Alternative Dice Rolls

>I can't figure out how defenses would work with this in a way that's fair to either side.
What do you mean? Can't the opponent roll 2d10 to defend?

> What others, that don't involve custom dice such as FATE Die, have you run into that work well?
UESRPG. And Dark Heresy, which, I think, was the inspiration.
1d100 rolls for skills. If you start with low amount of points to spend, you still have TN of ~40 for your preferred skills, then you can go over 100 for TN. This makes game heavy on degrees of success but you will be sure that the task at hand will be accomplished by your character, unless you roll unlucky number (which is 0-5% chance).
The system is more about having/being a good GM than hoping for dice to be nice to you this time. Though, dice still fucks you up in combat, but the systems for it try to minimize that without going too far as well.

What do you think your image proves again?

The two distributions are practically identical. They have the same mean and almost the same standard deviation. If you look at the cumulative distribution, the difference never even reaches 5%.

But randomly getting a win or loss because dice is part of trpgs, isn't it?

So why would anyone not use the more precise distribution that gently rolls out at the ends for rare and spectacular success/failure over 1d10+5? 1d10+5 runs into trouble once you get modifiers such as +4 or +6 or even higher.

I like 2d10 because all the likelihood of each outcome is 1% higher/lower than the next.
11 is target number and the roll is modified by attribute, skill, gear, and my arbitrary understanding of how difficult stuff should be.

And just because I'm autistic, fate dice are d3-2 and not special at all.

>The two distributions are practically identical.
Shit bait but this is Veeky Forums so: No it's not you fucking retard, not even in theory. When the chance of rolling 3, 4, 5, 16, 17, and 18 in one system is more than infinitely more likely in the other it's not the same.
And you're missing the main point of the bell curve. Making training low skills worthwhile and mastering advanced skills difficult.

>the difference never even reaches 5%.
Then they're not practically identical, retard. Fuck off.

The dice system for the upcoming Sentinels of the Multiverse RPG is pretty interesting.

Each character has a number of Traits and Powers, along with a Condition die. When you take a test, you roll one die from your Traits, one die from your Powers and one die for your Condition, all of which can have different sizes based on the strength of the associated trait or power, or your physical condition- Although, interestingly, not all characters condition dice get smaller, as you'd expect. More mundane human characters do, but some more supernatural ones actually get more dangerous the more beaten up they get.

From these three dice, you take the middle result as the value of the check, but various different abilities within the game might shift it so you take the high or low result, or let you use two or more of the results for different actions on the roll. It's a pretty interesting idea.

Lol! I hope people like you exist for real and this wasn't just an amateur attempt at baiting.

>twice the chance of rolling a 6 or a 15
>durr they're almost the same guise

I once read a system where you had a small d10 pool + modifiers and you could exchange a d10 for a +5 modifier, which was interesting considering most attacks had a target value AND a quick test of skill against the enemies resistance, so you could allways tactically choose wether you wanted to go for a safe success that might not pass the quick test of skill with the enemy, or if you wanted to go for the full swingyness because otherwise you couldn't surpass the enemies defense. So basically trading a safe sucess against the chance for both critical failure and massive damage. Low DC utility spells for mages would often just be automatic, because there was no benefit to a high roll and the wizard was just too competent to fuck up such a little thing. That really accelerated play sometimes.

3d12 middle

Counting-free 3d6 alternative.

It's faster to compare values at a glance than add them.

Roll 3 d12. Take the middle value.

Doubles happen ~8.3% of the time. Figure it's a crit fail or success with those.

Triples happen ~0.06% of the time. Hell if I know what happens then.

Something I've been playing with is 3d6, Replace with skill.
Roll 3d6 for a melee attack with a skill of 4 for example.
> 4, 2, 5 = 11
Your skill is 4, so you can replace that 2 making your roll
>4, (4), 5 = 13
Rolling all three dice above your skill does nothing extra for you.

Now as I've used it you need traits such as Soldier, Criminal, Hunter, to apply the skill to two of your dice rolled for Melee attacks, Pickpocket checks, and Ranged attacks, respectively.

