D&D 5th Edition: Thread #7

[Kurald Galain]

Comments on Individual Spells

That was an insightful summary.

Your comment on Disguise Self shows the fundamental flaw in 5E's core mechanic, though. DC 15 is intended to be a hard check, since normally you'll have only a +0 to +3 from your attribute, so this is not something most characters can do reliably. But it's something that will frequently happen randomly: as you say, 30% of commoners will see through your disguise immediately. This is because a 0-3 spread on attribute modifiers pales in comparison to the 1-20 spread on the die (and 5E has no notion of "trained only" skills, and being trained makes only a minor difference anyway).

That's right, this is the game where ordinary commoners can randomly succeed on a Very Hard check one time out of ten. Very Hard is defined as something that "only especially talented individuals need even try", such as physically breaking out of manacles, or recalling esoteric information known only to a few. I repeat, one out of ten commoners can do this! Frankly this system is absurd.

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

[wadledo]

That was an insightful summary.

Your comment on Disguise Self shows the fundamental flaw in 5E's core mechanic, though. DC 15 is intended to be a hard check, since normally you'll have only a +0 to +3 from your attribute, so this is not something most characters can do reliably. But it's something that will frequently happen randomly: as you say, 30% of commoners will see through your disguise immediately. This is because a 0-3 spread on attribute modifiers pales in comparison to the 1-20 spread on the die (and 5E has no notion of "trained only" skills, and being trained makes only a minor difference anyway).

That's right, this is the game where ordinary commoners can randomly succeed on a Very Hard check one time out of ten. Very Hard is defined as something that "only especially talented individuals need even try", such as physically breaking out of manacles, or recalling esoteric information known only to a few. I repeat, one out of ten commoners can do this! Frankly this system is absurd.

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

Use the same rule that every game has (including D&D 5e).
Only roll for important things.
Otherwise you get such wonderful things as the merchant not being able to see you because he rolled a 1 on his spot check, or the guard is incapable of hearing you because he rolled a 1 on his listen check, or the experienced blacksmith falling into his forge because he rolled a 1 on his check to move across the cluttered floor of his workshop.

[Menteith]

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

Use the attribute instead of the attribute modifier, and increase the DCs of all check by 10 to compensate. 20 Strength adds a +20 to a check, instead of +5, and a relatively easy task is DC20 instead of DC10. This will allow very high attributes to scale faster than the difficulty (giving a sense of actual heroics at higher power levels - A Raging Barbarian 20 might actually be capable of Kool-Aid Manning through walls, for example) while increasing the range of bonuses in a meaningful way. It also helps keep odd values on attributes relevant.

EDIT

Wait, how did your post get above mine? Are you from the future?

[Tehnar]

That was an insightful summary.

Your comment on Disguise Self shows the fundamental flaw in 5E's core mechanic, though. DC 15 is intended to be a hard check, since normally you'll have only a +0 to +3 from your attribute, so this is not something most characters can do reliably. But it's something that will frequently happen randomly: as you say, 30% of commoners will see through your disguise immediately. This is because a 0-3 spread on attribute modifiers pales in comparison to the 1-20 spread on the die (and 5E has no notion of "trained only" skills, and being trained makes only a minor difference anyway).

That's right, this is the game where ordinary commoners can randomly succeed on a Very Hard check one time out of ten. Very Hard is defined as something that "only especially talented individuals need even try", such as physically breaking out of manacles, or recalling esoteric information known only to a few. I repeat, one out of ten commoners can do this! Frankly this system is absurd.

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

The quick fix (that still has problems of its own), is to increase all DC's by 10 (while further increasing DCs for hard and very hard checks), and to roll d20 + Ability SCORE instead of d20 + Ability modifier.

Use the same rule that every game has (including D&D 5e).
Only roll for important things.
Otherwise you get such wonderful things as the merchant not being able to see you because he rolled a 1 on his spot check, or the guard is incapable of hearing you because he rolled a 1 on his listen check, or the experienced blacksmith falling into his forge because he rolled a 1 on his check to move across the cluttered floor of his workshop.

That is not a simple solution, that is just handwaveing things away, leaving the resolution mechanic to the DM's whim.

