thread). Then the heart of it is how do you fairly resolve many rolls all under one roll fairly, taking into account many factors (ability, skill, circumstance etc.). What happens when multiple people gang up on any roll? There should be a bonus. But how much? Group rolls should be ingrained in the system completely. Some possibilities:

• Highest skilled person rolls with advantage
• Lowest skilled person rolls with advantage
• Highest skilled person rolls, each other person rolls vs a DC then if successful adds +2
• Highest skilled person rolls, each other person adds +1, to a maximum number of people based on the task
• Like flanking in some way... which is in essence a group roll...

Trained only skills are tough. Not everyone is a surgeon. Certain skills people should have almost no ability to roll in. The rules can be tricky though because if you were a surgeon, you would know a lot of ancillary things about medicine. So imho, to model a system like this there should usually be a root untrained skill and then add the surgeon on top. So for instance, first aid, might be the root skill and then surgery an adder. The surgeon adder would unlock new uses of the first aid skill. In this way, the specializations to skills could come in from a module for those who wanted to use them.

There is one other type of rolling of checks that should be written right into the core: multiple skill rolls to complete a task. Some of these are obvious. Make multiple checks to climb a large cliff (distance divided by climb move rate). But what about a very complex lock, you might need multiple successes. What about banging a door down you might need 3 successes to knock it down (either from multiple people or stacking them round after round by bashing into it). Lifting a heavy slab might require 2 successes (2 or more people making a STR roll each round).

Multiple successes could be a way to handle group rolls too, in theory. Everybody rolls their check if you get a certain number of successes in the group then you are successful... Hmm have to think about how that would work...

 

[AbdulAlhazred]

This is something I would totally like.

Those who aren't trained just don't bother, unless the check is basically a save (so they better try even if the chance is low, such as in the case of a Balance check) OR unless the check is easy like low-level challenges.

We have bounded accuracy for combat, why shouldn't we have also bounded accuracy for checks? Why should we always have to assume that low level PCs struggle to jump a 10ft pit, and high level PCs never have an issue with that? Personally I totally hate systems where all PCs automatically get better at doing everything.

So going back to your questions?

Q1: Why bother if you're untrained?
A1: Don't bother. This is exactly what I want: that those who aren't trained don't bother, because if they bother it means soon they'll figure out that it's always worth that all PCs try their luck at every Knowledge check or Profession check, Listen/Spot, Survival, Search, Track etc and every other skill the results of which affect the whole party and failure doesn't carry consequences.

Q2: What happens if nobody has skill X?
A2: A wonderful thing happens: that the party has to find another way to solve the problem! Once again, I hate games where the party covers everything, and I hate adventures where the DM requires one and only one skill or ability in order to continue. That's bad adventure design, thus if nobody has skill X and there is no other way to continue the adventure, this immediately shows the DM is a bad DM.

How do you go if you feel the opposite about these two?
1) You can always house rule that everybody is trained at everything, so everyone can always try the check. This is much easier than having a default where everyone can try everything and then having to figure out how to house rule which skills should allow this and which not, and with what limits.
2) You can always grant everybody bonus skills so that the party covers all the skill list and more. This is always easier than taking skills away.

edit: Also notice that at the moment the 5e system in fact IIRC already has some embedded limit of DC25. That's the highest DC you can beat with a skill check if you're untrained, because (without magical aid) you're bound to max 20 in any ability score thus +5 in bonus, therefore you're never going to roll more than 20+5 unless you're trained in the skill. Perhaps this is going to be good enough for my purposes, but at the moment it still feels quite high to me.

IMHO you're using the skill system to accomplish the wrong thing. You want PCs to work in terms of a PC concept. "This guy is strong and athletic, he can leap things, climb, wrestle, etc effectively". You don't need to exclude all the other PCs from a chance to attempt these things.

First of all there are consequences for most failures, so it isn't as if in most situations it is a good idea for a poor athlete to try to swim the raging torrent.

Secondly you should be thinking in terms of 'fail forward' and 'success with complications', so that being able to clean blow through something is a boon, but situations can be overcome without needing to be a great expert. The guy with a poor comprehension of ancient writing takes an educated guess at what the map means. Of course he thinks that funny icon is a slide, pity he can't read the "here be giant snake" runes...

I voted for stat + skill + roll, but I also strongly like stat + roll, gain advantage if skilled.

Problem is, with the latter you can get some weird results where the talented (high stat) noob outperforms the skilled artisan, and I don't quite like that.