This keeps the maximum result at a 18, but raises the minimum roll and shrinks the bell curve as the character advances. Meaning that difficult lock-picking and such is possible with no skill but has a much greater chance for higher skilled characters.

>using anything but a flat distribution
>ever

I assume you're a creator in this?

Looks pretty neat in principle. I like the example of +1 swords not being useful for someone who can't use a sword to begin with; it's something I've tried to do in my homebrews (without much success). Only real criticism right now is that some of the text is unreadable when it overlaps with white/light-grey images.

Yes, just getting this baby rolling. A sample chapter will be released in coming weeks and at some point in Q1 the complete Quickstart rules. If you want to make sure you don't miss out, there's newsletter set up to get the latest info, please feel invited to sign-up.

>Only real criticism right now is that some of the text is unreadable when it overlaps with white/light-grey images.
Agreed, it's a bit difficult to read.All I can do for now is make the outline around the text a bit more pronounced. Done! Hope it helps, thanks for the feedback. I appreciate it.

Yeah, 3d6 systems have treated me well in the past.
I don't know. It kind of feels like, if the opponent has a roll as well, it means that there are some cases where you either miss or you do MASSIVE DAMAGE but with no step in between. That seems a bit off.
That sounds like fun to play with.
I don't know, to me that sounds a lot like it's just accepting mediocrity or padding for failures. How well does it work in practice?
You know, that's a fun way to play it, but what happens when you have shit for a skill but high rolls all round?

I'll believe this when I see it.

This is actually counter-intuitive. People don't understand percentage based chances of success or what a percentage bonus is.

>bell curved d100
Please explain.

What do you mean people don't understand percentages? How stupid are they?

I'll agree that percentage bonuses are very vague at times, but with some basic maths you can determine that a 10% improvement is not the same as a '+10% modifier'.

>Path of Exile: TTRPG Edition

Pretty sure it's just 2d10?

The 2d10 method for emulating a d100 is still a uniform distribution. In a d100, both digits are random and (hopefully) independent. In 2d10, both dice, which correspond to the digits, are random and independent, so the digits themselves are also random and independent. There is zero difference between their distributions.

anydice.com/program/e651

Only dice you'll ever need is d%.

anydice.com/program/e655

If I could only have 1 sort of die, I'd want d30s or knucklebones.

:\

If that's it I hope OP's game fades to obscurity before his brain-dead term for my favourite method of generating random numbers catches on.

step it up

What about for damage?

how can you roll less than 6 or greater than 15 with 1d10+5

THIS!!!!!!

1d6 + 1d6*10 + 1d6*100

It's a shit graph because it's drawn continous while the values are discrete

What the fuck is the point to this? I know it has a "nice" curve, but what is there to gain from it? It lacks both the simplicity of D20/D100 and the visceral feel and inuitivity of Dicepools.

>Some basic maths
Percentages are not basic maths, unless you are dealing with single steps, like 50 or 100%. You have made a crunch game.

Having a curve allows you to figure out which results are going to be the most likely to occur, which also helps you to plan around it when either designing a system or making changes to an existing one.

Not to mention, it means that every time you roll, you'll generally get the avg. of your roll more often than not, which helps maintain consistency and avoid the D&D problem where a folk hero of legend can miss while the peasant boy can score a critical hit just because of the way their luck played out.

I think you're getting confused. The user talking about his new game and bell curve d100s is not the op. He is however, a fucking arsehole for derailing the thread to shill his shitty failure of a game.
> my game is simple to learn and intuitive
> calculate granular percentages.

That's just because d20 based systems are fucking dogshit though. You're sleeping a bandaid over an infected wound. It's still rotting, it's just not as noticeable.

>Percentages are not basic maths
It all boils down to (100/X) man. Hell, the only difficult part of higher level math is that it usually requires more steps.

I know what curves do, but why use 3D20PickMiddle and not a quicker one or one that people are simply more used to?
How intuitive the resolution mechanic is is really fucking important because the GM has to apply modifiers on the fly.

You're misunderstanding the practice. 2d10 is rolling the tens and ones place separately rather than on one die, not adding them together like most other rolls. If your ones die rolls a 7 and your tens die rolls a 2, you get a 27, not a 9.

