[Skills] Solutions to the oblivious rogue problem

[Chris_Nightwing]

We've discussed here before the problem of the low wisdom Rogue. He has expert knowledge in finding traps, but that darned Cleric of Pelor shows up and just knows where they all are, better than him!

In the most recent L&L this is mentioned, along with a solution that the team is testing: A skill is used in place of an ability modifier, instead of in addition to. For example, instead of the trapfinding skill being +3 to find traps (typically a Wisdom check because it involves perception), it would be +5 flat, instead of your usual perception check (number modified so that this system makes sense).

This way, the untrained Cleric has +3/+4 and is still perceptive, but the Rogue has +5 because he knows what he's looking for. This sounds like it might work for this problem, but I forsee some problems:

• Firstly, it moves directly away from the original concept that abilities aren't raw physical quantities, but are also representative of your general skill and ability in certain areas.
• If two characters are trained in tightrope-walking, they are equally good, despite their dexterities potentially being rather different - which would still affect their AC and missile attacks.
• In fact, it separates skills from 'combat skills', so should proficiency with a weapon be a flat bonus and completely ignore your Str/Dex?
• Why would you want a skill which you were already pretty darned good at due to your ability modifier?

I don't think this system will function well in practice. Is there another modification to the skill system that would solve this problem, or is it really just down to changing how we adjudicate perception?

I will reiterate my previous solution - wisdom is for passive perception, intelligence is for active perception. You look for traps (Int), but you walk past secret doors obliviously (Wis).

 

[Hypersmurf]

I think there are a couple of questions that should be answered before a Skills mechanic is considered, and then a Skills mechanic should be judged on whether it gives the correct answers to those questions.

I'm not saying what I think the answers should be; rather, I'm saying that whoever's designing the rules should have the answers in mind before a rule is evaluated for workability. If the rule doesn't give the answers you determined to be your goal, it's not a workable rule.

---

Let’s say we have three PCs, Angus, Bob, and Chris, who are all attempting to weave a basket. The DM calls this a Dexterity check.

Angus has an 18 Dexterity.
Bob has an 18 Dexterity and is trained in basketweaving.
Chris has a 10 Dexterity and is trained in basketweaving.

Q1. Is it okay if Bob’s best possible basket is no better than Angus’s best possible basket?

Q2. Is it okay if Bob’s best possible basket is no better than Chris’s best possible basket?

Q3. Is it okay if Angus’s best possible basket - given that he’s never made a basket before but has plenty of natural ability - is better than Chris’s - who has been making baskets all his life but has only average natural aptitude? Is the reverse okay?

---

Current playtest system:
Untrained: + stat mod
Trained: + stat mod + 3

A: +4 (Best: 24)
B: +7 (Best: 27)
C: +3 (Best: 23)

1. B's best is better than A's best.
2. B's best is better than C's best.
3. A's best is better than C's best.

---

Proposed Alternative:
Untrained: + stat mod
Trained: + 5

A: +4 (Best: 24)
B: +5 (Best: 25)
C: +5 (Best: 25)

1. B's best is better than A's best.
2. B's best is no better than C's best.
3. C's best is better than A's best.

-Hyp.

 

[Chris_Nightwing]

Great example [MENTION=1656]Hypersmurf[/MENTION].

With my simulationist hat on, I would say that the alternative is better than the null system - basketweaving is a complex skill that requires more than just nimble fingers. On the other hand, Bob is a natural, his baskets should be better, and he probably found it easier to learn basketweaving. I can envisage more complex systems that incorporate these features.

With my gamist hat on, dexterity isn't just nimble fingers, it's doing things with them. In which case the null system suits my needs. Poor Chris was never that good at basketweaving, but overcame his disability to be better than most. Still, he envies the all-rounder Angus who can do almost anything with his hands.

 

[ComradeGnull]

Wisdom is, and always has been, a rubbish stat. Your ability to find favor with the divine shouldn't have anything to do with being highly skilled at finding Waldo.

Are you taking into consideration, though, the Skill Mastery feature that the Thief has? The Thief has +3 to Find Traps AND can't get lower than a 10 on the check. That equates to 'the talented but untrained person may be able to do better than the trained and untalented person, but the trained person's worst effort will be better than the untrained person's worst effort.'

The phrasing 'minimum die result' is interesting as well; by one reading, that means the Thief can't botch on a natural one, if you're playing with that option. That could be a nice way to reflect the difference in trained/untrained trap seekers- on a lucky day, the priest might find something the thief overlooks, but on an unlucky day, the priest might set off a trap in the course of looking for it. It makes it much safer to use the Thief first, and then the Priest if you think the Thief might miss something.

 

[Hypersmurf]

An alternative way to read the proposed system:

Skill checks are 1d20 + ability mod. When using a trained skill, you are considered to have an ability score of 20.

-Hyp.

