The advantage die - anyone else tried this ?

[MoutonRustique]

I will be converting [combat advantage] into an [advantage die] (d4) - the idea is simply to help with bookkeeping a little bit and also adding a more tangible aspect to getting combat advantage.

It's only this : when you have [combat advantage], you get a d4 to add to your attack roll. That's it.

Anyone else using this little trick? Thoughts?

[Wasteomana]

I will be converting [combat advantage] into an [advantage die] (d4) - the idea is simply to help with bookkeeping a little bit and also adding a more tangible aspect to getting combat advantage.

It's only this : when you have [combat advantage], you get a d4 to add to your attack roll. That's it.

Anyone else using this little trick? Thoughts?

I wouldn't use it in any game I played in 4e, but if you want to use it at your table have at it. How would that work for things like Nimble Blade? Would it become a d6?

Personally I like how CA is a flat number in 4e, but as long as all your players know you are doing this up front and you are able to deal with the issues it causes, enjoy!

[ThePurple]

Anyone else using this little trick? Thoughts?

It tried it for a bit and, while I and my players enjoyed the additional level of unpredictability (since you could no longer guarantee the payout), it ended up being a lot more work than it was worth to do the additional typing/math. After a couple games of doing it that we, we pretty quickly got rid of it.

If 4e used 3d6 instead of d20, an awesome way to do advantage is adding an additional d6 but still only taking 3. It doesn't increase the top end, but has a very noticeable effect.

I *really* want to find a way to use 3d6 instead of d20 since I like the options it opens up (4d6 take 3, straights, 2 of a kind, 3 of a kind, explosive dice, mayhem dice), but it's just *way* too hard to get a default critical in those circumstances. I have a feeling it would work, but you'd need to make a lot of those "get a crit in other interesting ways" mechanics so common that players often end up with multiples.

[Wasteomana]

It tried it for a bit and, while I and my players enjoyed the additional level of unpredictability (since you could no longer guarantee the payout), it ended up being a lot more work than it was worth to do the additional typing/math. After a couple games of doing it that we, we pretty quickly got rid of it.

If 4e used 3d6 instead of d20, an awesome way to do advantage is adding an additional d6 but still only taking 3. It doesn't increase the top end, but has a very noticeable effect.

I *really* want to find a way to use 3d6 instead of d20 since I like the options it opens up (4d6 take 3, straights, 2 of a kind, 3 of a kind, explosive dice, mayhem dice), but it's just *way* too hard to get a default critical in those circumstances. I have a feeling it would work, but you'd need to make a lot of those "get a crit in other interesting ways" mechanics so common that players often end up with multiples.

The amount of effort required to turn 4e into a d6 system would be somewhat insane. But I would also prefer it.

[MoutonRustique]

Huh. I'd never have guessed that it might cause additional work - were you guys playing with software ?

But it's good to know that it did! Thanks for the heads up.

Yeah, 3d6 (or 3dX) is incredibly enticing... I have a personal fondness for 2d10 in theory (because 1-10 is intuitive), and 2or3d8 (because I like d8 the best!) There's also 2d12 (because d12 need the love...)

If you dial back on what can be added-on to crits (since the base 4e crit is quite manageable), I imagine that, on 2d10, having a crit when you get doubles wouldn't be too crazy (1 in 10). I wonder just how hard the game would snap by using 2d10 in this way... ?

On the other hand, 4e is already fairly "tight" in it's range - and it also has a massive base range! - so having bell curves might make it even stricter! Perhaps this kind of thing would be better applied to something like 13th Age, where the range has been purposefully narrowed?

[ThePurple]

Huh. I'd never have guessed that it might cause additional work - were you guys playing with software ?

We played it on roll20. It just got annoying to keep adding the additional +d4 to a roll (and, if someone forgot that they had advantage for the roll, they'd have to roll a d4 on top of the previous total rather than adding 2).

