Bell curve rolls, and (dis)advantage

[Cicciograna]

I am particularly fond of the 3d6 bell curve mechanics, as opposed to the straight d20. I generally appreciate the idea of favoring "average" results for dice rolls, giving thus more importance to the actual bonus attached to the roll, rather than having a flat 1/20 possibility to achieving incredible success, or total failure at whatever task at hand.

The 3d6 mechanics would be easily adapted to advantage and disadvantage rules: rolld 4d6 and drop the lowest/highest if you have advantage/disadvantage.

Now, I have only recently taken up 5th Edition, and only recently learned of "bounded accuracy": I find this a very interesting concept, and generally a good idea. How would a 3d6 bell curve with the proposed (dis)advantage mechanics interact with BA? Do you see any particular pitfalls to such an approach?

[KorvinStarmast]

I'd recommend asking that question at this place (https://rpg.stackexchange.com/). Loads of good answers on game/dice, and people who understand it.

If you'd rather not, I'll probably lift your question and ask it there. It will help me understand bounded accuracy better.

How do you intend to handle critical hits? (Only on an 18?)
The Champion's added chances for Critical Hits? (On a 17/18 at level 3?)

[Malifice]

If I were to do it I would use 2d10 instead of 3d6.

I would only use the 2d10 for ability checks. Advantage and disadvantage would be roll an extra d10 and take the highest or lowest two.

Seeing as AC can routinely get over 20, Having attack rolls made with a Bell curve would result in unhittable player characters. A standard mook with +5 or 6 to hit almost could never hit an armor class over 20.

I would also probably change expertise to a flat +3.

[Asmotherion]

This may come off as a not so desirable thing to hear, but: If it aren't broken, why fix it?

Think of this in terms of philters:

-Is this affecting your playtime as much as to become a bother to you?
-Is this more of a personal matter, or does it reflect the oppinion of all your table/other people you know? (It would be pointless to make effort to create a "fix" that will never see actual play, because everyone you know prefears the classic d20 than your solution).
-In case you think about it as a marketing solution, do you have a marketing plan? Is it solid? I wouldn't bet on changing D&D marchendise to 3d6, it's always been a d20 game classic (also, good luck with copyrights, unless you don't do it for actual profit). If you're thinking on basing a whole new game on D&D, that's an other story.

[Knaight]

I'd recommend asking that question at this place (https://rpg.stackexchange.com/). Loads of good answers on game/dice, and people who understand it.

If you'd rather not, I'll probably lift your question and ask it there. It will help me understand bounded accuracy better.

How do you intend to handle critical hits? (Only on an 18?)
The Champion's added chances for Critical Hits? (On a 17/18 at level 3?)

There's no shortage of people who understand dice probability here - and while I'm a little pressed for time and can't do an analysis right now, I can at least mention the glory that is Anydice. It's a neat little specialized programming language for dice probabilities with a range of built in displays, which can give you a pretty good idea of how mechanics actually work out.

[MrStabby]

So I am trying to work out why you would want to do this. It isn't to change the probability of success or failure - as you could more easily do that my changing the DC. Presumably it is to change the relative value of bonuses between different characters for different tasks. What this will do is it will make static bonuses much more valuable for the most challenging tasks. At the same time it will make it very difficult for any half competent PC to fail an easy task.

So the major things to worry about are probably:

1) How do you keep things interesting for people who have invested a little into skills? The barbarian who chose to not dump int, so they could be a little bit better at arcana and other knowledge checks for example - now their investment is pretty much worthless. You need to find a way to protect the interests of those people who want to be OK at things and to still preserve their chance to pull out the highest roll in a group. If you don't take care of this then you either end up with DCs that are trivial and have no drama for the best PCs in that realm or the DCs are so high that you can only contribute to that skill if you have also maxed the appropriate stat. Fewer PCs can contribute in fewer situations and the game suffers for it. Consider grapple checks - there is a massive advantage to proficiency, strength and all that but there is still a chance for the underdog, still some tension as the wizard can wriggle free for the dragon. With 3d6 rather than d20 this chance pretty much disappears and takes the tension away with it.

