D&D 5E D&D Next Blog - Wizards Like to Roll Dice Too

[ArcaneSpringboard]

This is incorrect. The standard deviation of 1d20 is 5.916 and the standard deviation of 2d20 (or 1d20 - 1d20) is 8.165. Opposed rolls have a higher standard deviation, so I would call them more swingy. They also eat up more time at the table.

Also, you're not comparing apples to oranges here.

Sure, the 'standard deviation' that you calculate for 1d20-1d20 is higher...but the range of values is also higher. 1d20 only goes from 1 to 20. 1d20-1d20 can go from -19 to +21.

 

[The Shadow]

Hey - sudden thought I just had. (Partly inspired by ArcaneSpringboard pointing out we don't care about the size of the difference in the opposed rolls.)

What if we did care? What if this is the way out of the Save-or-Die and Save-or-Suck dilemma?

"In order to turn someone to stone, you have to beat their roll by X. If you don't, they are only slowed. If already slowed, you only have to beat their roll by Y."

This is sort of how M&M 2e does things (only there it's just a save, not an opposed roll) and it works pretty darn well. Typically a Save-or-Suck power has three tiers of effect; one when they fail by 0-4, one when they fail by 5-9, and true suckage when they fail by 10 or more. If they've got the first tier of effect already, the other two get bumped down by 5, and so on.

M&M 3e combined (almost) all such powers into a new Afflict framework that let you pick the three tiers to taste, and that also works well.

Use the 5e opposed roll the same way, and things could perhaps get quite interesting...

 

[the Jester]

How about all rolls (including attack rolls) are player-facing?

Meaning the players roll all the dice, always, and the DM never does. The DM always takes 10 on the roll.

Player A casts fireball. The monsters' defenses are 16 -- player A rolls 1d20 + something vs. the 16 DC.

Player B gets attacked by a lightning bolt. The monster's spell power is +7, taking 10 is a 17. Player B has to make a Dexterity check that beats 17.

Hell no! I'm the dm and I want to roll dice too!

 

[Libramarian]

Mechanical resolution is ultimately abstract and gamist, not something that matters 'within the story.' You'd never talk IC about any mechanical difference between saving throws and attack rolls, for instance, so having differences for the sake of difference is pointless.

No way.

The story doesn't matter on its own. It's the meaning of the story that matters, and that's constantly in a dialogue with the psychologies of the players.

The symbols and procedures involved in mechanical resolution have meaning that can be consonant or dissonant with the gameworld fiction. It's something that should be considered.

 

[GM Dave]

It does, for me. There are many time the big, bad necromancer/pyromancer/othermancer has gotten the drop on the PCs, wins initiative, and has time to cast one meaningful spell before the fighters close and attempt to have their way with him. IME, more often than not, the wizard botches his attack role or the PCs all make their saves. It's probably somewhat selective memory, but, over nearly 30 years of gaming, I can only think of one time the caster pulled off what he needed to. I can recall a half-dozen times where the caster spent round two trying to regain the advantage and round three bleeding out. Round one should bloody well count.

I agree, let the attack do what it is going to do.

This goes for fireballs by NPC wizards and dragon breath by dragons.

Let the damage roll determine if the attack was 'on target' or the players caught a lucky break or were in the center of the flames.

Thieves and monks might get a class feature to reduce the damage like the opponent has to roll the damage dice twice and they get hit by the lower of the two totals.

It is fast and simple for players and GM to use and apply.

 

[GM Dave]

Hey - sudden thought I just had. (Partly inspired by ArcaneSpringboard pointing out we don't care about the size of the difference in the opposed rolls.)

What if we did care? What if this is the way out of the Save-or-Die and Save-or-Suck dilemma?

"In order to turn someone to stone, you have to beat their roll by X. If you don't, they are only slowed. If already slowed, you only have to beat their roll by Y."

This is sort of how M&M 2e does things (only there it's just a save, not an opposed roll) and it works pretty darn well. Typically a Save-or-Suck power has three tiers of effect; one when they fail by 0-4, one when they fail by 5-9, and true suckage when they fail by 10 or more. If they've got the first tier of effect already, the other two get bumped down by 5, and so on.

M&M 3e combined (almost) all such powers into a new Afflict framework that let you pick the three tiers to taste, and that also works well.

Use the 5e opposed roll the same way, and things could perhaps get quite interesting...