Yeah, 3d6 (or 3dX) is incredibly enticing... I have a personal fondness for 2d10 in theory (because 1-10 is intuitive), and 2or3d8 (because I like d8 the best!) There's also 2d12 (because d12 need the love...)

2dX is nice but it doesn't allow for straights or 3 of a kind (which are the most interesting dice effects imo) and makes additional dice to select from a lot more powerful (whether explosive or to choose from). It just doesn't play *nearly* as well with all of those interesting dice effects as 3d6 does (or 3d8 even).

If you dial back on what can be added-on to crits (since the base 4e crit is quite manageable), I imagine that, on 2d10, having a crit when you get doubles wouldn't be too crazy (1 in 10). I wonder just how hard the game would snap by using 2d10 in this way... ?

A 10% chance of automatically hitting though? If you require it to be a double that would also hit the target, it brings it down to the same ~5% that 4e manages by default though you still have to deal with additional dice advantage being so much stronger in this context since it's increasing hit chance and crit chance *significantly*.

On the other hand, 4e is already fairly "tight" in it's range - and it also has a massive base range! - so having bell curves might make it even stricter! Perhaps this kind of thing would be better applied to something like 13th Age, where the range has been purposefully narrowed?

That's one of the biggest problems I've seen. The bell curve makes everything a lot more average so fighting enemies that are even moderately higher or lower level than you gets completely screwed up. If we design NPCs such that PCs have a ~66% chance to hit an even level enemy by default, that would mean they require either a 9+ (74.07% chance) or 10+ (62.5%) in order to succeed; with simple scaling (which is a major advantage of 4e, and is something we definitely want to keep around), those chances drop to 11+ (50%) and 12+ (37.5%), which is a *huge* change. And 2 levels below translates into 90.74% (7+) and 83.79% (8+), which is basically trivial.

The bell curve allows for a lot of *really* cool stuff, but it normalizes performance *so much* that it narrows the level range that you're allow to use so much it's almost painful. If you're willing to having a game where it's only appropriate to fight enemies within 1 level of you, it could work but, as a GM, I like having a bit more variability in performance than that. It also makes penalties and bonuses to those 3d6 rolls *insanely powerful*; you'd basically need to turn all of them into +1 rather than +2 mods and then prevent stacking for a *vast* majority of them. It also requires *significantly* tightening up ability mods for all chars since any deviation the expected math is going to have *massive* repercussions.

I have a feeling that there's a way to make the 3d6 math work *somehow* in a game with the simple NPC math and balance PC math of 4e, but it's going to require a lot of finagling with special dice mechanics in order to normalize it, and I just don't have that energy atm, sadly.

[Kurald Galain]

This reminds me of the "step dice" mechanic from Alternity (D&D in spaaaace, that also has you roll d20+d4 or d20+d6 for most tasks; or sometimes d20-d4).

My experience is that it substantially slows down gameplay for no noticeable benefit. If you want to 'curve' your die rolls, replacing 1d20 by 2d10 or 3d6 has a better effect.

[Yakk]

3d6 in a roll over/under game has no in-game noticable "curve" outside of the 5% "only hit hit/miss on a critical" cases.

The effect of 3d6 vs d20 is almost completely captured by changes in standard deviation.

This change just changes the "size of modifiers" and "Distance DCs are from average" by a linear factor.

The remaining "curve" effect is tiny.

If you graph 3d6 "roll to hit a target" vs d20 "roll to hit a target", you get a graph that looks impressively different and is utterly irrelevant because you are rolling to hit "at or under a target". The "roll to hit a target" is a graph of the derivative of what you care about.

Almost nobody believes me when I say this unless I pull out graphs and spend multiple posts on explaining it. Can we skip that this time and just take it on faith?

[Jaeda]

The multiple dice also lets you try out some meta-mechanics. The Planet Mercenary RPG uses 3d6, but one of them is colored differently (red) which causes MAYHEM if it is high. Some of the guns also have special effects on things like doubles of the other two blue dice or if you get a straight and stuff like that.

