3d6

[Talakeal]

My major complaint about 5e so far is that the game is really too swingy, bounded accuracy went a bit too far and made the game more about the dice roll than character skill, particularly in the case of skill checks or NPC on NPC fights (including summoned minions).

What would happen if instead of running the game on 1d20 you instead rolled 3d6? This gives you a bell curve with most results in the middle as well as a higher minimum and lower maximum roll. This seems like it would eliminate the swinginess while still keeping the averages the same.

I can see a couple problems with this, mostly involving critical hits becoming rarer (and thus abilities that affect them becoming either more or less powerful) and at high levels when you get a bunch of attacks each round it might be hard to keep all your dice straight, but those are fairly easy fixes.

Anyone have an opinion on this? Anyone who knows 5E better than I do see some problems (or benefits) that I do not?


Also, another problem with 5E IMO is the all or nothing nature of advantage or disadvantage. With this system every source of advantage could allow you to roll an extra dice but only keep the best three, and every source of disadvantage could cause you to roll an extra dice and only keep the worst three (advantage and disadvantage dice would cancel each other out on a one for one basis if you have both.)

What do you think?

[Grod_The_Giant]

I tossed around this stuff a while back. My conclusion was that 3d6 or 2d10 would work but feel off. My preferred solution is to approximately double Proficiency and add half the value to AC and nonproficient saves; that improves matters considerably vis-a-vis reducing the impact of the d20.

[JoeJ]

Using 3d6, if you allow a critical hit on an unmodified roll of 16 or better you'll have approximate the same chance as an unmodified 20 on a d20. A 15 or better would give you roughly the same as a champion fighter's Improved Critical ability, and a 14 or better would approximate Superior Critical.

[Grod_The_Giant]

Using 3d6, if you allow a critical hit on an unmodified roll of 16 or better you'll have approximate the same chance as an unmodified 20 on a d20. A 15 or better would give you roughly the same as a champion fighter's Improved Critical ability, and a 14 or better would approximate Superior Critical.
Alternately, rolling double 6's has a pretty comparable probability to a natural 20 (7.41% vs 5%).

[2D8HP]

5d4 would work as well (and I hardly get to use them anymore).
I like this "bell curve" option.
Thanks!

[Foxhound438]

only issue becomes that it disproportionately rewards stacking up AC, since monsters having to roll 15+ on 3d6 makes PC's nigh invincible. Be prepared to see everyone have plate and shield with defense style and shield of faith. The end result there is that combat would take a lot longer.

[Jamesps]

5d4 would work as well (and I hardly get to use them anymore).
I like this "bell curve" option.
Thanks!

You might run into trouble as your bell curve will be centered around 12.5 instead of 10.5, making success on most rolls easier.

[Warwick]

3d6 shortens the distribution, obviously. It also makes bonuses have non-uniform impacts. Whereas a +1 bonus right now is an additional flat 5% chance to succeed on a standard roll, what that +1 bonus does for you under 3d6 depends on what your skill bonus is and what your target is. Though that will be true for any non-flat RNG.

I kind of like the idea of 3d20, take middle. Same range of results, but it normalizes the distribution. Also syncs easily with advantage/disadvantage. Never actually used it, though, so it's purely speculative.

[Jamesps]

3d6 shortens the distribution, obviously. It also makes bonuses have non-uniform impacts. Whereas a +1 bonus right now is an additional flat 5% chance to succeed on a standard roll, what that +1 bonus does for you under 3d6 depends on what your skill bonus is and what your target is. Though that will be true for any non-flat RNG.

I kind of like the idea of 3d20, take middle. Same range of results, but it normalizes the distribution. Also syncs easily with advantage/disadvantage. Never actually used it, though, so it's purely speculative.

The 3d20 method will still have the same issue with bonuses meaning different things depending on the target number.

Granted, that's only a problem if you define it as such. I do like how it syncs with advantage/disadvantage though. Nice and simple.

[Segev]

You could do 2d20, take the average. The curve looks like this (http://anydice.com/program/8154). Well, that's for rounding down; you could make them round up and slosh it slightly towards the higher end, or have them actually be allowed to get values in the .5 range (resolves more ties). It keeps the average around 10, and has the advantage of making them roll 2d20 all the time.

When they have Advantage, they instead take the better one. When they have Disadvantage, they take the worse one.

[eastmabl]

The benefit of using 3d6 is that Wizards has already instructed us how to interpret bell curve rolling in a d20 system.

http://www.d20srd.org/srd/variant/adventuring/bellCurveRolls.htm

I'll put my 2 cp in a spoiler so you can choose to ignore it.

Personally, I don't mind the swingy-ness of 5e. Instead, I've just changed my definition of failure in non-combat situations. I use the "fail forward" technique sometimes, where failure can be success at a cost. Other times I'll use technique of "save or suck" spells and will treat the first failed check as the first step towards the colossal failure.

[Segev]

I'll put my 2 cp in a spoiler so you can choose to ignore it.

Personally, I don't mind the swingy-ness of 5e. Instead, I've just changed my definition of failure in non-combat situations. I use the "fail forward" technique sometimes, where failure can be success at a cost. Other times I'll use technique of "save or suck" spells and will treat the first failed check as the first step towards the colossal failure.

Another thing to do is consider that you should only have them roll when success is in question. If they surely can't or definitely can do something - or SHOULD be able to if the game is more or less grinding to a halt if they don't - then just let them narrate it.

Also, you can have failure not be "you weren't good enough," but rather "and a bird pooped on your hand just as you were about to pick his pocket, causing you to flinch and jerk away" or something less silly but equally interfering. I've had repeated failures to use a tool with a particular skill in a Rifts game be translated as the stupid tool being broken and it taking me that long to realize it, for instance.

[Grod_The_Giant]

Okay, let's try to keep the side arguments about whether or not 5e is too swingy/acceptably swingy/functional/dysfunctional to one of the, oh, four ongoing threads about that?

[Warwick]

The 3d20 method will still have the same issue with bonuses meaning different things depending on the target number.


Yeah. If you want a curved distribution, bonuses will have a variable impact and that's pretty much how it's going to be. 3d20 (pick middle) was just an alternative to sum(3d6) that doesn't require recalibrating all your numbers to accommodate a different range. And it doesn't require us to rename the system.

I'd personally only use it for skill checks, since combat is already a (relatively) large number of checks, so things will mostly shake out as the averages would suggest. Skill checks are more likely to be a single roll and consequently more prone to generating weird results where the veteran ninja trips over his shoelaces into a closet full of pans because he muffed the stealth roll.