Advantage is Fun. Here's a Better Alternative.

This might be a long post, so I'll skip right to my idea and then go back and explain how I got there:

Advantage should revert to flat +2/-2 modifiers, and the 2d20 mechanic should be moved into class choice.

To quote myself on the topic of the 2d20 mechanic for advantage and disadvantage:
146406243 wrote:
It works well at its intended goal, to significantly increase the chance of someone who only has a moderate bonus while only slightly helping those with a very high or very low bonus.

The collateral is that it's (frequently) too powerful and doesn't stack with itself, and that it greatly magnifies any other bonuses that are present.  Because you only need one source of advantage, people will find whatever way is most convenient (probably flanking again) and then ignore anything else that should help them but would just provide redundant advantage - instead, they'll flock toward any small thing that gives a flat bonus, and the combined effects of advantage and a few flat bonuses will reduce the chance of failure to almost nothing.

Disadvantage is a similar issue.  A lot of great tactics (blinding, tripping, slowing, etc) become meaningless once you've already applied your one source of disadvantage.

The biggest problem right now is that the idea is already out there, and people like it.  I honestly can't fault anyone for enjoying the advantage mechanic, because rolling two dice really is more fun - especially when you can see that you would have rolled a 1, but you get to drop that die and only take the 17.  I'm not entirely convinced that people enjoy disadvantage, since dropping a 20 and taking a 4 just feels terrible.  I do get the sense that players would somehow end up benefiting from advantage far more often than they would end up suffering from disadvantage... but that's another topic entirely.

Although I expect that there will be an easily-abused condition for gaining advantage, the alternative is worse.  If the only way to grant advantage is for one character to spend an action, then we risk giving rise to the advantagebot - as boring as a healbot, but considered just as necessary for any serious group.  And just as with healbots, I'm sure there will be some players who enjoy being an advantagebot (myself included), but it should never feel like a mandatory part of the game.

Since I feel that players will be operating with advantage most of the time anyway, it would be less dis-honest to just build that into the way the game is intended to be played.  Don't hide the mechanic in the situational combat modifiers in order to make experienced players feel smart when they can manage an 80% up-time on it.  Make it new-player friendly, so people can get a sense of when they're doing well as they play.  I suggest baking it directly into classes, as an irrevocable part of what it means to belong to a specific class.

I propose that each class has specific circumstances under which it gains the 2d20 (advantage) mechanic, which should highlight the expertise of that specific class.  I will call it mastery so as to distinguish it from combat advantage or the specific current iteration of the D&DN rules.  (It's just a word - the exact term isn't important right now.)

Mastery would come into play when you act accordingly with your class features.  Fighters would gain mastery to attack rolls whenever they use the kind of weapon they have chosen for weapon specialization.  Rogues would gain mastery to attack rolls in situations that would traditionally allow them to sneak attack.  Wizards would specialize in a particular school, and gain mastery whenever casting a spell or ritual from that school.  Clerics would gain mastery when they cast a domain spell (or something).  Everyone gets mastery when they use the local equivalent of class skills.  (If the rogues seem to get the short end of the stick, well, they do benefit a lot more from the skills portion of mastery, and I've never seen a rogue denied a chance to flank consistently over multiple encounters.)  

There are some assumptions here based on rules that I just don't have access to yet, but the core of the idea should be clear: you are intended to roll 2d20 when you do the things at which you excel.  You are not supposed to miss when you are in control of the situation.  It is not fun to miss when you're doing everything right.  Everyone has a basic chance to do anything, based on ability scores alone, but special training gives you the 2d20 mechanic which is the only benefit of that training.  You could and probably should back up your mastery with good stats, but if you choose to have a fighter with moderate strength (for example), the mastery will actually help you more than if you have high strength - it's a kind of diminishing returns on effectiveness.  Meanwhile, the rogue can spot traps much better due to his training, even if he isn't normally the most perceptive guy on the team.

As far as enemies go, the only ones with mastery would be the ones who were always intended to be awesome.  In 3E, these were the monsters who had levels in barbarian or ranger instead of warrior.  In 4E, these were the elite and solo monsters.  Not every two-bit goblin will be rolling mastery, but the goblin warchief probably will.  (Tied to what? that can go in the stat block.  These guys are intended to be special, so they could probably stand to have a longer-than-average entry describing how they work).  When the players see the DM reach for that second d20, they should know that things are about to get serious.

