How feasible is it to convert the d20 to 3d6 in D&D 5e?

Advantage and crits are going to be a bigger problem than you think because of the difference between result distributions.

>>50127750
You want to introduce the Bell Curve into 5E? Good luck?
http://www.d20srd.org/srd/variant/adventuring/bellCurveRolls.htm

>>50127750
the average of 3d6 occurs a lot more often than 1d20

you'll have many many more average and near average rolls than you will good or bad rolls

I think its a bad choice, it removes some of the randomness and fun

>>50127822
For advantage I'm thinking either roll 4d6 drop lowest or simply roll 3d6 twice, sadly i'm not
enough into statistics to do the calculations to know the characteristics of each.
You raise a good point for crits which ties us to...

>>50127838
That's kind of the point, to have a more 'realistic' representation of character skill.
Also the rarity of the extremes should allows for more impactful consequences which in turn will compensate for reduced occurence. However, rolling crits feels good and might not be worse that the rarer but epic-er counterpart, I'm sure there are players which prefer the constant, low effort gratification.

>>50127750
>What do you guys think?
You are insane or don't know even basic math.

>>50128080
i studied math, don't listen to this moron, OP.

>>50128106
Good for you. Maybe you should tell OP that he will need to rebalance not only magic items and proficiencies but also every other instant where a modifier is added to a d20 roll ?

>>50128156
this is much less of a problem with bonded accuracy. GURPS has a much wider range to the upside

>>50128778
GURPS doesn't care much about heroics (at least in base) and has other ways of keeping it's shit together. Still high numbers is a problem even there.

D&D is not designed to work with 3d6. For example you'll need to remake all armors. And even if you ignore armor ability to easily have +5 modifier over your opponents is insane with 3d6.

>>50127750
Do you want there to be NO crits, or just fewer?
The odds of getting a straight 3, or a straigt 18 will both be 1/216 - less than half of a percent.
If that is the required roll for the crits, that's really the same as "they don't exist", functionally.

>>50128917
Although, I would recommend trying it if you want to. You can always bug-fix and work out kinks as you go along, as long you make it clear to the players that you WILL be doing this.

>>50128917
If you make it so that a crit is "17 or higher", with the 3d6 it will mean a 4/216, or 1/54, or just around 2% chance of crit.
It would be the same, for crit-fails, if you called it as "a 4 or lower"

>>50128917
As stated >>50128048, I'm thinking of re-flavouring crits. The logical conclusion probably being a vorpal-like ability to instantly kill a target. Alternatively if in practice it does not feel right I can just change it to a range of numbers as >>50129095 said.

>>50128829
Actually I think I overstated the problem of modifiers. Since the whole curve scales with the bonus, it should give me a good, instinctive idea on how to modulate monster bonuses and stats.

>>50127750
On average, a roll with advantage is 3.325 points higher than a roll without. This figure deceives us because about half of all rolls with advantage will be 15 or higher (median). Success matters for d20 rolls, rather than the number other than 20.
The chance of a critical hit with advantage is an impressive 9.75%, rather than 5%.

We all know that 3d6 has a mean and median value of 10.5. 4d6, drop lowest has an average result about 12.25 with a median of about 13.5 (lazily-calculated number, but should be about right).

Obviously, critical hits are an issue here, as the chances of rolling an 18 are 1/216, rather than 1/20. I'm frankly too lazy to calculate the "advantage" figure on this, but we know it's not pretty.

To make crits work with 3d6, I would propose something like... if you have 2 6's, then crit! The odds are 7.41%, 13.19% with "advantage."
Including the provision "if the attack hits," into your crit calculations should approximate d20 crit chance.

As for +X bonuses from magic an proficiency... I assumed you wanted a normal curve for the base distribution. Low numbers will be significantly less likely to succeed than d20, just as high numbers will be significantly more likely to succeed than d20.