Of course you should not roll for every thing that comes up, but rolling for guards to see through a disguise should be one of the things you roll for.

[Kurald Galain]

Use the attribute instead of the attribute modifier, and increase the DCs of all check by 10 to compensate.

That is a good and clear start, but a str-10 commoner would still have a 10% chance of breaking a Very Hard (DC 29) chain. It would probably work if you spread the DCs a bit more; if you think about it, "Hard" has no business being only 3 points (15%) more difficult than the next easier category.

Use the same rule that every game has (including D&D 5e).
Only roll for important things.

That doesn't work, because seeing through a disguise spell of someone trying to fool you, and breaking out of manacles that are holding you, are important things. And commoners (or low-perception low-strength 1st level characters) can do these things WAY more often than they should.

[Menteith]

That is a good and clear start, but a str-10 commoner would still have a 10% chance of breaking a Very Hard (DC 29) chain. It would probably work if you spread the DCs a bit more; if you think about it, "Hard" has no business being only 3 points (15%) more difficult than the next easier category.

Tentatively - use DC20 as a base and increase the DC by 5.5 (rounding down) for each catagory?;

DC20 for everyday things (under no pressure, a common person can accomplish the task by taking 10, and it might not even be possible for an exceptional individual to fail.)

DC25 for harder, but still possible, tasks (normal individuals will need to take 20 to reliably succeed, while exceptional individuals will be able to succeed a reasonable amount of time on their first try.)

DC 31 for very hard tasks (impossible for normal individuals, and even exceptional individuals will struggle with the task. Only the best of the best can accomplish these tasks quickly and reliably.)

And so on. I think it makes more sense to have a DC31 be present for Hard tasks (a normal person can't take 20 to succeed), but maybe it's just me.

[Ashdate]

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

I'd try and "fix it" by re-wording the spell to indicate that you don't get to make a "check" to break the illusion unless an NPC has a particularly good reason to be doing so.

If you're walking down the street as some random "off-the-top-of-your-head" human, then there is no reason for a commoner to look at you twice. If you're walking down the street as a particular non-famous person (perhaps the brother of the Captain of the Guard), you would only attract attention from people who would want to talk to that person, which likely means that commoners ignore you, but the actual Captain of the Guard can make a check. And if you stroll down the street as the King? Well, everyone's going to be paying attention to you, so you deserve the DM pulling out an online dice roller and making 100d20 checks to see if anyone rolls above a 15.

Granted, it doesn't solve the larger issues I have with the skill system, but there should be some common-sense language built into each spell/item that could create the possibility of multiple opposed checks.

[Craft (Cheese)]

Use the same rule that every game has (including D&D 5e).
Only roll for important things.
Otherwise you get such wonderful things as the merchant not being able to see you because he rolled a 1 on his spot check, or the guard is incapable of hearing you because he rolled a 1 on his listen check, or the experienced blacksmith falling into his forge because he rolled a 1 on his check to move across the cluttered floor of his workshop.

You cast the spell on your entire party to disguise yourselves in order to enter the count's estate. There are two guards at the door, and 5 PCs in the party. Not at all an unreasonable situation. Assuming both guards get to make separate checks for all of you as you pass by individually, and assuming they get no bonus, the chance that you all pass by the guards undetected is (7/10)^10 = 2.8%.

[Dienekes]

That's right, this is the game where ordinary commoners can randomly succeed on a Very Hard check one time out of ten. Very Hard is defined as something that "only especially talented individuals need even try", such as physically breaking out of manacles, or recalling esoteric information known only to a few. I repeat, one out of ten commoners can do this! Frankly this system is absurd.

Question to the Playground: can we come up with a simple fix for this problem? I'm sure it's not a problem to everybody; but if WOTC truly wants to maximize their market share, they should at least include a module or optional rule for those of us that are bothered by this.

Honestly if this bothers you the easiest is probably still just to convert to 3d6. Rolls of 19-20 are now impossible to the commoners, and it makes everything less swingy so especially talented individuals are the only ones who can even attempt the hard stuff.

Of course some say the swingyness is part of the fun of the system, but you can't please everyone.

