D&D 5E How easy are skill checks?

[the Jester]

How, exactly, do you help someone spot that bugbear hiding in the shadows when none of you know he is there in the first place?

 

[wellerpond]

Perception: do they hear a rustle?

If not, then they think it's a squirrel or something not threatening.

Nature or intelligence: do they know it's a bugbear or just something scary?

 

[GMMichael]

Back at the Bat-Cave (or On Topic Cave):

Let's consider a 16 ability score pretty rare, since if you're rolling 3d6 (as most NPCs do?) you have a 4.5% chance of getting that 16 or better. This goes back to the "specialist" argument.

There is a bit of a scaling problem in the difficulty table. Notice that very easy things are a failure for the average person 20% of the time. Gonna set your wine glass down without breaking it? Well you've already set it down successfully four times...the fifth time might not turn out so well.

WotC can justify this problem with the combat chaos explanation: everything is treacherous in combat, even simple things. The problem comes from the swinginess (new word?) of a d20. You have equal odds of rolling a 1 or a 20. Which means that average results, say 10 or 11, only happen 10% of the time.

The difficulty table seems to start too hard, but then becomes more balanced toward the top. 64% odds seems pretty reasonable to me for the specialist attempting something hard. Now, what must one do to accomplish something "nearly impossible?"

 

[77IM]

Setting down your wine glass is trivial, DC 0. If you have a low Dex you might fail 5-20% of the time. I'd grant advantage to a person with a lifetime of experience drinking wine (remember when you were little how awkward wine glasses were? the clumsy kids probably DID spill 5-20% of the time until the got the hang of it). It's a routine, automatic check, so I'd use passive Dexterity, which turns the advantage into a +5 modifier. So that's why ordinary people don't spill frequently, are able to tie their shoes, drive to work, etc.

The system works if you want it to.

 

[77IM]

The difficulty table seems to start too hard, but then becomes more balanced toward the top. 64% odds seems pretty reasonable to me for the specialist attempting something hard. Now, what must one do to accomplish something "nearly impossible?"

A first-level character with a 16 stat and Expertise and advantage and Guidance has a 7.2% chance of success. This sounds right for heroic fantasy: that would be an epic and memorable roll. The same character at 20th level with a 20 stat succeeds 40% of the time by himself and 77% with advantage and Guidance. Seems reasonable.

 

[Paraxis]

Let's consider a 16 ability score pretty rare, since if you're rolling 3d6 (as most NPCs do?) you have a 4.5% chance of getting that 16 or better. This goes back to the "specialist" argument.

I am pretty sure using 3d6 isn't mentioned anywhere ever in the rules for ability generation, npc or not.
16's are just not that rare, of the npc's listed in the back of the MM only 3 have any ability scores lower than 10 so I doubt they use even the array method.
Also of note, 10 out of 20 of the listed npc's have at least one 16 in an ability score, some have more than one 16.

Specialists in 5e, pretty much are anyone with the expertise feature, or a feat that commonly gives them advantage on the roll. Having at least a 16 and being proficient just makes you competent in that area.

 

[The Crimson Binome]

WotC can justify this problem with the combat chaos explanation: everything is treacherous in combat, even simple things. The problem comes from the swinginess (new word?) of a d20. You have equal odds of rolling a 1 or a 20. Which means that average results, say 10 or 11, only happen 10% of the time.

That "swinginess" (very old word, but never official) is a known feature for pretty much any d20 system. Especially at low levels, the range of the die exceeds the range of possible modifiers, so luck is the overwhelming factor in a lot of things.

But sometimes you want unlikely things to happen, because it keeps things interesting. This system allows an unskilled commoner to rarely (5%) succeed at a difficult task, while generally assuring that even the most talented of experts will occasionally mess up.

That doesn't sit well with everyone, and those people may do better with a system that involves rolling more dice (such as substituting a 3d6 or 2d10 in place of the d20).

 

[The Crimson Binome]

I am pretty sure using 3d6 isn't mentioned anywhere ever in the rules for ability generation, npc or not.

If it shows up anywhere in this edition, it will be in the DMG. Remember also that NPCs have racial modifiers, so that's going to skew their numbers higher.

 

[GMMichael]

I didn't find 3d6 - but I found something more sinister.

The Commoner in the basic rules has only 10s for ability scores. I think that qualifies as "most NPCs," per my earlier post.
Ranking slightly higher than the commoner's 0 Challenge is the Guard at Challenge 1/8. His highest score is a 13.
Slightly higher than that - the Thug, who has two non-average scores: 15 and 14.

So I'm sticking with 16s-are-rare. However, I'd hear an argument that D&D's difficulty class table isn't based on difficulty for the average person - it's based on the difficulty for an adventurer. After all, the basic rules say "player or monster," neither of which is a commoner, right?