That's really all there is to it.  I can easily see this expanding to 3d20 for epic levels, and a multi-class mechanic that lets you pick up the mastery of another class.  It's much easier to expand upon than the advantage and disadvantage mechanics, and I honestly think it would be more fun - especially for the critical new-player demographic, who would otherwise risk being overshadowed to an uncomfortable degree by the more experienced players.  

Just to get a jump on one of the perceived flaws with this system, though, I don't expect that flat modifiers will break bounded accuracy.  As far as I understand it, the idea of bounded accuracy is only that +hit and +AC won't increase with levels.  For the purposes of hitting an opponent, then, it will be like perpetually being first level in 3E or 4E.  I wish I had time to go into further analysis of those situations, but those games never really suffered from flat modifiers at first level; a +2 here or there was useful, and only sometimes felt necessary for hitting a very tough foe.  Under my proposed system, anyone fighting with mastery shouldn't need the flat bonus (it shouldn't feel mandatory, at least), but those without it might be able to take advantage of circumstances in order to contribute where they otherwise could not.

The metagame is not the game.

Very well put together there. I agree with you on most of it. I'm afraid as it stands, that advantage and disadvantage are going to be thrown around all over the place. That's not so bad, except that they represent much larger swings in effectiveness than most people seem to realize.

I find myself thinking that if your version (mastery) is class based, then it should probably be focused on combat-actions strictly. I say this only because if it's extended to specific tasks beyond that (yes, even detecting traps) then it effectively means that rogues will be the ONLY trap finders; Even if other classes have the appropriate skills, they won't compare.

How about for non-combat functions, apply it to your specific Background related tasks. Thus, anybody with a Thief-like background might benefit with traps and lock picking and such. Rogues aren't the only kind of thieves there are in the world, just like Fighters aren't he only soldiers. That's not to say that rogues can't get that one as a default regardless of background, but I kinda think anybody who wants to devote a part of their character to it aught to be able to do so competitively.
Very well put together there. I agree with you on most of it. I'm afraid as it stands, that advantage and disadvantage are going to be thrown around all over the place. That's not so bad, except that they represent much larger swings in effectiveness than most people seem to realize.

I find myself thinking that if your version (mastery) is class based, then it should probably be focused on combat-actions strictly. I say this only because if it's extended to specific tasks beyond that (yes, even detecting traps) then it effectively means that rogues will be the ONLY trap finders; Even if other classes have the appropriate skills, they won't compare.

How about for non-combat functions, apply it to your specific Background related tasks. Thus, anybody with a Thief-like background might benefit with traps and lock picking and such. Rogues aren't the only kind of thieves there are in the world, just like Fighters aren't he only soldiers. That's not to say that rogues can't get that one as a default regardless of background, but I kinda think anybody who wants to devote a part of their character to it aught to be able to do so competitively.

Look through the playtest and tell me how many times you need a 10-13 on the dice to hit.
I really like this idea, and I think it fits really well with the design goal that "class X should be really good at Y".

This recently article
www.wizards.com/DND/Article.aspx?x=dnd/4...
says they want to find a better way of handling skill training instead of just a bonus.  I don't like the proposed idea of having skill training replace your ability modifier because someone with skill and talent should benefit from both.  I think it would make a lot of sense if skill training gave you "advantage" with a skill instead of a bonus.

Similarly, the Rogue "Skill Mastery" could be changed from a minimum of 10 on skill rolls to rolling three d20s for Rogue skills.
I have to admit that I like this idea on first glance. It better regulates the awarding of advantage (mastery). For combat actions, the key is to make sure that whatever class feature the mastery is tied to, it cannot be chosen every round, otherwise it becomes stale fast. It should apply when a charcter is doing something exemplary for thier class, and not their basic attack/action.

As far as non-combat, I think this makes a good way to implement the skill training in a background. This makes sure you are good at what you master without needing a high ability score or requiring a separate calculation for background skills (the proposed skill bonus instead of ability bonus for the non-wismatic rogue to spot traps).

As mentioned, this reduces the drive to pump up one or two key stats to be effective, which I also like.

Mother may I isn't fun. Advantage is mother may I
...whatever
Mother may I isn't fun. Advantage is mother may I



except that currently the only way to assuradly get advantage is completely not mother-may-I in any way, and is in fact the inverse where you tell them exactly what is happening to them.
Mother may I isn't fun. Advantage is mother may I



except that currently the only way to assuradly get advantage is completely not mother-may-I in any way, and is in fact the inverse where you tell them exactly what is happening to them.