>That's just because d20 based systems are fucking dogshit though.
And part of the reason why that is, is because the system it's built around doesn't take into account how swingy the rolls can be when using a single die for the result.

For example, using the system from older editions where you had to roll under your character's ability score in order to do the thing, with how much you beat the result determining how well you did the thing, worked out well because it was built around the d20.

As opposed to now where the d20 is always what gives you the biggest modifier to your action while the modifiers only determine how much higher the result could be.

>I know what curves do, but why use 3D20PickMiddle and not a quicker one or one that people are simply more used to?
Because if tabletop game design was limited to what most people were used to, we'd be playing with nothing but d6's and maybe a deck of playing cards. Besides, The Dark Eye uses this method of resolution and from what I heard it's just as popular as D&D in its country of origin.
>How intuitive the resolution mechanic is is really fucking important because the GM has to apply modifiers on the fly.
It takes about as long as rolling any other dice pool would and you'd simply add a modifier to the avg. result. Hell, it'd be faster since you'd be able to go "that's really low, and this one is higher than that one, so the result is X."

OP here. I've basically just given up on this thread. It kinda quit being about interesting ways of handling dice and more about graphs and averages.

I don't know enough about statistics to get hard over bell curves and it just doesn't seem like percentiles are something that people enjoy.

m-maybe it's 2d50?
or 5 d20, but that would be awkward to add up.

Bell curves aren't better than linear distributions, they're just different.

If you want reliability, a bell curve is great, sure. But it also takes a lot of the drama out of rolls by making extreme results less likely. It works in games where you want everything to be very predictable, but in terms of what people actually enjoy, that's often not the case.

The d20 is a good die for what it does. And what it does is create a lot of drama and risk on each die roll. You can fuck up the system around it and make the die matter too much, but at the end of the day it's popular for a reason- People like the fact they can succeed or fail at any moment. It creates tension and drama in the game.

Now, some people don't like that, and for them a bell curve system is better. But don't try and assert your personal preference as objective fact.

There must be some amazing design space for modifiers to a default roll of 1d4+1d6+1d8+1d10+1d12 but to my knowledge it's never been tried.

...What would accomplish? That just seems like an unintuitive clusterfuck.

You clearly have even less imagination than I have.

So you've got nothing and just came up with a dumb gimmick for the hell of it? Got it.

If you want to prove me wrong, then actually post one decent idea you could do with that mess that wouldn't work just fine in another dice system.

>The Dark Eye uses this method of resolution and from what I heard it's just as popular as D&D in its country of origin.
I am from its country of origin and it's an autismal piece of shit that is only played by people that think they are better for playing an even shittier and way more autistic version of D&D but with German origin.
And fuck off, I know Das Schwarze Auge isn't really like D&D but it was made to fill that niche.

>If you want reliability, a bell curve is great, sure. But it also takes a lot of the drama out of rolls by making extreme results less likely.
That only matters if extreme results have special effects, and that can be emulated.
I use two d10 and count doubles as crits. That gives a 10% crit chance, same chance as rolling a nat 1 or 20, and whether it's a crit success or failure depends on whether the skill check was passed.

So like every d100 die roll?

Cause that' how you roll it in CoC and 40k RPG's.

No? It can matter without all that.

The very fact that the dice is swingy makes the act of rolling it tense. The increased uncertainty of the result, and the awareness of it, is something that can tangibly affect the experiences of the players.

This is something that needs to be used correctly, which a lot of systems fail to do, but it's still something that can provide a distinct advantage compared to more reliable dice systems.

For an attack roll, the sum of all the die results plus one set of modifiers determines whether you hit or miss, while each single die result added to different modifiers also determines other aspects of the attack, like how much damage, critical hits, added "proc" effects like a debuff, chance of a free combo attack, etc. If it's a gritty lethal game, the single die results could instead be linked to hit location tables.
Can it be done in another dice system? Of course, but that doesn't stop anyone from designing new games.

I'm all for innovative dice mechanics. I'm against gimmicky bullshit that doesn't provide an advantage or exist for a reason because 'Nobody has done it before'.

If you're going to do a weird dice mechanic, you should be able to explain what the advantages of it are by default, what aspects of the resolution system are meaningful.