Comparative worst case example: players are fighting against full plate + shield in tier 1 (levels 1-4).
Attacks have +5 to hit (3 skill, 2 proficiency)
Target has 20 AC.
D20: attacks have 6/20, or 30% chance to hit.
3d6: attacks have 20/216, or 9.26% chance to hit.
Full plate + shield is worth about 3x more "staying power" versus basic attacks than in core.

The moral of the story is that core mechanic conversions are really fuckin' intense and require a lot of consideration.

You need to understand that on a bell curve, cumulative bonuses have exponentially greater benefits

So a +2 would be like 10x better than a +1, and a +3 would be like 100x better than a +2

>>50128048
>That's kind of the point, to have a more 'realistic' representation of character skill.
The problem with this is that NOTHING else in the game is there to support any kind of realism.

Like a few others have suggested, critting on dubs is a good way to convert. The AGE system has a stunt table system you can easily swap in with a few adjustments.

>>50127750
Why? Rule sets that use a D6 are all over the place. A lot of work that your players are just going to hate

>>50129690
>You need to understand that on a bell curve, cumulative bonuses have exponentially greater benefits
totally not exaggerated.
and yet that is how GURPS does it

>>50129668
This is the most useful post so take a look OP >>50127750

>>50128048
>to have a more 'realistic' representation of character skill.
I was going to ask what the purpose is, but if this is all: use a system that already does that. D&D has basically nothing to support it and even if you work on it for a decade you won't come close to a balanced end product. The monsters, classes, races, everything needs to be rebalanced and then what reason is there left to use D&D?

>>50127750
Why would you want to?

>>50133445
the closest to this 3d6 realism is GURPS. is there a direct D&D port for GURPS?

alternatively there is MYFAROG

>>50133809
GURPS Dungeon Fantasy.

File: charidimos-bitsakakis-charidimos-bitsakakis-drgn2-logo.jpg (386KB, 1701x1025px) Image search: [Google]

386KB, 1701x1025px

Thanks everyone, great ideas and counterpoints. After some consideration I find 3d6 could work nicely by rescaling the difficulty scale as such: 11(easy), 14(moderate), 17(hard), 20(very hard). This is not quite as sharp as the standard scale so a bit easier at lower +Xs. Since st, skills and attack rolls are functionally equivalent everything can abide to this scale so not much fiddling is required.

>>50134256
honestly, glance over how GURPS does. just to get an idea of proper scaling and looking for potential pitfalls. people have developed 3d6 system before you.

>>50129410
Attack bonuses scale faster and higher the. Defense bonuses in 5e, this will be even more prominent on a 3d6.

>>50127750
I can see tons of problems with it. At which point I'd just run AGE.

File: d8 vs 3d6 prob graph.png (25KB, 768x655px) Image search: [Google]

25KB, 768x655px

>>50127750
The distribution of 3d6 is actually much closer to that of a d8 or a d10 than a d20. Each 1 pt. move away from the middle of the range (where you have a 50/50 chance of success) is worth and average about 2 pt. on a d20. So a +1 sword is giving you the equivalent of +2 to hit if you're using 3d6.

If monster has high ac and you need to roll high to hit then your players are fucked

They will get average rolls more often but wont get the hogh rolls needed against certain monsters

File: 3d6 v d20 with double-size steps (smaller).jpg (50KB, 669x428px) Image search: [Google]

50KB, 669x428px

>>50136011
Here's a comparison of 3d6 and a d20, with each "step" being 1 point on 3d6 and 2 points on the d20. In other words, both systems have the same chance to roll an 11 or over. 1 step up from that is 12 (11+1) on 3d6 and 13 (11+2) on the d20. Another step up is 13 (11+1+1) on 3d6 and 15 on a d20 (11+2+2).

>>50136057
>1 step up from that is 12 (11+1) on 3d6 and 13 (11+2) on the d20.
Err, That would be 1 step down. I'm used to calculating on a roll under system. But you get the idea.