Some other games allow you to have extra benefits or complications separate from the success or failure of the roll. You can make Good Stuff happen if the red die is high and Bad Stuff happen if the red die is low. If the advantage die was a second red die, you would get Good Stuff more often in advantageous situations, and Bad Stuff more often in detrimental situations. Stacking multiple advantages might not be too bad if they all had to be the red die; you essentially guarantee Good Stuff by getting a red 6, but the blue dice are still random and can tank your roll.

[Yakk]

Yes, having mechanics tied to something other than the total can make multiple dice quite fun.

I'm just objecting to people rolling multiple dice, adding them up, and hand-waving about how "bell curves make it better".

Bell curves work reasonably well on roll to hit a target; for example, old school D&D encounter charts. Well, at least a bit better than roll-under, because integration isn't smoothing the bell curve out.

[Jaeda]

It took me reading that a few times to understand the distinction you were trying to make. The chance of getting exactly X on multiple dice is represented by a bell-curve, but we are usually interested in getting at least X, which is represented by an S-curve; the probability changes incredibly quickly near the middle, but very slowly near the extremes.

[Yakk]

Yes, and that S-curve can be approximated to a ridiculous degree of accuracy by a strait line. The slope is a bit steeper on the 3d6 than d20.

The only parts that are off are the "tails" at the end, which correspond to rolling a 1 or a 20 on a d20.

[ThePurple]

The only parts that are off are the "tails" at the end, which correspond to rolling a 1 or a 20 on a d20.

If your goal is to have a roughly 66% chance to hit an even level target (the DCs for a 3d6 need to be 1 lower than the same ones for 1d20), the chance drop off becomes pretty extreme pretty quickly when you start fighting higher level enemies. For +2 enemies, the chance to hit for a d20 is 55% while 3d6 manages 37.5%; for +3 enemies, you get 50% and 25.93%. It doesn't take much deviation from an even level fight in order for the differences between the flat line progression of the d20 and the s-curve of the 3d6 to start generating significant problems. 4e default requires engaging foes within a comparatively small level range (-2 to +4) and requires consistent and predictable increases in bonuses, but, with the 3d6, this shrinks it even more (-1 to +2) and doesn't easily accommodate the natural variation in gear acquisition that will occur since enhancement bonuses are *even more important* to acquire early.

I appreciate the elegance of 3d6, but it's farcical to claim that the negative impacts only appear when you're 3 standard deviations away from the mean. Going from a 10+ to a 13+ to hit reduces your chance to hit by half.

Now, if we're going to just admit that players aren't supposed to fight anything outside of the even more narrow range that 4e already uses, this could work without much problem, though, in order to actually provide players with a full range of foes, you're going to have to basically admit that each adventure will require a unique set of enemies created specifically for it (at the very least, you'll need to create 2 tiers of each type of enemy that you previously only needed to create 1 for).

[Kurald Galain]

As I recall, one of the advantages of using 2d10 or 3d6 is that it makes a more skilled character more likely to beat an unskilled character at the same task. With the normal skill system it happens too often (IMO) that e.g. the friendly untrained cleric beats the menacing trained bard at an intimidate check, so reducing that is a good thing for me.

[ThePurple]

As I recall, one of the advantages of using 2d10 or 3d6 is that it makes a more skilled character more likely to beat an unskilled character at the same task. With the normal skill system it happens too often (IMO) that e.g. the friendly untrained cleric beats the menacing trained bard at an intimidate check, so reducing that is a good thing for me.

First off, we'll determine the exact likelihood of those events happening in the first place before we make a decision as to whether it "happens too often" or not.

Let's assume that the "friendly untrained cleric" has a CHA of 10 (since WIS is on the same NAD, it's not likely to be a stat that is heavily invested in) and the "menacing trained bard" has a CHA of 18 (CHA is the primary stat so it makes sense). This gives a +9 bonus to the bard and a +0 bonus to the cleric (this is also the *best case* since the cleric's mod is only going to go up by half-level while the bard's is going to go up based upon half-level, ability score increases, and, in all likelihood, some gear and powers as well).