I'm a fan of tier systems for these types of spells but I'd rather they apply their tiers over multiple rounds rather than one save/fail and result.

The one roll is faster but the multiple turns to final effect allows the intervention of healing, using magic, or simply killing the person inflicting the curse before it gets to the next step.

 

[Oldtimer]

Except we're not actually dealing with the size of the difference.

If we were talking about damage, sure.

But we're not. We're only talking about the sign of the difference (ie is my roll minus your roll positive or not?)

In that case, the distribution is symmetrical and the variability doesn't matter.

Well, I said that I would call them more swingy. Maybe you won't.

But it's incorrect that we only care about the sign of the difference of the rolls. We care about the sign of the difference of the results (roll + modifier). If the modifiers are different, this makes the standard deviation important.

Also, if the partioning of the rolls is entirely balanced (i.e. the fail/success divide is exactly on the statistical average), it doesn't really matter if you roll 1d20 or 100d20.

That's even if you can really talk about the standard deviation of a 1d20. That's only supposed to be used when the distribution in question approximates a normal curve, which 1d20 defnitely doesn't, and 2d20, while better, doesn't really fit that curve either.

Maybe we took different courses in statistics, but I've used standard deviation on values nowhere near a normal distribution.

 

[Falling Icicle]

One of the biggest advantages of 4e's roll vs static defense mechanic is that it allows all spells to not only miss, but to crit also. I like being able to crit with a fireball. In a system where the enemy rolls the reflex save, that doesn't happen. I'd also like to see a "crit" mechanic for non-damaging spells. This is an area of design space that static defenses makes possible and that I would like to see explored. For example, maybe a charm spell charms a monster on a hit, or dominates it on a crit. A sleep spell might slow on the first round, and put to sleep on the second, but a crit would put the target to sleep immediately. There are many possibilities.

Another thing I think is nice is that it gvies a certain degree of uniformity to the rules. Spells can still do fantastic things, but they still work in the same basic way that melee attacks do when it comes down to rolling to "hit." As for the player desire to roll saving throws, I get that. But is it really that big of a deal? Players don't roll their AC, even though that is just as important as saving throws. And as 4e demonstrated, you can use static defenses and still have rolled saving throws be a big part of the game.

So my vote is to do it the 4e way. Against save-or-suck effects, let the player roll to save on following rounds. That save doesn't need to be coin toss either, it can have a DC that is set by the caster. I think that gives the best of both worlds. Players still get to roll saves to break out of lasting effects, but the initial attack roll would be against static defenses.

 

[KidSnide]

So my vote is to do it the 4e way. Against save-or-suck effects, let the player roll to save on following rounds. That save doesn't need to be coin toss either, it can have a DC that is set by the caster. I think that gives the best of both worlds. Players still get to roll saves to break out of lasting effects, but the initial attack roll would be against static defenses.

I think the right solution is a modified version of the 3e approach. I would have wizards roll attacks when they are physically trying to hit the target with a magical blast or ray. This is similar to the melee and ranged touch attacks of 3e, except against better designed defenses. However, instead of making a minority of spells "attack" spells, make rolling to hit the standard procedure for damaging magic and Harry Potter style aimed curses.

Attacks that are resisted are appropriate for saving throws. This would mostly include unaimed AoE attacks plus traps (that really shouldn't have been rolling to hit in the first place). It would also include almost any lasting effects, whether or not the initial spell involved an attack roll. Lastly, particularly nasty-but-resistable attacks could involve a combination attack followed by a saving throw (much like the 3.5 Phantasmal Killer).

-KS

 

[Primal]

So my vote is to do it the 4e way. Against save-or-suck effects, let the player roll to save on following rounds. That save doesn't need to be coin toss either, it can have a DC that is set by the caster. I think that gives the best of both worlds. Players still get to roll saves to break out of lasting effects, but the initial attack roll would be against static defenses.

This; I think the ideal way would be an attack vs. 4E style defenses. For example, a paralyzing/petrifying attack might be vs. Fortitude; if it hits, you become dazed or slowed, but you still get to save at the end of your next round (Fort DC 18). It'd also be nice if not all saves were of the "save ends" type; just as with 4E diseases, I think it'd be cool if you could improve your condition (e.g. from paralyzed to slowed), if the effect has several "steps" (dazed/slowed/petrified, for example).