You pretty much prove the point. There's only one way to assuredly get it, and anything else is in the realm of Mother may I.
...whatever
I like this idea.  It would have to be implemented very carefully, but divorcing Advantage/Disadvantage from universal circumstantial instances and instead using it to represent/encourage actions iconic to a particular class/build is quite appealing to me, at least in principle.

You pretty much prove the point. There's only one way to assuredly get it, and anything else is in the realm of Mother may I.

I don't know why you'd assume that the playtest represents the final word in how the mechanic will work, and that it won't be refined or expanded to work in other ways.  I expect it could change pretty dramatically before release, or even before the next playtest.

"I want 'punch magic in the face' to be a maneuver." -- wrecan

Mother may I isn't fun. Advantage is mother may I



except that currently the only way to assuradly get advantage is completely not mother-may-I in any way, and is in fact the inverse where you tell them exactly what is happening to them.


You pretty much prove the point. There's only one way to assuredly get it, and anything else is in the realm of Mother may I.




That assumes they won't invent other ways of spending actions in order to gain advantage.  Especially given the examples in the most current blog post I could definately see some sort of attack that has a decent negetive to the attack roll to gain advantage on your next attack.  People will number crunch them to determine if it is worth it and that is fine tuning, but I see no indication that there won't be other ways to assuradly gain advantage .  Although I am sure that it will follow the spend an action to get advantage on your next attack model.  Precisely so that people can't sit there with advantage all the time.
I like your idea of using two dice rolls in a different way. I think that your everyday advantages/disadvantages (A/D), eg. higher ground, blind, slippery ground, etc, should be represented, as a general rule, by flat bonus or penalties. A/D that allow a double roll should be very rare, maybe, as you said, attached to class abilities.
I do thing that A/D look very significant. I'm also under the impression that the present method is not very good to reflect different kinds of A/D, for example, I think attacking some one that is paralysed is easier than attacking someone that is prone, and the rules should reflect that.
Thinking about it now, It might be better suited to go along with Theme. The role of any class can vary a lot in combat based on build and focus, so there's no ONE circumstance that would fit every type of Fighter, or every Wizard, or every Rogue. Linking it to Theme however, would give you a way to apply it to your chosen role more so than class would.

Defender would apply Advantage to certain defensive combat maneuvers and/or abilities, where as Slayers would be offensive. More to come as we see more themes.

Just so long as it's established with the creation of the character, I think it'd work better that way.
I'm not a fan of this proposal.  You properly identify that advantage is a problem if it is too ubiquitous and your example is what would happen if you got advantage with flanking (which at the present, you don't). 

But in your proposal, fighters would get mastery all the time because they will always be weilding their specialized weapon.  Rogues would get advantage when flanking, which you specifically identified as a problem.  Wizads would almost always cast spells int heir specialized school.  So I think characters are more likely to have 2d20 in your system than the current one.

I think the fear theat advantage could become too ubiquitous is justified.  Developers would have to be very careful to limit how you get advantage.  But tying it too class would, in my opinion, accentuate that very problem.
I posted here "CLICK" on the effect of adv/dis. Basically, it makes it much easier to hit things that are easy to hit, and  marginally easier to hit things that are hard to hit.

So tieing it to a class feature that is "expected" to hit is probably a good thing. In 4E, firing off that daily only to have it miss against mooks really sucked. Even if it had miss on half, against minions you wanted to cry. The diminishing returns against a hard to hit creature make it so its not a "I win" effect.

So IF tied to class, I would like for a liited set of abilities: Dailies or some other limited situation. If this were a 1E dragonlance game, Black robe wizards would gain advantage when Nuitari was waxing eg.
I'm not a fan of this proposal.  You properly identify that advantage is a problem if it is too ubiquitous and your example is what would happen if you got advantage with flanking (which at the present, you don't).

The real problem I was trying to address was that, because it was attached to so many different things, it created a nightmare by causing most situational modifiers to not stack.  It made itself very valuable and desired, and then stopped adding any benefit at all - you need exactly one source of advantage, or else you're suffering from it or it's being wasted.  (Granted, there aren't a ton of circumstances described yet, but we've barely seen a fraction of the system so far, and every indication has been that they would be relying heavily on mechanics that grant advantage or cause disadvantage - because they're fun.)