2) Time, rerolls/advantage. This is going to slow things down at the table. Throwing two sets of 3d6 and keeping them apart is just a little bit more fiddly than chucking a couple of d20s - do addition, no errors no need to separate the two sets.

3) Design of content and preparation. It is pretty easy to judge what is easy or hard in a d20 system - even for a wide load of characters including NPCs. This isn't going to break the game but you will need a lot more prep time for most games and it is pretty difficult to wing unexpected encounters, work out what perception modifiers are needed vs what stealth rolls and so on. Just be aware of the extra time adventure design will take.

[Cicciograna]

Note for everybody: I am not DMing any game, this is just pure speculation on my part. The idea for the 3d6 bell curve comes, of course, from the namesake variant (http://www.d20srd.org/srd/variant/adventuring/bellCurveRolls.htm) from Unearthed Arcana for 3.5, as well as the various implications for crit ranges, Taking 20, etc.


I'd recommend asking that question at this place (https://rpg.stackexchange.com/). Loads of good answers on game/dice, and people who understand it.

If you'd rather not, I'll probably lift your question and ask it there. It will help me understand bounded accuracy better.

Feel free to ask the question there, by all means. This is indeed only just an exercise in though, and I'll confess that understanding better bounded accuracy was my objective as well.



How do you intend to handle critical hits? (Only on an 18?)
The Champion's added chances for Critical Hits? (On a 17/18 at level 3?)
If I were to implement this mechanics I would follow the rules as described in the link I posted on top. However, Malefice pointed out an interesting issue...


If I were to do it I would use 2d10 instead of 3d6.

I would only use the 2d10 for ability checks. Advantage and disadvantage would be roll an extra d10 and take the highest or lowest two.

Seeing as AC can routinely get over 20, Having attack rolls made with a Bell curve would result in unhittable player characters. A standard mook with +5 or 6 to hit almost could never hit an armor class over 20.

2d10 could indeed be more suitable: it's less "bell-y" than 3d6, but at least it has the bonus of preserving the maximum value of the sample, 20.
My understanding on BA is that one of its aspects is that it is very important that numbers end up being bound by certain limits, one of these being 20 for certain scopes. Thus, the 2d10 could be better suited to the numbers as ran in 5th Edition.


This may come off as a not so desirable thing to hear, but: If it aren't broken, why fix it?

Think of this in terms of philters:

-Is this affecting your playtime as much as to become a bother to you?
-Is this more of a personal matter, or does it reflect the oppinion of all your table/other people you know? (It would be pointless to make effort to create a "fix" that will never see actual play, because everyone you know prefears the classic d20 than your solution).
-In case you think about it as a marketing solution, do you have a marketing plan? Is it solid? I wouldn't bet on changing D&D marchendise to 3d6, it's always been a d20 game classic (also, good luck with copyrights, unless you don't do it for actual profit). If you're thinking on basing a whole new game on D&D, that's an other story.
Not planning to do anything in particular, this was just some speculation on my part. I don't feel the 1d20 mechanics is broken per se, it's just that I liked (back in 3.5) the idea that when attempting something, everything, the average results would come up more often than the extreme results, be these extreme success or extreme failure. The 1d20 has the inerent 5% probability for all the outcomes and this seems somewhat artificial to me. It works, for sure, it was just a matter of personal taste.


So I am trying to work out why you would want to do this. It isn't to change the probability of success or failure - as you could more easily do that my changing the DC. Presumably it is to change the relative value of bonuses between different characters for different tasks. What this will do is it will make static bonuses much more valuable for the most challenging tasks. At the same time it will make it very difficult for any half competent PC to fail an easy task.

So the major things to worry about are probably:

1) How do you keep things interesting for people who have invested a little into skills? The barbarian who chose to not dump int, so they could be a little bit better at arcana and other knowledge checks for example - now their investment is pretty much worthless. You need to find a way to protect the interests of those people who want to be OK at things and to still preserve their chance to pull out the highest roll in a group. If you don't take care of this then you either end up with DCs that are trivial and have no drama for the best PCs in that realm or the DCs are so high that you can only contribute to that skill if you have also maxed the appropriate stat. Fewer PCs can contribute in fewer situations and the game suffers for it. Consider grapple checks - there is a massive advantage to proficiency, strength and all that but there is still a chance for the underdog, still some tension as the wizard can wriggle free for the dragon. With 3d6 rather than d20 this chance pretty much disappears and takes the tension away with it.

