Anything I missed with this Dice Pool System?

jakinbandw

Registered User
Validated User
#1
So I've been trying to come up with a good resolution system for a game I'm designing, and after lots of trying to just use degrees of success I've found myself drawn to a dice pool system.

The thing is that I prefer a bit more of a flat curve than most dice pools give. I like having a number of successes a player can use on abilities, but I don't like the curve. So this was my solution. Please let me know if it is too complex, or I'm missing something.

When you make a skill check you roll a number of d6s equal to your power and count the number of successes on them. Each 1 gives you no successes, each 5 or 6 gives you 2 successes, and everything else gives you a single success. These are your Power Dice. Next your roll a d6 and add your skill level. If the result is 7 or greater you gain a single additional success. This is your skill dice. You can roll them both at the same time if you have ways of telling your power dice and skill die apart
This is the anydice code for the probabilities. Go into summery mode because it wasn't really written for anything else.
https://anydice.com/program/126bd

This is my first attempt working with dice pools, so let me know if I'm missing something. I chose the numbers I did because it means that a player has an 83% of success to roll a number of successes equal to their power level when they have a skill of 1 (outside of power 0 and power 1). This allows an easier time eyeballing the number of successes.
 
Last edited:

ThornyJohn

Registered User
Validated User
#2
There isn't a lot there for you to be missing anything, yet. Of what you have, I'm wondering about the fiddliness of the extra die added to the roll. Most dice pool systems either just count successes or have sixes explode for an extra roll. I'm sure there are a few out there that add extra dice for one reason or another (I've done it myself on a few attempts), but most tend to keep it simple.

In any case, I'm not sure where any of this (the dice pool and the extra die roll) is making things "flatter." If you want flatness, you want simple rolling, such as a single d20 or d%, both of which give you a simple percentage chance of succeeding or failing while keeping things about as flat as they can be. If you want something less flat but still far away from rolling a very bell-curvy bunch of d6s and counting successes, you might want to consider something like rolling two or three larger dice, say d10s, d12s or d20s and picking the single highest result.

Ultimately, it's going to depend heavily on what feels best for your system, though. Don't try to squeeze a dice mechanic into the system just because the numbers "look right." Best to see how they actually play.
 

jakinbandw

Registered User
Validated User
#3
There isn't a lot there for you to be missing anything, yet. Of what you have, I'm wondering about the fiddliness of the extra die added to the roll. Most dice pool systems either just count successes or have sixes explode for an extra roll. I'm sure there are a few out there that add extra dice for one reason or another (I've done it myself on a few attempts), but most tend to keep it simple.

In any case, I'm not sure where any of this (the dice pool and the extra die roll) is making things "flatter." If you want flatness, you want simple rolling, such as a single d20 or d%, both of which give you a simple percentage chance of succeeding or failing while keeping things about as flat as they can be. If you want something less flat but still far away from rolling a very bell-curvy bunch of d6s and counting successes, you might want to consider something like rolling two or three larger dice, say d10s, d12s or d20s and picking the single highest result.

Ultimately, it's going to depend heavily on what feels best for your system, though. Don't try to squeeze a dice mechanic into the system just because the numbers "look right." Best to see how they actually play.
The system I'm working on was using degrees of success to get things done. For example, a wizard in combat casts a spell and needs a certain amount of degrees of success total before it goes off. In the end, I decided that just using number of successes was easier to understand. My system is also a bit on the exponential side of power. Before this meant that after a certain bonus you were now in the super human tier (but also had a ranking in it). A dice pool preserves that while making sense.

Finally, as I said, I don't like the way that dice pools curve out. So this lets me track power tier, power within that tier, and is fairly flat.

My big concern is if this is too fiddly, what with rolling two different types of dice.
 

kenco

Registered User
Validated User
#4
So I've been trying to come up with a good resolution system for a game I'm designing, and after lots of trying to just use degrees of success I've found myself drawn to a dice pool system.

The thing is that I prefer a bit more of a flat curve than most dice pools give. I like having a number of successes a player can use on abilities, but I don't like the curve. So this was my solution. Please let me know if it is too complex, or I'm missing something.



This is the anydice code for the probabilities. Go into summery mode because it wasn't really written for anything else.
https://anydice.com/program/126bd

This is my first attempt working with dice pools, so let me know if I'm missing something. I chose the numbers I did because it means that a player has an 83% of success to roll a number of successes equal to their power level when they have a skill of 1 (outside of power 0 and power 1). This allows an easier time eyeballing the number of successes.
Not sure that I understand the description:
- what is 'Power' (as contrasted with 'Skill')? What is the range of values for 'Power'?
- how do I roll a 0 on a d6?

I'm also not sure what you need this procedure to achieve. It sounds like you want a Skill 1 player to generally score successes equal to their Power rating, you need a minimum number of Successes to activate certain effects (and probably extra Successes can be spent for flourishes or enhancements of the base effect?). Beyond that, I'm not very clear. So it's hard to judge suitability.

Presumably you are happy with the probability distribution, and are asking for feedback mainly about what it will be like to play.

Overall it sounds a little bit fiddly. I think having to sort the Power dice into three(?) different piles to count them might be painful (although I still can't see how to roll a '0'). Custom dice would fix that problem. Or maybe you could simplify the reading of the power dice to two categories?