The chances of a d20+0 beating a d20+9 are 13.75%. The chances of a 3d6+0 beating a 3d6+9 are 5.40%. It reduces the chances of it happening by ~61%. While it seems like a lot, from a practical standpoint, it's not that big of a difference since it's already pretty uncommon (and will tend to only get worse at higher levels; remember: this is a *best case* for the cleric). Once you hit the paragon tier, it's basically *impossible* (cleric has to crit and bard has to fumble, even then, since crits and fumbles aren't automatic successes, the cleric would never be capable of matching the bard) because the bard's bonus so quickly outpaces the cleric's, regardless of whether you're using a d20 or a 3d6.

If you're remembered enough cases where the low-stat untrained character beat the high-stat trained character at a skill check, I have to wonder about those memories because, chances are, they didn't happen as often as you seem to think (I could go into the psychology of why this is, but this isn't the place) since it's pretty uncommon at low level and basically impossible at higher levels.

This does bring us into the discussion of whether an untrained and/or low score character should even have a *chance* at a task that would challenge a trained and optimized character. Some would argue that it doesn't make sense. Others would point out that even novices have taken out living legends with all the skill in the world in a single test. Luck is a *huge* factor in life, which is why the only way to get an accurate judge of a person's capabilities, you need to put them through a battery of tests rather than a single one.

And, as I previously stated, if you dislike luck having a chance of spoiling the most likely outcome, why are you playing a game that uses random numbers in the first place? The entire point of random numbers is that things don't *always* go as expected.

Of course, all of this is why I champion getting rid of binary success altogether and adding a separate roll for quality of success: an untrained person might be able to succeed only slightly less often than a trained individual, but a trained individual will not only succeed more often but, more importantly, virtually *always* generate more success than an untrained individual. An untrained carpenter might be able to make a table, but a trained carpenter will do it faster, better, and/or more efficiently. The same is true of most skills, in fact.

[Kurald Galain]

The chances of a d20+0 beating a d20+9 are 13.75%. The chances of a 3d6+0 beating a 3d6+9 are 5.40%. It reduces the chances of it happening by ~61%. While it seems like a lot, from a practical standpoint, it's not that big of a difference since it's already pretty uncommon

One out of seven is not what I would call uncommon.

Anyway I'm fine with randomness in general, but I prefer it to be a smaller factor than e.g. the bonus for skill training.

[Yakk]

3d6 makes modifiers bigger and makes target numbers vary more away from 10.5. So yes +1 or +3 means more in 3d6 vs in d20.

Var(KdX)=K*(X-1)(X+1)/12
SD(KdD)~=X*sqrt(K/12)
SD(3d6)~=6*sqrt(1/4)=3
SD(1d20)~=20*sqrt(1/12)~=5.8

The standard drviation of d20 is about twice the standard deviation of 3d6. So in 3d6, a +1 to DC or skill modifier is worth about 2x what it is in d20.

Or, you can emulate using 3d6 4e by making d20 scale at a rate of *2* per level instead of 1, and keep the d20.

[ThePurple]

One out of seven is not what I would call uncommon.

One out of seven *at best*. Keep that in mind. It's at the *very start* of the game. Once you start getting into higher levels, that 1/7 precipitously drops as you level up; by mid-heroic, it drops to ~1/11 (+10 difference) or ~1/15 (+11 difference).

For me, 1/7 is definitely "uncommon", since it's ~1 standard deviation out of the norm, which I see as it should be for an untrained person just starting their career outperforming a specialist just starting their career simply by random chance (in a single action).

Of course, now we're getting into perception and opinion on what "the best" ratio necessary for a highly trained person to demonstrate their superiority is (the behaviorist in me feels compelled to bring up that a 1/5-1/6 ratio is right around the point where people feel that they have a *chance* even if it's terrible; less than that and people tend to become extremely unmotivated to actually try).