Weird; in contrast to the general reaction around here, I was impressed by this latest playtest packet overall. Sure, there are still dozens of little problems, but this is the first round of rules I've seen that gives me hope that Next could turn out to be a better overall game than 3e Core.

It's what you focused on I think. Seems you went to magic, I went for martial.

And frankly, martial needs work. I still can't believe anyone thought 1d10 to 3 opponents at level 10 was in any way a good deal to make. And that blows all your Exp Dice

[Stubbazubba]

Use the same rule that every game has (including D&D 5e).
Only roll for important things.
Otherwise you get such wonderful things as the merchant not being able to see you because he rolled a 1 on his spot check, or the guard is incapable of hearing you because he rolled a 1 on his listen check, or the experienced blacksmith falling into his forge because he rolled a 1 on his check to move across the cluttered floor of his workshop.

You don't auto-fail a skill check on a nat 1, though, that's only on attacks and saves.

[Menteith]

And presumably everyone's taking 10 on stuff they can take 10 on. Under zero pressure, an experienced blacksmith isn't going to fall into their forge. If they're being chased by a ravening ghoul, well, then it's possible, and a really low roll might result in that, even without autofailing on a 1.

[Clawhound]

I think that wadledo had it wright. Roll for what's important.

That "what's important" criteria is important because many rules are written with that presumption in mind. The rules aren't a mechanical simulation of the world's mechanics. Instead, they are a mechanical means of inserting uncertainty into the narrative. Failure must be a possibility if you are to have tension.

In 3e terms, losing out to a 1st level commoner is an insult. In Next terms, losing out to a 1st level commoner is exactly what the designers want. It may not happen often, but it is always possible.

The designers also aren't designing for consistency. In this edition, they want variation at the tables. They want each group implementing the game in a way that works for the group. They exactly want the DM making decisions on the fly.

[Kurald Galain]

In 3e terms, losing out to a 1st level commoner is an insult. In Next terms, losing out to a 1st level commoner is exactly what the designers want. It may not happen often, but it is always possible.

The point is that by the current rules, it will happen often.

[AgentPaper]

For disguise self, I'd only roll if the person knew the person you were disguising yourself as, and only if they're actually interacting with you, or studying you closely. If you walk by a commoner on the street as someone they know, no check is made. Similarly, walk by a guard as someone they don't know, again no check. Walk by a guard as a noble they've seen a few times, and they get to make a check.

I'd also rule that if you're disguised as someone they know very well, IE a close friend or family member, then they can make the check even if they just see you in passing, and if they are interacting with you then they get advantage on the check.

With this, the rules work out much more logically, but I agree that this kind of thing should be written explicitly in the rules.

[Tehnar]

For disguise self, I'd only roll if the person knew the person you were disguising yourself as, and only if they're actually interacting with you, or studying you closely. If you walk by a commoner on the street as someone they know, no check is made. Similarly, walk by a guard as someone they don't know, again no check. Walk by a guard as a noble they've seen a few times, and they get to make a check.

I'd also rule that if you're disguised as someone they know very well, IE a close friend or family member, then they can make the check even if they just see you in passing, and if they are interacting with you then they get advantage on the check.

With this, the rules work out much more logically, but I agree that this kind of thing should be written explicitly in the rules.

Its not a question of criteria, when you should roll, or when you shouldn't.

The point is that even against untrained commoners (when you decide they need to make the check) the disguise spell does nothing. And that is far away from the only example.

[AgentPaper]

Its not a question of criteria, when you should roll, or when you shouldn't.

The point is that even against untrained commoners (when you decide they need to make the check) the disguise spell does nothing. And that is far away from the only example.

Untrained commoners have a fairly low chance of detecting you. It's only when they get to roll a lot, IE when every commoner you run into gets to make a check for every person that'd disguised, that the issues start to appear.

If you're only getting a check rolled on you 3-4 times over the course of your disguise, you've got pretty good chances to get by.

[Menteith]

Untrained commoners have a fairly low chance of detecting you. It's only when they get to roll a lot, IE when every commoner you run into gets to make a check for every person that'd disguised, that the issues start to appear.

If you're only getting a check rolled on you 3-4 times over the course of your disguise, you've got pretty good chances to get by.