And yes, this convinced me that it would not be an especially good idea, or rather that it would possibly be against the spirit of bounded accuracy. As you say, with the 3d6 - or perhaps better the 2d10, one would be gimping the probability for the underdog to achieve an outstanding result, whereas one of the tenets of BA is that even Goblin McGoblinface should have a possibility, as lowly as it can be, of stabbing Lord Almight Von Uberful Emperor of the Universe, even if for just 1 hp of damage (or none at all).
As stated in the 3.5 SRD in the "Metagame analysis" panel for the bell curve rolls,

Another subtle change to the game is that the bell curve variant awards bonuses relatively more and the die roll relatively less, simply because the die roll is almost always within a few points of 10. A character’s skill ranks, ability scores, and gear have a much bigger impact on success and failure than they do in the standard d20 rules.
Emphasis mine. BA has (in my still rudimentary understanding of it) somewhat reduced the focus of the bolded aspect of the game - at least when it comes to skill ranks and gear, meaning that the die roll plays instead a bigger role in the determination of success/failure: in all the examples that I have found online pitting the 1st level character to the 20th level demigod, the former has usually a bonus of +5 to whatever he's trying to accomplish, coming from ability score and proficiency, versus a +11/+12 of the demigod (+5 from ability score, +6 from proficiency and an eventual +1 from magic items), meaning that the dice roll in whatever challenge between the two will play a big role. Going with a system that favors average values could severely lower whatever already low probability the 1st level character could have against the demigod.

[Diebo]

As others pointed out, bounded accuracy for armor class would require some rebalancing.

It is easy enough to balance critical hits - 4.63% of the time crit on a 16, 17, or 18, and 9.26% on a 15-18. You could do the same for critical miss.

For AC, it is a bit harder. Take the following examples:

Orc, attacking AC 16, has +5 to hit. Both methods (3d6 and 1d20) result in hitting 50% of the time. Now consider someone with 21 AC. Using 1d20, orc hits on a 16-20, or 25% of the time. With 3d6, orc hits only 9.26% of the time.

Hill Giant attacking AC 16, has +8 to hit. The hill giant hits easier with 3d6 (84% of the time, versus 65% for 1d20). Ouch! But flip things with 21 AC. Hill giant with 3d6 only his 26% of the time versus 40 percent with 1d20. Just +2 more AC (to 23) makes the giant only hit 10% of the time in 3d6 versus 30% in 1d20.

3d6 REALLY benefits any extra points put into AC. If you are a DM, perhaps limit magical armor.

To help monsters, you could boost things, perhaps using the monster's CR rating. For example, increase monster AC and attack bonus by 1/2 their CR rounded down.

[napoleon_in_rag]

One of the things I liked about BECMI is how it used 2d6 + CHA tables to resolve social situations. For example, when negotiating with a merchant over price:






2
Merchant is insulted, refuses to sell item. -2 to future negotiations with merchant.


3-5
Failed negotiation. Price is now 150%


6-8
Deadlocked negotiation. No change in price


9-11
Successful Negotiation. Price is now 50%


12
Merchant is impressed. Price is now 50%. +2 to future negotiations with merchant.





What I like is that it is not a pure pass/fail roll like in 5e. That makes it more interesting. This can easily be translated into 5e by using 2d10. It might look something like this.






2
Merchant is insulted, refuses to sell item. -4 to future negotiations with all merchants in town/guild.


3-4
Merchant is insulted, refuses to sell item. -4 to future negotiations with merchant.


5-8
Failed negotiation. Price is now 150%


9-13
Deadlocked negotiation. No change in price


14-17
Successful Negotiation. Price is now 50%


18-19
Merchant is impressed. Price is now 50%. +4 to future negotiations with merchant.


20
Merchant is impressed. Price is now 50%. +4 to future negotiations with all merchants in town/guild.