You can't just get rid of the 2d20 (advantage) mechanic, so I thought of where else it could be placed.  I have experience running a system (of my own design) which is similar to what I proposed here, and though it suffered from many flaws, this aspect of that system met with resounding success.  It is new, and different from anything that has come before in D&D, so I thought about making it more familiar by tying it into the iconic class features from ancient D&D: weapon specialization, backstab, school specialization, etc.

This might mean an uncomfortable up-time on the 2d20 mechanic - you would get it most of the time under this system - but the more I thought about it, the more I became okay with it.  Missing was never fun anyway.  This is a way of highlighting the superior skill of player characters without making them un-hittable or letting them one-shot anything.  To be honest, not having advantage feels a lot worse than it actually is - a rogue who couldn't get into position might drop from 90% accuracy down to 60% accuracy, but that less-skilled goblin is never going to benefit from mastery so this just puts you back on even footing (similar to how things have always been, in 3E or 4E).  Because it's a system that usually only works for the players, it's rewarding you for playing well instead of punishing you for playing badly.

If it was tied to themes, or some other character-creation detail, then that works too.  The main goal is to take advantage out of the realm of situational combat modifiers without removing the 2d20 (advantage) mechanic entirely.

The metagame is not the game.

I like the advantage.  It helps mitigate the insane randomness of the d20.  There's nothing I hate more as a DM than a player who misses 5 times in a row because of bad luck.  It is immensely frustrating and invokes a feeling of powerlessness in everyone.  It's no fun for players either, of course.

In this sense, it is a lot better than a +2 or even a +5 bonus, since they still have that ridiculously random element of one d20.

I'd rather have slight advantages of roll 2d20, take the highest and subtract 2 than lessen the use of the advantage in general.  That's still a net gain around where players should be rolling for success (similarly, take the lowest and add 2 is a net loss on average).

I'd accept something else that helps mitigate randomness in d20s in a less profound way than the Advantage.  Heck, I might just use a 3d6 system in my games, though the addition there can be annoying, I suppose.  With a very flat progression curve, that should work pretty well.  There are other ways to handle it too, I suppose.  Roll 2d20, with a main die selected, if the main die is 5 or less, then use the other die if higher...

Hmm.  Let's see.  if you need an 10 to hit.  Then there's a 55% you hit on the main die, a 20% chance you miss and don't use the other, a 25% chance you miss and use the other.  Then there's a 55% chance the other hits.  So that's 25% of 55% chance that you hit when you roll less than 5.  Or 13.75% chance to hit above your base 55% chance.  That's a bit short of a +3 bonus.

Though, that's also equivalent to reroll if you rolled a 5 or less.  If it is reroll if you rolled 4 or less, then that's an 11% boost or just over a +2.

I'd rather have a reroll mechanic of some sort.  That would work for a minor advantage, yes?  As a bonus, it can stack if you want it to.

Anyhow, I'd rather this or something like a Rogue's Skill Mastery (minimum die result) than just a static bonus.  Also, it is better that way, since they are trying to keep static bonuses to a minimum.  Start having some of those and you might get awkwardness from stacking.
A d20 is not insanely random. It's a uniform distribution for goodness sake. If you're missing 5 time in a row it's because the difficulties are set too high. This an issue with the system that's completely independent of the die that's being rolled.

Adding more dice or "roll X/pick Y" mechanics only obscures the probability of what's happening even more, and actually makes the static modifier even more important and the higher AC's and DC's even more punishing. Trust me, I used to make this mistake in some of my earlier designs. I fall back on the uniform distribution out of necessity, not preference.
Mother may I isn't fun. Advantage is mother may I


I sincerely doubt we've seen all the nuances of advantage before it's even out & finished
A few guidelines for using the internet: 1. Mentally add "In my opinion" to the end of basically anything someone else says. Of course it's their opinion, they don't need to let you know. You're pretty smart. 2. Assume everyone means everything in the best manner they could mean it. Save yourself some stress and give people the benefit of the doubt. We'll all be happier if we type less emoticons. 3. Don't try to read people's minds. Sometimes people mean exactly what they say. You probably don't know them any better than they know themselves. 4. Let grammar slide. If you understood what they meant, you're good. It's better for your health. 5. Breath. It's just a dumb game.
I was actually hoping they would get rid of static bonuses with skill training. I was thinking more along the lines as someone skilled in climb would be able to climb at full movement speed, but someone not trained can only move half. I would think advantage would play into it as well instead of just a flat +3 bonus.
A d20 is not insanely random. It's a uniform distribution for goodness sake. If you're missing 5 time in a row it's because the difficulties are set too high. This an issue with the system that's completely independent of the die that's being rolled.