As a counterpoint,

You cast the spell on your entire party to disguise yourselves in order to enter the count's estate. There are two guards at the door, and 5 PCs in the party. Not at all an unreasonable situation. Assuming both guards get to make separate checks for all of you as you pass by individually, and assuming they get no bonus, the chance that you all pass by the guards undetected is (7/10)^10 = 2.8%.

[AgentPaper]

As a counterpoint,

With my rules for how the checks are made, this would mean that the 5 PCs are all disguised as 5 different notable nobles, who the guards have seen before fairly often (IE: They live there), AND the guards are alert and actively looking at every single person coming in. In that case, yes, you have a pretty damned low chance of getting by, as you probably should.

On the other hand, if you plan better, you'll instead disguise one member up as a noble, and then disguise the other 4 party members as retainers/servant/etc to that noble. Now, each of the guards gets to do a single check, giving you a 56% chance of getting by.

Of course, this is still assuming that the guards are alert and actively searching. If they aren't, then they don't get to make a check at all, or at best, they get one check at disadvantage. Or you simply dress everyone up as washmaids and no checks are made in the first place, unless you run into a bunch of other washmaids, or mistakenly dress up someone as a washmaid who happens' to be one of the guard's wife.

Anyways, I agree that the spell is broken, I just don't think changing the DC will fix the issue. Even if you make it so that the guards have to roll a 20, you'd still only have a 60% chance of getting through. Pass through a crowd of 10, and you've got a 7% chance of not being spotted at all.

The only way to change that would be to make the roll impossible altogether for commoners to make, but I don't think anyone thinks that's a good idea. Instead, it'd be much better for the check to be made only in special circumstances. That way, you can retain the tension of not being sure that the disguise will work when you do something crazy, but if you're just walking around normally, nobody will give you a second glance.

[Tehnar]

Even if each guard makes one check for the entire group, you only have 49% chance of success.

The very basic mechanic of 5E is very bad, thats why this causes so much problems. 1d20+X, where X <= 5, against DCs that start with 10, is a bad resolution mechanic since its no better then throwing a coin to determine the outcome of a action.

[wadledo]

You don't auto-fail a skill check on a nat 1, though, that's only on attacks and saves.

And if your modifer is +4 or less?
Then you fail on a one, auto doesn't come into it.

[Kurald Galain]

Like I said earlier: not everybody considers this a problem, but clearly numerous people do. So it would be a good idea if WOTC wrote a module or optional rule to draw in those people. And by "good idea" I mean "more marketshare".

[Menteith]

Even if each guard makes one check for the entire group, you only have 49% chance of success.

The very basic mechanic of 5E is very bad, thats why this causes so much problems. 1d20+X, where X <= 5, against DCs that start with 10, is a bad resolution mechanic since its no better then throwing a coin to determine the outcome of a action.

To be honest, I feel like most of my issues can be put under this box. Sure, I wish that (Dis)advantage had functional stacking, or that the mundane classes stacked up a bit better, or that the magic item loot rules weren't pants-on-head dumb, but at its core, my issue with D&D Next is that a 1d20 resolution mechanic doesn't work for me when the static bonuses are small. It's also a part of why I don't enjoy playing low level 3.5/Pathfinder as much as mid levels. If they offer a solid 3d6 module, or alter the system to use Attribute+1d20 instead of Attribute Mod+1d20, or pursue another option that doesn't let a palsied cripple knock down a door a Raging Half-Orc Barbarian can't 10% of the time, then I might enjoy the system more.

Randomness is fine, and is integral to many systems. Over the top randomness, which clashes both with what I find fun and what would make sense in the world, is not fine.

[obryn]

The DCs are one issue. The number of rolls is another.

I don't know what can be done about the DCs in a bounded accuracy system. They're screwed up, and without inventing extra subsystems (really, 1d20+Stat might as well be 1d10+stat or 1d6+stat; it's a non-unified resolution mechanic either way) I don't know if it's possible to ... screw them down?

The fail chance based on number of die rolls is easier to reconcile.

Ideally, something like "sneaking past guards while disguised" should either be a simple, unopposed skill check ... or part of an extended skill check like a skill challenge. Use something like 4e's passive perception as a DC; rolling for the guards just increases the chaos in the system

-O.