The diceroll creating a result while also giving you values for lookup tables is a vaguely interesting notion, although I can see it having some kind of odd implications, but it still doesn't necessitate such a weirdly varied yet static pool.

3d6 was created because people assumed 3d6 could emulate 1d20, and use common dices to play.

Now people use 3d6 for curves.

Since the curve is the only thing important now, a middle 1 of rolling 1d20 3 times is better, you go back to a min 1 to max 20 roll like people were trying to emulate, and you dont need to do math, just look the middle value

3d6 is a lot closer to a bell curve than 3d20m

anydice.com/program/2798

>discussing dice
>without translating them into the universal language of math
Yeah, no. That's stupid.

What aspects of 1d20 are meaningful, other than being simple as shit? What aspects of 3d6 are meaningful, other than MUH BELLCURVE?

The d20 is quite easy to understand, the probability increasing in 5% increments, and while you dismiss it the simplicity of one dice plus modifiers is an advantage that makes the system easy to understand. As I've mentioned, the variability of the dice tends to add tension to the game, leading to unexpected successes or failures that can keep things exciting and dramatic in play. This needs to be used correctly, as it can also be anticlimactic or annoying, but that's more a matter of utilising it correctly than anything else.

3d6 is also relatively easy to understand, if slightly less intuitive in terms of probability, and is simple enough in play. The bell curve nature of it brings a lot of reliability, making it more appropriate for games where the sort of unexpected success or failure I talked about above isn't desired.

The two are honestly very comparable in many respects, and function as two of the most basic, standard and simple dice systems out there, just with a difference in dice tone- 3d6 is stable and reliable, 1d20 is exciting and tense. Which you want depends on context.

Having to roll a 10 or higher on a d20 or 11 or higher on 2d10 is both a 55% chance of success. The only difference is that with the bell curve initial skill improvements have greater impact and diminishing returns for further improvements.

That's kind of my point. The different in how probability changes also changes how the system feels to play.

>People don't understand percentage based chances of success or what a percentage bonus is.
I respectfully disagree. In my estimation they understand
>roll 80 or lower on this d100
much better than
>roll 5+ on this d20

This isn't me you're talking to here though.
>You have made a crunch game.
Yes, medium-crunch. Inbetween D&D 3.x and D&D 5E, that's where the sweet spot is. I am pretty confident of that.

No, it's d100, not 2d10. That would be a misnomer.

As the bell-curved d100 guy, for me the biggest benefit of the bell curve is that
a) the mid-range benefits from bonuses more than the lower or higher ranks,
b) because of diminishing returns/penalties, probabilities never go below 0% or over 100%.
A nice side-effect of that is that is makes already powerful characters less greedy in getting yet another bonus on top.

Correction. My reply was actually directed at >But it also takes a lot of the drama out of rolls by making extreme results less likely.
I kinda disagree. The less likely a result is, the more meaning players attach to it because it's so rare. A 3 in GURPS is a way more spectacular event for players than a 1 in D20. Likelihood and drama are directly correlated. But everyone has a different preference where the sweet spot lies - for some the 3 in GURPS is too rare. There's no accounting for taste.

>As the bell-curved d100 guy, for me the biggest benefit of the bell curve is that
Your system doesn't even have a bell curve, stop saying it does. You just copied CoC and 40k.

anydice.com/program/e67c

If that's what he's rolling, then the -1 is entirely superfluous ontop of that. Even if he wants the system to be 0-99 instead of 1-100

Why is that?
The -1 is to reduce the range of the Ones digit die to 0-9 instead of 1-10. Which is what they do in 40k and CoC. You read the die literally instead of assuming that the 0 is a 10.

Single die provides an equal opportunity for each.

Additional dice totaled, lead to an increasing likelihood of an average result. (A bell curve.)

Dice pools tend to take the top 25~% chance of the dice coming up. (Shadowrun 5&6 on a d6, WoD 7&8 on a d8.) And expect a person who is good at something to have enough of that thing to have a likely chance of coming up with a number of successes. (Thus how, with 4 being a common "difficult" goal, people advise having 12 at a minimum in a skill your Shadowrunner is supposed to be good in.)