Not at all.  If you need an 8 or better to hit, that's a 35% chance of a miss, or .52% chance you miss 5 times in a row.  Seems small, right?  With 5 players, that's about an 2.6% chance it happens to one or more of them in a single 5-round period.  If combats are 5 rounds and you have 3 a night, then that's a 7.6% chance it happens any given night.  If combats are 6 rounds, this is a 12.2% chance it happens in a given night one or more times.

Not common, certainly, but it is a common enough to be a problem.  Of course, if you only consider 3 or 4 rounds of absolute uselessness, then the probabilities go up considerably.  To nearly a 50% chance that you'll have a character in a 5 person party be useless in any 3-round period through no fault of their own.

Adding more dice or "roll X/pick Y" mechanics only obscures the probability of what's happening even more, and actually makes the static modifier even more important and the higher AC's and DC's even more punishing. Trust me, I used to make this mistake in some of my earlier designs. I fall back on the uniform distribution out of necessity, not preference.



No, it makes it more of a bell-curve, which is far, far more realistic for modelling most things, and it plays better too if you control static modifiers.  D&D hasn't traditionally had much control over static mods, which is why this hasn't worked for D&D in the past.  In Next, however, they are taking this very, very seriously, so less random resolution mechanics will work a lot better.

So while I accept your point is correct with older versions of D&D, I don't see how it applies to Next.
Drachasor: I'm already well aware of how binomials work. My point is the systems you're suggesting are objectively worse at this and just obscures that with complexity. My degree is in statistics. My focus is in games. Please listen to me.

Let's say you take 3d6 as your core mechanic and try to map it to difficulty levels. At DC 11 you'll succeed 50% of the time. Neither of us like the 50/50 split so let's drop the DC by one to 10: Now you have 62.5% chance of success . . . just from a measily +1! Still, I know you want something a little more reliable . . . Here's the basic relevant ranges for you:

DC: Success Rate
13: 29.5%
12: 37.5%
11: 50.0%
10: 62.5%
  9: 74.1%
  8: 83.8%
  7: 90.7%
  6: 95.4% 

As you can see, with the mechanic you propose a single difference in the target number makes a huge difference in probability unless you're setting the success rates to close to 95%.  What exactly was the benefit of this?
Basically, my point is that you get more reliable results from focusing around 75% on a uniform distribution since then you have the reliabiilty that players desire without practically eliminating the chance of failure. When you're dealing with a d20 on a DC 6, a +2 or -2 modifier to the roll only makes a 10% difference in probability and a 13.3% change in general effectiveness. Compare this to the -24.1% or +11.4% that a -2 or +2 brings from a 3d6 system.  A static modifier suddently can cost you about 33% of your general effectiveness or grant you an additional 15%. Essentially, the tyranny of accuracy is actually more dramatic in a bell-curved distribution. This is a general mathematical truth.  Furthermore, such a system is far from intuitive and this can cause many unintended consequences from GM's and game writers tinkering with them later.
This is why I think that with bounded accuracy set to around 50-60% classes and themes should grant advantage to show mastery instead of handing out flat bonuses. This way the fighter can have a class feature that grants advantage with weapon attacks, the rogue can get advantage with certain skills instead of the count roll as 10 ability, etc. The classes that are true masters in their field will have between a 70-85% chance to succeed at tasks the rest of the party succeeds at only 40-60% of the time at. Then for tasks where the expert is succeeding at 50% of the time someone else in the group still can succeed at around 30% of the time. This also keeps numbers lower in general and helps to remove numbers bloat and auto success/auto failure situations that can come up in 3e and 4e.
As you can see, with the mechanic you propose a single difference in the target number makes a huge difference in probability unless you're setting the success rates to close to 95%.  What exactly was the benefit of this?



You're proposing a completely different rolling mechanic. Sure, you've got yourself the GURPS probability chart there, but that's rolling the SUM of 3 dice, not applying the highest of multiple dice. You may have a degree in statistics, but you're a little confused clearly... I'm not trying to be a jerk in saying so, but if you're rolling two dice and applying the highest, it doesn't make it any more difficult to hit a higher target number. It does if you're applying the sum of multiple dice, but that's not at work here. It's quite literally impossible with this mechanic for it to be MORE difficult to hit a higher target number than it would be without Advantage, even without applying mathematical theory: You have two chances to succeed at the same task. Now, please feel free to correct me if i'm wrong, because I certainly don't have a degree in statistics.