 

[Li Shenron]

Skill checks are 1d20 + ability mod. When using a trained skill, you are considered to have an ability score of 20.

It could work, but I then I wouldn't like the fact that there is no difference between a Dex 1 basketweaver and a Dex 18 one.

I think the simplest solution is to make some skills trained/untrained so that if you just don't know how to make a basket, you just don't make a basket.

That's what would happen if I tried to weave a basket! It would take a month of trying before I finish one that doesn't fall apart straight away, and after that month if I succeed it means I have in fact trained myself into basketweaving.

So whether the skill system is flat-bonus based (4e, 5e) or point-based (3e) just make it so that you need at least +1 competence/training bonus to be considered trained, everyone else cannot use the skill.

Skills like climb, jump, perception etc. can still be so that they don't need training.

 

[Herremann the Wise]

Excellent questions and analysis, I tried to XP you both but only partially succeeded.

Looking at Hypersmurf's questions, I would propose the following to provide the granularity and differentiation between the performers I prefer:

* There are two levels of training: Standard and Expert. Standard gets you that +3 bonus while Advanced gets you a +6. All performers utilize a suitable ability (adding that ability's modifier).
* There are three types of tasks based on DC: Basic Tasks, Standard Tasks and Advanced Tasks.

Basic Tasks
Anyone can attempt basic tasks. Those trained (standard or expert) automatically succeed at basic tasks.

Standard Tasks
Anyone can attempt standard tasks. Those expertly trained automatically succeed at standard tasks.

Advanced Tasks
Only those trained can attempt advanced tasks without automatically failing.

Now lets look at Angus, Bob, Chris, and Don (Dex 14;Advanced Training +6):

* Basic Tasks
A: +4
B: Auto
C: Auto
D: Auto

* Standard Tasks
A: +4
B: +7
C: +3
D: Auto

* Advanced Tasks
A: Fail
B: +7
C: +3
D: +8

By doing this, we re-introduce the earlier auto-success concept but separate it from the actual ability score to DC comparison that previously did not quite work. In this way, we give a good nod of the head to training and experience while still allowing the naturally gifted to easily succeed. Training then simply picks up the slack of the old +10 or +20 rules.

However, Advanced Tasks are too difficult for the high natural ability novice. These tasks require a certain amount of know-how. The highly intelligent fighter has loads of potential, but they are never going to just happen to know and be able to perform Mordenkainen's Seven Arcane Principles and Preparations for the Ordered Mind. This is certainly a task left to those with appropriate training.

I could also see a further training category of Mastery. The Master does not necessarily get more bonuses (you want to keep the overall flatness of the system) but they should get advantage when performing Advanced Tasks.

This pretty much gets the numbers right to my eye while giving adequate weight to training. In doing this it most probably separates the ability scores to a more raw concept but hopefully not too much. A side benefit is that this hopefully reduces the necessity for maxed out ability scores with training now picking up most of the slack (I'm not a fan of the "I must have 18 in my primary stat" view and the effects it has on character expectations).

Thoughts?

Best Regards
Herremann the Wise

 

[Chris_Nightwing]

Thoughts?

Best Regards
Herremann the Wise

I like this direction, but I have two concerns. First, the complexity of the system, but that may be unavoidable if we want to represent skills and training correctly. Second, though I like auto-success as a time-saver, it does require a lot of codification of tasks, or DM fiat as to how hard something is.

As an extreme simplification, I suggest taking your first thought and making a simple rule from it. For any task complex enough (ie: this will not include movement, simple perception or shall we say, things an animal can do), you suffer disadvantage on the skill check if you are not trained with that skill. Training gets rid of this and provides a +3 bonus. You could extend this to providing advantage for mastery, and a further +3 bonus.

Angus: +4, but disadvantage because he's never weaved a basket before.
Bob: +7
Chris: +3
Don: +8, but has advantage because he's a basket legend.

So yes, Angus might make a better basket, but he's really unlikely to. Training in these complex tasks is translated into doing them better and more reliably. It also intersects with the Rogue ability nicely.

 

[Pickles JG]

People have far more attributes than the 6 D&D ones & those ones cover far more aptitudes than plausibly correlate. They are already mega fudgy kludge.

So I am happy breaking away from them & going with the some "features" allow you to use a trained skill with a flat bonus that is better than +3+attribute. I might not make it the default but it sounds good as a rogue feature or a background one that you have a skill at which you are very apt & roll with a +8.

 

[Nifft]

You can't hear anything if you are clanking around in heavy metal armor. Thus, put a ceiling on the Perception result which is equal to Stealth check.

Now the Thief is perceptive because he's sneaky, Rangers have a reason to wear light armor, and the Cleric is very seldom worst (thanks to Wis bonus), but also very seldom best (thanks to Stealth cap truncating his high rolls).

Cheers, -- N