Alright, if we're talking programming it into something like roll20, and you absolutely want it to be 0-99, then yeah, it isn't superfluous. For teaching players at a table, it is easier to just say 0 is always 0 on the die.

If they really want an actual bellcurve for a "d100", 10d10 would be best, especially since the floor would be 0 with the die actually reading 0=0. (Though, I swear I've seen one that had a 10, but most dice have 0=10.)

>If they really want an actual bellcurve for a "d100", 10d10 would be best, especially since the floor would be 0 with the die actually reading 0=0. (Though, I swear I've seen one that had a 10, but most dice have 0=10.)

10d10 would give a curve of 10 to 100. That wouldn't be a smart idea.
If you want a "bell curve" centered around 30, the best would the d10x system where you do d10 x d10. Has a mean of 30.25 and an STD of 23.82, which is less than a d100

anydice.com/program/e680

Well, you're right, regardless of stating that the die can come up 0.

If he's doing 0=0 then it would have a ceiling of 90. Though, that would get the full spread with an 11th d10.
anydice.com

When to use bell curves:
- when you're using degrees of succes/failure
- if you're using some kind of critical succes/fail and want skillful characters to have less disastrous fails.

Otherwise just stick to a univariate distribution. Each stat increase is a steady percentile gain in your chance of succes. This is good, since it in itself doesn't encourage optimization of stats as bell curves do; With bell curves, the highest marginal gains are around the center so optimally you should increase multiple stats instead of specialize, while with a single roll it really doesn't matter since it's always linearly increasing.

>a) the mid-range benefits from bonuses more than the lower or higher ranks,
So independently of anything else in your game, the dice already supply a method for optimization. I really dislike this.
>b) because of diminishing returns/penalties, probabilities never go below 0% or over 100%.
Are you implying that single die rolls have this problem, because I don't see it?
>A nice side-effect of that is that is makes already powerful characters less greedy in getting yet another bonus on top.
That's the same point as a) and it actually makes no sense since you arrive at the same percentage of succes sooner with a bell curve than with a univariate distribution due to higher early gains. It actually incentivizes powerful characters to be alike in terms of stats.

Am I missing something obvious here? Why do people like bell curves so much?

anydice.com/program/e681

I've come to find rolling D6+D3 pretty satisfying. The middle bulk of rolls are all equally likely, leaving only the extreme options as rarities.

2 - 1/18
3 - 2/18
4, 5, 6, 7 - 3/18
8 - 2/18
9 - 1/18

I'm not sure about your theory.

A single die has an average return of [value/2]+0.5 because the equal odds on everything means that the return is between the two middle results.

With each additional die, you increase the likelihood of the middle few results. Vastly higher than the end results.

In both cases, static increases are generally better. Unless adding dice is much easier in the system than static increases. Although if the static increases are larger than the average of the die added then they're still probably preferable from a reliability perspective.

Static increases are actually better for the multidie rolls, as the chances to spike the highest result becomes vanishingly small after even a few additional dice. (3d6 puts 3 and 18 at less than half a percent. with 9-12 being ~50% of the potential results.)

As for spreading out bonuses versus specialization, a multidie roll will ensure that things of average competence level happen more often, while things of a rarer nature happen less so. But static bonuses remain, static bonuses.

For example, there's a variant of D&D3.5 in Unearthed Arcana that subs out a d20 for 3d6.

I guess you're subjectively correct.
I remember having no trust in my character's abilities and expecting to fail simple tasks in d20 systems.
Even if the rolls were practically the same.

>if you're using some kind of critical succes/fail and want skillful characters to have less disastrous fails.
The game does just that.

>So independently of anything else in your game, the dice already supply a method for optimization. I really dislike this.
In that regard it's really the same as GURPS or Traveller or any other additive dice method. I like it but your mileage may vary.

>Are you implying that single die rolls have this problem, because I don't see it?
Yes, in Deathwatch RPG it's entirely possible to get over 100%. And in all d100 based systems it is easily possible to rack up negative modifiers that take you below 0%. Many d100-based games therefore have a cap-off, aka a min 5% and max 95% success chance or something in that range.

>That's the same point as a)
Well, it's a consequence of a), that's why I didn't list it as c).