I'll really be impressed if you can throw me up some statistics for roll probabilities from Deadlands (not the Savage Worlds version)

Another poster put together a great post on it that's actually informative and (as far as I can tell) accurate.
community.wizards.com/dndnext/go/thread/...
Gurka, Drachasor was talking about replacing his d20 with 3d6. My statement remains the same.
Apologies. I did miss a couple posts.
Also, original Deadlands is such a mess that it is my favorite evidence that 1990's game designers had no clue about statistics and systems design.  The very fact that several exploding d4's were generally better than d6's but cost less XP?

I should mention that I do have a soft spot for it though. They experimented with a lot of interesting mechanics, but they had almost no idea what effect those mechanics would have on the game.

Drachasor: I'm already well aware of how binomials work. My point is the systems you're suggesting are objectively worse at this and just obscures that with complexity. My degree is in statistics. My focus is in games. Please listen to me.

Let's say you take 3d6 as your core mechanic and try to map it to difficulty levels. At DC 11 you'll succeed 50% of the time. Neither of us like the 50/50 split so let's drop the DC by one to 10: Now you have 62.5% chance of success . . . just from a measily +1! Still, I know you want something a little more reliable . . . Here's the basic relevant ranges for you:

DC: Success Rate
13: 29.5%
12: 37.5%
11: 50.0%
10: 62.5%
  9: 74.1%
  8: 83.8%
  7: 90.7%
  6: 95.4% 

As you can see, with the mechanic you propose a single difference in the target number makes a huge difference in probability unless you're setting the success rates to close to 95%.  What exactly was the benefit of this?



Setting the difficulty near the center of the distribution is a bit crazy, imho.  You'd want an average die roll of 9 or better to do the job (75% success rates on actions generally feel about right, in my experience).  A +/-1 is just about 10% there, which isn't that bad.

The benefit is average results are more common, which can matter in skill checks when an especially bad check can produce worse effects as Next has.  It also means that if you outclass something, then you'll perform more reliably against it.  It's a lot less given to streaks of low numbers.  As you yourself indicate, the chance of getting a 6 or less is just about 10%, whereas in the d20 system it's 30%.  There is definitely something to be said for getting more consistently average results.

And you can't argue that this would work much, much better in Next than it would in 3.X which had numbers all over the place.  In a world were people are typically going to be within 1 of each other (5 or 6 to hit corresponding to a 16 or 18 ability score with a +2 bonus).  This, again, isn't that bad and you get the benefit so that if the hit is relatively easy, then you're not going to still suffer a massive miss chance of 30% or so.

Of course, I was specifically advocating a reroll on a low number -- 4 or less -- (the 3d6 was an idle comment).  This produces less significant jumps.

DC: Success Rate
13: 48% (from 40% on a straight d20)
12: 54% (from 45%)
11: 60% (from 50%)
10: 66% (from 55%)
  9: 72% (from 60%)
  8: 78% (from 65%)
  7: 84% (from 70%)
  6: 90% (from 75%)

Pretty nice for a moderate advantage.  It also has the benefit of making repeatedly failing for 3+ rounds far, far less likely since half the bad rolls or so give you another chance to hit with the second die.

Gurka, Drachasor was talking about replacing his d20 with 3d6. My statement remains the same.



Yeah, if you pick ONE sentence out of my posts and ignore everything else.  Yeah, I was really advocating that.  Seems more like you went after a convenient straw man rather than the main thrust of what I was saying.
Thanks for the responses, everyone.  Just to get things back on topic, I'm not making a case for whether the 2d20 (advantage) mechanic is inherently good or bad statistically.  That has been discussed previously, in other threads, and (as far as I can tell) removing it entirely would create significant backlash from the majority of play-testers who enjoyed it.

I'm mostly looking for feedback on whether this new alternative would be a better, more intuitive, more enjoyable, and/or less disruptive method of utilizing that mechanic.  (Of course, I would greatly appreciate discussion of how this shift would change probabilities in regards to characters performing tasks for which they are or are not optimized - I'm not nearly as good at math as I feel would be necessary to perform a fair analysis.)

The metagame is not the game.


Setting the difficulty near the center of the distribution is a bit crazy, imho.  You'd want an average die roll of 9 or better to do the job (75% success rates on actions generally feel about right, in my experience).  A +/-1 is just about 10% there, which isn't that bad.