[Kurald Galain]

I don't know what can be done about the DCs in a bounded accuracy system.

That's actually very easy. You increase the range on the static modifiers, so that they are no longer eclipsed by the 20-point spread on the die. Then you do some easy math and write down new numbers for easy/moderate/hard DC based on statistics rather than gut feeling.

rolling for the guards just increases the chaos in the system

It doesn't, actually. Having multiple rolls is less random than having a simple roll. That's because if you roll 1d20 you have no idea what the result is going to be, but if you roll 3d20 then you have very good odds of getting a result in the 30-33 range.

[wadledo]

It doesn't, actually. Having multiple rolls is less random than having a simple roll. That's because if you roll 1d20 you have no idea what the result is going to be, but if you roll 3d20 then you have very good odds of getting a result in the 30-33 range.

Isn't that contradictory thinking?
If you roll 3d20 and can expect to get a result somewhere in the 30-33 range, then a single d20 can be expected to get you a result between 10-11, on average.

[Kurald Galain]

Isn't that contradictory thinking?
If you roll 3d20 and can expect to get a result somewhere in the 30-33 range, then a single d20 can be expected to get you a result between 10-11, on average.

That's incorrect. It's not about averages but about standard deviation. Lower SD means less randomness, and more dice means lower SD.

[wadledo]

That's incorrect. It's not about averages but about standard deviation. Lower SD means less randomness, and more dice means lower SD.

Yes, but in this case you are not combining the total rolls, you are taking each individual roll and determining the result from there.

So 3d20 has the same odds of each individual dice (the only thing we really care about in this scenario, I would think) rolling any particular number as in a 1d20.

[obryn]

That's actually very easy. You increase the range on the static modifiers, so that they are no longer eclipsed by the 20-point spread on the die. Then you do some easy math and write down new numbers for easy/moderate/hard DC based on statistics rather than gut feeling.

It's the bounded accuracy bit that's the issue. In order to do this, you also need to have a wider range of bonuses and DCs.

It doesn't, actually. Having multiple rolls is less random than having a simple roll. That's because if you roll 1d20 you have no idea what the result is going to be, but if you roll 3d20 then you have very good odds of getting a result in the 30-33 range.

If both rolls are on the same side of the table, this is true; in that case, you have a curve, whether you're adding your numbers together or taking the better of two. If the rolls are on opposite sides, it's just increasing variability; there's no curve involved. Unusual results become more likely, not less.

-O

[Kurald Galain]

Yes, but in this case you are not combining the total rolls, you are taking each individual roll and determining the result from there.

I think the suggestion was to roll a player's skill against a guard's passive value, instead of rolling player's skill against a guard's skill check. The former is actually more random.

[obryn]

I think the suggestion was to roll a player's skill against a guard's passive value, instead of rolling player's skill against a guard's skill check. The former is actually more random.

No, it's not. Here's the basic illustration.

Let's say I have a +7 Sneak. The Guard has a +4 perception.

If we use a Passive value, and I need to meet or exceed the guard's skill check, I need a 7 or better to succeed. I succeed 70% of the time.

Let's say we're making opposed rolls. I can get from an 8 to a 27 on my d20 roll. The guard can get from 5 to 24 on his. There's no curve involved for either of our rolls - they're straight d20 - so each of these values has an even frequency for each of us.

Look at it in terms of my DC. My DC is no longer set; it's floating. It averages to a 14.5, but that's not an average with any real central tendency to it, since it's based on a single flat-curve d20 roll. It's even chances my DC will be 5, as it is it will be 24. On average, my success rate is still 70% or so; it hasn't changed. But I have a greater degree of uncertainty in my results. The variability has increased.

Move to the guard rolling 3d6 instead. Now he has a central tendency to hit right around 14.5. The variability has decreased from a flat d20 roll.

Now the guard is taking the average result of 200 1d20 rolls. It's very, very likely my DC will be around 14.5 because the curve has again increased its central tendency, and it will continue to do so as the curve narrows.

The final state is 100% central tendency ... or setting the DC to 14.5. At that point, we're back at Passive DCs, and the overall variability in the system is vastly reduced from opposed d20's.

-O