In the end, I think I'd actually prefer stat + skill + roll, gain advantage with the roll if skilled. Max the skill bonus to +5 and have it mean more than be "just a number" - go with the old guild rankings; initiate (+1), novice (+2), journeyman (+3), master (+4), grand master (+5). If you aren't of the right "rank" when making a skill test, you suffer disadvantage - negating the advantage you normally get for being trained in the skill. Adhocing the appropriate rank (i.e., where the DM doesn't deliberately set the level) might involve, say, subtracting 12 from the DC (so a DC 15 skill challenge would be a "Journeyman (+3)" rank test).

Rogues would still get their "skill dice", allowing them to add, say +1d6 to a given skill roll.

Skills in the sense used in 4e and 5e really should NOT include things like crafting items which are simply activities you undertake until you succeed. There's no need for a carpenter to make a check to see if he makes a door, clearly carpenters know how to make doors. Likewise 'workman like' tasks such as opening a lock. These things can simply be indicated by possession of some sort of option on your character like a background of "occupation: carpenter".

 

[Li Shenron]

Group rolls and trained only skills.
...

There should be a bonus. But how much? Group rolls should be ingrained in the system completely. Some possibilities:

Those are nice example for the standard game, maybe even a little too complex for that too.

But for the basic game I'm afraid something simpler is needed. Maybe just +1 for every person in the group for Listen and Spot checks.

OTOH, the game could indeed just grant everybody a chance to roll Listen & Spot (although DCs should generally be increased to compensate).

Perhaps Listen & Spot are the only 2 skills which really need to grant a roll to anyone even untrained (essentially, they are wisdom perception checks...). All other "group skills" like survival, knowledge, profession could follow the rule that only trained characters can roll, and the others can't (in Basic, at least).

By "group skills" I mean all those skills that ONE character succeeding equals success for all the party. Only these skills are a problem. Notice that athletic skills and physical skills in general are not "group skills" in the sense that ALL characters generally need to make the check.

This problem also has some kind of subtle correlation with retries.

 

[AbdulAlhazred]

Two issues that are valid in any skill system.



Group rolls and trained only skills.

Group rolls need to be a standard rule and not some sort of poorly thought addition. They need to be right up front and easy to do. I'd like to see something where a group of 8 orcs can make a single roll to spot the sneaky halfling. There are a few issues at work here. As construed currently, there would be a lot of contested rolls, hide/spot and move silently/listen. This is very unwieldy and would be nice for both sides to make one roll and resolve the action. The other and most dramatic problem is the contested roll system. Opposed d20s are very swingy (I wont go into detail here but see my arguments in this thread). Then the heart of it is how do you fairly resolve many rolls all under one roll fairly, taking into account many factors (ability, skill, circumstance etc.). What happens when multiple people gang up on any roll? There should be a bonus. But how much? Group rolls should be ingrained in the system completely. Some possibilities:

• Highest skilled person rolls with advantage
• Lowest skilled person rolls with advantage
• Highest skilled person rolls, each other person rolls vs a DC then if successful adds +2
• Highest skilled person rolls, each other person adds +1, to a maximum number of people based on the task
• Like flanking in some way... which is in essence a group roll...

Trained only skills are tough. Not everyone is a surgeon. Certain skills people should have almost no ability to roll in. The rules can be tricky though because if you were a surgeon, you would know a lot of ancillary things about medicine. So imho, to model a system like this there should usually be a root untrained skill and then add the surgeon on top. So for instance, first aid, might be the root skill and then surgery an adder. The surgeon adder would unlock new uses of the first aid skill. In this way, the specializations to skills could come in from a module for those who wanted to use them.

There is one other type of rolling of checks that should be written right into the core: multiple skill rolls to complete a task. Some of these are obvious. Make multiple checks to climb a large cliff (distance divided by climb move rate). But what about a very complex lock, you might need multiple successes. What about banging a door down you might need 3 successes to knock it down (either from multiple people or stacking them round after round by bashing into it). Lifting a heavy slab might require 2 successes (2 or more people making a STR roll each round).

Multiple successes could be a way to handle group rolls too, in theory. Everybody rolls their check if you get a certain number of successes in the group then you are successful... Hmm have to think about how that would work...

I know! There could be this system where different PCs can make checks with various skills and when they reach a certain total they succeed, or if they fail 3 checks they fail!

More seriously though I think you can't disentangle a 4e-like skill system from SCs. The whole thing is part and parcel of a whole.