The benefit is average results are more common


That's absolute nonsense. Your chance of rolling a 9 or higher with 3d6 is 74.1%. Your chances of rolling a 6 or higher with a d20 is 75%. We use the percent probability as objective comparisons.  If you use 3d6 then a +/-1 is over 10% and it varies significantly based on the base accuracy of user. If you instead used a d20 set for 6 then a +/-1 is only 5% and it's the same for everyone. Plus, if you want something closer to 10% you can just +/-2. You have more control over what's happening, more reliability from your system, and it's easier for everyone to understand.


Yeah, if you pick ONE sentence out of my posts and ignore everything else.  Yeah, I was really advocating that.  Seems more like you went after a convenient straw man rather than the main thrust of what I was saying.


Sorry if it seems like I'm picking on you. This is really just meant as an illustration for why bellcurves don't actually increase reliability of a system when you're trying to roll over "X." It's a common mistake that I've personally made so I share it as a cautionary tale. It's a lesson that sadlly carries over to almost all complex distributions too. I love complex dice mechanics in theory, but they generally cause more problems than they solve in reality.
Also, original Deadlands is such a mess that it is my favorite evidence that 1990's game designers had no clue about statistics and systems design.  The very fact that several exploding d4's were generally better than d6's but cost less XP?

I should mention that I do have a soft spot for it though. They experimented with a lot of interesting mechanics, but they had almost no idea what effect those mechanics would have on the game.




You're pretty much right, there. It's still one of my favorite systems, because it's just plain fun to play. It wasn't very well balanced, and generally took a lot of houserules to fix up, but it just had one of the most entertaining sets of mechanics of any game out there.

I still have a soft spot for die pools in serious or "realistic" settings however.

Even the absolute best in the world fails 5% of the time in D&D. This version should be a bit better than it used to be since DC's on EVERYTHING scaled up as you leveled in some earlier versions, it meant that getting "better" at something didn't actually mean you'd succeed more often, it just meant you could realistically attempt standard (for your level) tasks, and still have a more or less even chance of success.

Now atleast, getting better does mean that you'll succeed more often, but in my mind, if you want your character to have a shtick, focus, or talent for something, they should actually succeed reliably. That doesn't mean without fail, and it doesn't mean that very difficult tasks shouldn't still be risky. It means that tasks that they've performed hundreds or thousands of times to attain the skill in question shouldn't be something he worries about failing every few attempts.



Setting the difficulty near the center of the distribution is a bit crazy, imho.  You'd want an average die roll of 9 or better to do the job (75% success rates on actions generally feel about right, in my experience).  A +/-1 is just about 10% there, which isn't that bad.

The benefit is average results are more common


That's absolute nonsense. Your chance of rolling a 9 or higher with 3d6 is 74.1%. Your chances of rolling a 6 or higher with a d20 is 75%. We use the percent probability as objective comparisons.  If you use 3d6 then a +/-1 is over 10% and it varies significantly based on the base accuracy of user. If you instead used a d20 set for 6 then a +/-1 is only 5% and it's the same for everyone. Plus, if you want something closer to 10% you can just +/-2. You have more control over what's happening, more reliability from your system, and it's easier for everyone to understand.



That's certainly true about a d20 verses 3d6 regarding 6 and 9 respectively.  However, when you look at something that is then EASY compared to that, say 2 easier, things change a lot.  You still have a 15% chance of failure on something you are exceedingly good at, which is horribly unrealistic and this sort of thing has been immersion breaking many times in games I've been in or ran.  For the a system using a more bellcurve-like distribution, this chance drops off much, much faster.

I'll grant there are ways around this with a d20...the DM arbitrarily changing the bonus is a poor one.  Advantage or Skill Mastery are better...but I guess you're against any sort of reroll mechanic ever?  Or do you just agree with 90% of what I wrote and decided to make a big deal out of the 10%?

Again, though, I've spent the vast, vast majority of my posts NOT arguing for a 3d6 system, but rather a Minor Advantage that lets you reroll if you roll a 4 or less.  Pray tell me, what's wrong with that?  Or do you only bother to attack the 3d6?


Yeah, if you pick ONE sentence out of my posts and ignore everything else.  Yeah, I was really advocating that.  Seems more like you went after a convenient straw man rather than the main thrust of what I was saying.