>it actually makes no sense since you arrive at the same percentage of succes sooner with a bell curve than with a univariate distribution due to higher early gains.
I don't think that's entirely correct. If we take GURPS as example - if you have Guns-22, do you really feel compelled to expend the character points to take it to Guns-23? Maybe, but the incentive is less so that in comparable equal distribution games. You are much more compelled to take Guns from -12 to Guns-13. That is the point I was making. (It's a 3d6-based game, in case any reader is not familiar.)

>It actually incentivizes powerful characters to be alike in terms of stats.
Yes, that is a point that I will concede. Fortunately, differentiation does not have to come from the stat alone. You may have two characters who are approximately equally skilled but you have different specializations, different tricks. Given that this is a fantasy genre simulation game and just thinking of some notable fighters in Game of Thrones, I am at ease with this consequence - capable fighters with pretty much the same basic level of skill.

That's because of the 1d20.

With no bonus to a thing, and that thing being set at dc10 you've a 55% chance of success, and a 45% chance of failure on every die roll. Technically, this is why the take10/20 rules exist, but most of my play experience people have completely forgotten they exist. (That being, so that people don't fail at absolutely mundane tasks with no outside stressing factors.)

>univariate
You mean uniform, I think

>Percentages are not basic maths
...what kind of retard do you have to be to believe that?

>if you're using some kind of critical succes/fail and want skillful characters to have less disastrous fails.
Actually confirming crits solves this, but then you're not really using just a single die anymore.

>I don't think that's entirely correct. If we take GURPS as example - if you have Guns-22, do you really feel compelled to expend the character points to take it to Guns-23? Maybe, but the incentive is less so that in comparable equal distribution games. You are much more compelled to take Guns from -12 to Guns-13. That is the point I was making. (It's a 3d6-based game, in case any reader is not familiar.)
What I meant is if you want for example a 75% chance of success it'll cost you less points in a multidice system than with a single die system. So you'll have points to spend somewhere else sooner.

I'm revising a paper and it shows, I really need some sleep.

I'm really not sure what you're disagreeing with here? We seem to be making the same argument that a point increase in a skill/attribute near the average task difficulty level is worth more than a point increase when you already have extreme stat values. This encourages similar stat/skill spreads for characters since there's diminishing returns. I'm not a fan of this because I believe character specialization is important in making each character feel like they matter at some area of roleplaying. But of course dice mechanics do not a game make and there could still be plenty of niche protection built into the game elsewhere.

>Actually confirming crits solves this, but then you're not really using just a single die anymore.
Confirming crits is D20 game designer's admission that they have hit against the 5% granularity barrier.

>What I meant is if you want for example a 75% chance of success it'll cost you less points in a multidice system than with a single die system. So you'll have points to spend somewhere else sooner.
Doesn't that depend entirely on the cost of skill progression?

Bell curves seem to fit just fine with what you want from the game, no problems with that. It's just a minor pet peeve of mine when people want bell curves for their game for no apparent reason besides liking them or because they're "realistic".

>Confirming crits is D20 game designer's admission that they have hit against the 5% granularity barrier.
Confirming crits is a neat mechanism that makes sure your chance of a critical hit is always exactly 5% of your chance of a hit. I haven't seen any other mechanism that accomplishes this.

>Confirming crits is D20 game designer's admission that they have hit against the 5% granularity barrier.
Agreed, but it's a simple fix that many don't even find necessary. It's not worth changing the entire system just for this.

>Doesn't that depend entirely on the cost of skill progression?
It absolutely does.

My main point is that independent of any other game mechanics, and assuming equal importance between stats or skills, it encourages similar stat spreads.

In Harnmaster, every roll that is divisble by 5 (5,10,15,20,...) results in a Critical effect, either Success or Failure, depending on whether the test was passed or not. Very elegant and univariate.
In older CoC, you had a similar rule where 20% of your actual skill level was for critical success, iirc. In practice what you did was calculate your rollx5 and if that was still a passed test, you had scored a crit. A bit more clunky but not too difficult.
Anyway, I don't believe for a second that the intention of critical confirmation was any other than to make crits rarer than 5%. It's fairly obvious.