Replacing the d20 for skill checks/tests

[windgate]

A bit of context, the only Tabletop RPGS I have any real experience with are 4th and 5th edition D&D. I have roughly equal amounts of time spent with both.


Skill checks in both editions have always kinda bugged me in that they both felt overly skewed in opposite directions. Namely the contribution of the d20 roll versus skill roll modifiers.

Fourth Edition: IIRC, Every skill gained 1/2 your level as a bonus modifier (plus ability score, proficiency, specific powers and magic items). This lead to odd scenarios where the high level anemic wizard is breaking down doors a low level barbarian could not dent.

Fifth Edition: The 1/2 level thing is gone and replaced by a significantly weaker potential expertise. The DM needs to account for the random chance of any PC rolling a 20 on any given skill check. You want to be able reward players for investing in a skill heavily but the potential impact of a 20 being rolled by a person with no +modifiers makes it hard to set DC's effectively (at least for me)




I am considering experimenting with replacing the d20 with a smaller dice on skill checks for an upcoming homebrew session. Not sure what die to use, which is why I am bringing it up here.

My goals are as following:

(1) The average character (a maximum +2 bonus to the skill)
Should have a reasonable chance of achieving a moderate difficulty task with a small chance of achieving something "hard"

(2) A gifted character (High Ability score/proficiency)
Should reliably achieve something "moderate" with a good chance at something "Hard". At the top end, "very hard" successes might happen but are really rare

(3) The focused character (Max stat + Expertise)
Goes one step further than the gifted. They are the only one that can pull off something considered nearly impossible. They don't even bother rolling for moderate checks.


As an example:

Intimidation check vs:

(0) Very Easy: Peasant
(5) Easy: Random town thug or 1/4 CR humanoid/beast.
(10) Moderate: NPC Soldier
(15) Hard: Captain of the Guard
(20) Very Hard: Hill Giant
(25) Nearly impossible: Adult Red Dragon
(30) Requires Plot Bonuses: Ancient Red Dragon

I was thinking of using the above if the dice used was a d10. not sure about balance issues caused. At minimum I would probably need to 1/2 the values for the rogues 11th level feature. perhaps 2d6 instead?

[Kane0]

I'd say use 2d10 instead of 1d20 for ability checks. If you have advantage, roll 3d10 and use highest two results; if you have disadvantage roll 3d10 and use lower two results.

[OldTrees1]

It is useful to identify why you don't like the 1d20 for 5E. That will help inform how you change the RNG.

For me I dislike the relative lack of progression or differentiation. If characters vary by merely 7 points at 1st level, and a 20th level character only improved by 7 points, then I felt the 1d20 was too long a range. However I had no issue with it being a flat curve instead of a bell curve. Based on that I should probably either double the bonuses (and thus double any difference between bonuses) or halve the die (use 1d10+5 instead of 1d20).

Others don't mind the lack of progression but prefer to use a bell curve to make the outliers less common.

From the opening post it seems you had 2 critiques you highlighted.
1) A high level 4E character being better at their weakness than a low level character is at their specialty. This is due to a level scaling bonus that ignores whether the character is trying to improve at that task.
2) An unskilled 5E character regularly and reliably lucking out an besting the 5E specialist. This is due to the flat curve and ratio of bonus growth / size of the RNG.

It sounds like your critiques are similar to my issues. I think these 4 dice options might be relevant:
1d10 (Reliable Talent is a 5)
1d10+5 (Reliable Talent is a 10)
2d6 (Reliable Talent is a 7)
3d4 (Reliable Talent is a 7)

The benefit of 1d10+5 is you can use the same DC ranges as 5E (since 1d20 and 1d10+5 average 10.5). However it includes adding 5 to the die roll. Using 1d10+0 is simpler but means the DM would be converting from the 5E range to their DC-5 range.

If you like bell curves then usually 3 dice is an ideal starting place. 3d4 and 2d6 both work as bell curves that shrink the range of the RNG. In these cases the average is around 7 instead of 1d20's 10.5. You could convert DCs by subtracting 3, or maybe subtract 5 because it is easier even if it buffs the PCs.


If I had to guess, I think 1d10 (Reliable Talent is a 5) is a good fit for your concerns.


Balance Concerns:
Opposed Checks. There is some merit in having the skill roll modifier have less impact in opposed check so underdogs have some theoretical chance of surviving. Imagine a Mind Flayer grappling a PC. You want them to have a chance to escape the grapple. This might be a case where keeping the 1d20, or replacing it with 2d10/3d6 fits balance better.

Other than that there are no balance concerns.


PS: I don't recommend replacing Reliable Talent with 2d6. The main benefit of Reliable Talent is its floor. If using 1d10, then Reliable Talent would translate to a floor of a 5. If you replace it with 2d6 then the floor is a 2 (-3 penalty to Rogues) even if the average is 7.

[Elves]

- play 4e without the +1/2 level bonus to everything
- in 5e give people an additional +2 on trained skills
- play 3e