Sorry if it seems like I'm picking on you. This is really just meant as an illustration for why bellcurves don't actually increase reliability of a system when you're trying to roll over "X." It's a common mistake that I've personally made so I share it as a cautionary tale. It's a lesson that sadlly carries over to almost all complex distributions too. I love complex dice mechanics in theory, but they generally cause more problems than they solve in reality.



It feels like your MISREPRSENTING my proposal, by focusing on a side 3d6 sentence rather than the PARAGRAPHS that I wrote.

You want I should tackle each distribution seperately and explicitely list the pmf's and cmf's to demonstrate how with each one you're getting less control and making differences in characters harder to manage? Because that's a lot of work that I'm not getting paid for, just to educate a few gamers. I'm not completely against reroll mechanics, but they are of variable benefit to each character and can be problematic when players optimize for them. As such designers need to be careful which scenarios trigger rerolls.
You want I should tackle each distribution seperately and explicitely list the pmf's and cmf's to demonstrate how with each one you're getting less control and making differences in characters harder to manage? Because that's a lot of work that I'm not getting paid for, just to educate a few gamers. I'm not completely against reroll mechanics, but they are of variable benefit to each character and can be problematic when players optimize for them. As such designers need to be careful which scenarios trigger rerolls.



Well, I guess I rather expected you'd tackle the one I was actually focused on talking about, rather than the one I just wrote a sentence on.  That or not bother, really.  Either would seem to be what would make sense.




Anyhow, for those that missed what I said in the rather off-topic exchange, I was proposing a Minor Advantage system, where you reroll the die if you get a 4 or less.  It's about the same as a +2 bonus most of the time, but avoids potential stacking issues which is one of the goals of Next.
Anyhow, for those that missed what I said in the rather off-topic exchange, I was proposing a Minor Advantage system, where you reroll the die if you get a 4 or less.  It's about the same as a +2 bonus most of the time, but avoids potential stacking issues which is one of the goals of Next.

That might be too small of a bonus to matter, so at least that should avert the whole advantage-bot catastrophe.  I can see how it would kind of make for a more reasonable distribution and wouldn't be overpowered, but it would still not allow cumulative modifiers to stack - a blind, prone, stunned target would be no easier to hit than one you were merely flanking.  

could imagine a system of scaling minor advantage (basically move the flat modifiers from old editions into the reroll threshold of the new system, so what would have been a +8 bonus in 3E would become a d20 re-roll if the first die comes up 8 or less), but that seems kind of complicated.

The metagame is not the game.

What you are suggesting is that the fighter who specializes with the longsword always rolls 2d20 to attack, as the wizard who specializes in evocation. ect. You are making Advantage ALOT easier to get than I percieve  will be available in the final 5E rules.

It makes me think about the 4E avenger who got "advantage" on all his attacks as long as he wasnt adjacent to 2 foes, and could single out one chosen guy. And even that was more limiting than what you are suggesting.

I would like to see the +2/-2 system brought back into play for the basic stuff like flanking and attacking while prone. And make this advantage/disadvantage thing Alot more limiting or only in special cercumstances.

IE Advantage casting Fireball vs a whie dragon.

Advantage attacking a Pralyzed or seemingly crippled target.

Disadvantage using a ranged weapon through galeforce winds

ect.

I just dont like advantage system at all. I hope to easily houserule and remove it if I ever get into 5E

I just dont like advantage system at all. I hope to easily houserule and remove it if I ever get into 5E

 This would be ideal, yes.  As of current, though, it seems too popular for them to change and too integrated to be easily removed.

The metagame is not the game.

I've been considering the Dis/Advantage mechanic (and the suggested alternative), and the more I think about it the more Dis/Advantage not stacking across various conditions just doesn't bother me.  Because being prone or blinded each do more (and different) things aside from just cause Dis/Advantage.  So yes, Dis/Advantage could easily become quite common if not properly policed in the rules, but flanking someone is still strictly not as good as blinding them. I am not going to flank an enemy and then not bother blinding them if I'm able.  For one thing, they can presumably get out of the flank, but will have much more trouble overcoming blindness, which causes a host of other problems for them aside from Dis/Advantage.

If the mechanic seems too powerful, I agree with what was suggested earlier for "Minor Advantage," akin to the "brutal" mechanic/property for 4e damage dice, but applied to an attack roll.  So Advantage might be, "You reroll a natural roll of 5 or lower," and Disadvantage might be, "You reroll a natural roll of 15 or higher."  In principle, at least; I haven't done the math on those numbers specificically.

"I want 'punch magic in the face' to be a maneuver." -- wrecan