D&D 5e ability checks


Registered User
Validated User
I'm not very familiar with the D&D 5e rules, though I am, in the abstract, a big fan of most of the simplifications and streamlining that I've seen talked about in 5e.

That said, I would appreciate a deep dive on how well 5e handles non-combat task resolution. I.e., "skill checks" and the like. I ask because I think I'm going to design a modern/hard sci-fi rule set, and I would like 5e to be my starting point. Mostly because a lot of roleplayers like to see familiar things, and the D&D system is the ultimate in familiar things.

E.g., I think the trad Str Dex Con Int Con Cha base works fine for my purposes, so I figure why reinvent the wheel (though the specifics of bonuses are up for review)? Same goes for using the d20 as the go-to die for ability checks, skill checks, etc.

But looking at web search results for how 5e handles task resolution, I'm wondering how much people like it, and what thoughts people have on how they might improve on it, or re-implement it, for a modern/hard sci-fi system.

For reference, I plan to ditch D&D's levels, and maybe the whole class thing, too. And I don't think hit point bloat belongs in a modern/hard sci-fi system and am set on ditching it.


Registered User
Validated User
I would improve 5e resolution system in two ways.

First, I would broaden the range of skill/ability bonus. With 5e, a skill bonus from a 20th level character will only be 6 points higher than a first level character at best (unless he's a rogue or a bard), which is equivalent to the d20 standard deviation. That makes the game really "swingy", as the random element is more important than the skill. Another way to do it would be to chose dice with a smaller standard deviation (such as 3d6 or 1d10).

I would also put different thresholds for different success quality. For instance, if your roll is between DC and DC+4, it's a poor quality, between DC+5 and DC+9 it's average quality, between DC+10 and DC+14 good quality, and son on...

If you want games that use similar system, you could take a look at Talislanta, Traveller or Fuzion. None of those are level based. :)

John Out West

Registered User
Validated User
I didn't like it enough to change it when I made my own system. The proficiency system is a nice change from skill points. Its still annoying that, generally, there's no reason why a character cant "Try again" on any check (Besides maybe a Knowledge check), especially when there are no stakes. The d20, i think, is too large to use for skills when considering their bonuses. A +5 bonus is really good, but its competing against a random chance of 4x the bonus.

What I did, essentially, was make Unskilled checks a 1d6 roll, and Skilled checks a 2d6 roll. Getting a 6 on a roll was good, for an amateur, but some challenges can only be overcome by getting a 7 or higher. Easy tasks like jumping through a window are a 2. Anything above a 12 was simply impossible for a mortal. There were not modifiers, only unskilled and skilled, with advantage or disadvantage based on the situation. This eliminated the skill-point arms race that would inevitably appear, where Rogues would go Strength Grapplers with a +12 bonus at level 3, throwing everyone off ledges.
I also used a Stamina System, which essentially transformed the Character's HP into Stamina. Whenever they did any action, they would lose 1hp. With this, they were able to retry most mundane checks without it becoming tedious. It also gave some stakes to each check made, and made players always aware of how much effort their putting into each task. They would regain some of their HP with resting, with as little time sink as 5 minutes.

I personally love the idea of leveling and classes, but I didn't like how 5e handled it. The fact that you can only level up into your Class made leveling boring. They also had 20 levels, and some of those levels were just a +1 stat boost.
What I did was give players options when leveling. They could choose to upgrade their Class, Race, or Charisma. There were no modifiers, and each level gave an ability that could fundamentally change how the character is played. Players playing the exact same Race/Class combo we're, generally, fundamentally different characters by level 5 based on their individual playstyles. The question of "Do you get Echolocation to see enemies through walls, or, the ability to grapple as part of an attack," got a lot of players excited about each level. My leveling system that leaned hard into width (Ability Variety and utility) rather than height (Modifiers and bonuses) was super fun and surprisingly balanced.

That being said. Its fine, and it'll do in a pinch. But its no longer innovative. If you're okay with an unskilled person being able to do something that a skilled person can't, it'll be fine, but if you think its far fetched that the group's mule knows more about hacking than a computer scientist, even for a moment, i would probably change it.

I hope this helps.


Electronic Thing
Validated User
I like DnD 5e. But the skill system is a weak point. I can say that it IS better than the 3.x system or 2nd edition. But the 5e system is still pretty meh...

It is too swingy, the bonuses too low in many cases, and the DCs set too high in many cases. For instance, it is not uncommon to encounter a DC 15 for tasks that aren't all that difficult. And even at level 12, our party's current level, a DC 15 is still often difficult to make. Take a group of 10 PCs, all needing to hit the same DC 15 in one try, and watch most of them fail.

For the d20 heartbreaker I was working on, I used something closer to 2e's NWP. It was a pared down skill list of broadly applicable skills, like 5e uses, but decoupled from attributes. Use your best attribute that will work for a given task. And it was Blackjack style roll low: roll below your attribute. The number of points you had in the skill boosted the effective attribute. So if you were trying to climb a 40' wall in a hurry, had 2 points in Athletics, and the highest relevant attribute you had (we'll say Dex) was a 15, you'd need to roll under a 17.

But maybe I'm weird in that I like starting characters that are broadly competent, instead of wildly incompetent.


Registered User
Validated User
John Out West, yes, it helps a lot, thanks. Your stamina idea is interesting, btw.

Stattick, thanks. I think a big part of starting characters competence is down to having a good GM. I don't go back for a 2nd session with a GM who forces a lot of skill checks on routine actions (unless the system is supposed to be comedic, like Paranoia or Toon something). But yes, I kinda like a Star Trek approach where if you're a PC you're pretty good and well cross-trained, even if you're at "level 1."


Registered User
Validated User
Alban, somehow I missed your comment the first time 'round, thanks for the info and suggestions. Degrees of success are a good idea, btw.
Like Alban said, a greater range would be better. One simple solution to this is make Expertise more common, treat it like advanced training in a skill. It wasn't clear if you're aiming for an 5E OGL game, but that would be a familar mechanic to incorporate more commonly if you are.

If every class can get Expertise in a particular skill or two, either through a class feature, spending two proficiency slots on the same skill, or through a feat, it makes a big impact. Your aptitude with skills then ranges from +2 to +12 from level 1 to 20 instead of +2 to +6.

The other way to increase the range in aptitude it is with features that grant advantage with particular ability checks (which mechanically, is roughly another +5, increasing the "skill bonus" range to +2 to +17 ).


up to no good
Validated User
Whole the rules to resolve skill checks is worded to roll, noting that deciding whether to roll is up to the DM, there is additional guidance to just let players succeed without bothering to roll if there isn't a point to failing just like having a skill/tool proficiency may be necessary for some tasks. The guidance could certainly be clearer.

So like real life, without any real pressure like time constraints it is perfectly reasonable to let someone autosucceed if they are proficient and/or the task isn't really difficult. Save the rolls for when under pressure or under a time limit to see if they overcome difficulty or do it fast enough. The old take 10 or take 20 rules from 3e are pretty good at estimating how long automatic success might take.


Validated User
Like many already said, I think its a great improvement over some of the previous editions (2nd and 3.x). Though, to my players' chagrin, they still fail DC 15 rolls at higher levels and get frustrated by how it breaks their immersion. After all, shouldn't the super smart wizard be able to recall basic arcana knowledge?

Maybe you could retool the math. Expanding expertise across other classes, as some said, could fix that problem while providing a way to differentiate from others in the class. Some think 5e provides little in the way of customization after you've picked your subclass, and they're probably right. If I could make my wizard super good at knowledge nature and consistently kill those rolls, I think I'd feel more confident calling my mage a master biologist.

However, I'm also curious if you could modify something other than the math. Because ultimately failure is what makes high-level 5e so much better than 3.5. I can't imagine a worse campaign than ones I ran in Pathfinder where a player always rolls over 20 on perception, another on investigation, and yet another on persuasion. Maybe, just maybe, there's something else to modify – like the severity of the failure – or the magnitude of success, without ever having to get rid of those other outcomes.
Top Bottom