Having said this, I find counting BODY on Champions dice pretty easy (1 = 0, 2-5 = 1, 6 = 2). But that's because I played it a fair bit back in the day. Maybe regular play would get people to that point with your Power Dice.

I notice that your Skill die method Roll + Skill >=7 is functionally the same as Roll <= Skill. The latter would be quicker and simpler to my mind.
 
Last edited:

spaceLem

Green haired rodent
Validated User
#5
Why not take a look at the Unisystem way of handling things?

In Unisystem you roll Stat + Skill + d10, and compare the result on a table. It's something like
  • 0-8 is a fail
  • 9-10 is 1 success
  • 11-12 is 2 successes
  • 13-14 is 3
  • 15-16 is 4
  • 17-20 is 5 (it goes a bit wobbly here)
  • 21-23 is 6
  • 24-26 is 7
(or something along those lines). It's a mostly non-linear curve but it basically works like a flattened dicepool.

Personally I'd adjust that a bit and say roll Stat + Skill + 2d6 (average 7), then
  • 0-8 is a fail
  • 9-11 is 1 success
  • 12-13 is 2 successes
  • 14-16 is 3 successes
  • +3 is +1 success.
Then your expected number of successes is 1/3rd of Stat + Skill, which is kind of like throwing Nd6 and counting 5-6 as successes, only with a flat distribution instead of a binomial distribution.
 

jakinbandw

Registered User
Validated User
#6
Not sure that I understand the description:
- what is 'Power' (as contrasted with 'Skill')? What is the range of values for 'Power'?
- how do I roll a 0 on a d6?

I'm also not sure what you need this procedure to achieve. It sounds like you want a Skill 1 player to generally score successes equal to their Power rating, you need a minimum number of Successes to activate certain effects (and probably extra Successes can be spent for flourishes or enhancements of the base effect?). Beyond that, I'm not very clear. So it's hard to judge suitability.

Presumably you are happy with the probability distribution, and are asking for feedback mainly about what it will be like to play.

Overall it sounds a little bit fiddly. I think having to sort the Power dice into three(?) different piles to count them might be painful (although I still can't see how to roll a '0'). Custom dice would fix that problem. Or maybe you could simplify the reading of the power dice to two categories?

Having said this, I find counting BODY on Champions dice pretty easy (1 = 0, 2-5 = 1, 6 = 2). But that's because I played it a fair bit back in the day. Maybe regular play would get people to that point with your Power Dice.

I notice that your Skill die method Roll + Skill >=7 is functionally the same as Roll <= Skill. The latter would be quicker and simpler to my mind.
Power is a 0-5 scale that measures your power level. The scale looks something like this (with a strength example in brackets):

0: Untrained (Normal human)​
1: Professional (Deadlifting a thousand pounds)​
2: Low Superhuman (Lifting cars)​
3: High Super human (Lifting skyscrapers)​
4: Low Divinity (Lifting cities)​
5: High Divinity (Lifting mountains or moving planets)​


Skill is a 1-5 scale which measures where you are on each scale. At power rank 2, skill 1 you may just be working hard to lift a single car, but at power rank 2, skill 5 you're tossing them easily, one in each hand (at least, that's the design goal). The idea of the skill dice is that it would also work as a tie breaker when the number of successes were tied in an opposed roll.

I also mistyped the post, when you roll a 1 on a d6, that counts as 0 successes. That's been fixed. As to just having two piles... I'll work on it. The problem I've been having is that the number of expected successes becomes odd (power 1, skill 1 averages one success, power 3 skill 3 averages 3 successes.) I'll mess around with different dice to see if I can get it to work better.

Roll under might work really well. I'll test to see if that works better. I can always reverse the power dice so that 1 and 2 give two successes, and a 6 gives no successes. My only concern is that it might be harder to use the skill dice to break ties then, but it's a good idea. Less math is usually a good thing.



Why not take a look at the Unisystem way of handling things?

In Unisystem you roll Stat + Skill + d10, and compare the result on a table. It's something like
  • 0-8 is a fail
  • 9-10 is 1 success
  • 11-12 is 2 successes
  • 13-14 is 3
  • 15-16 is 4
  • 17-20 is 5 (it goes a bit wobbly here)
  • 21-23 is 6
  • 24-26 is 7
(or something along those lines). It's a mostly non-linear curve but it basically works like a flattened dicepool.

Personally I'd adjust that a bit and say roll Stat + Skill + 2d6 (average 7), then
  • 0-8 is a fail
  • 9-11 is 1 success
  • 12-13 is 2 successes
  • 14-16 is 3 successes
  • +3 is +1 success.
Then your expected number of successes is 1/3rd of Stat + Skill, which is kind of like throwing Nd6 and counting 5-6 as successes, only with a flat distribution instead of a binomial distribution.
I'll play around with this. I don't think it will take over the entire system, but I could see using it for the power dice. Thanks for the idea. I'll see what I can do with this!
 

jakinbandw

Registered User
Validated User
#7
Not sure that I understand the description:
- what is 'Power' (as contrasted with 'Skill')? What is the range of values for 'Power'?
- how do I roll a 0 on a d6?

I'm also not sure what you need this procedure to achieve. It sounds like you want a Skill 1 player to generally score successes equal to their Power rating, you need a minimum number of Successes to activate certain effects (and probably extra Successes can be spent for flourishes or enhancements of the base effect?). Beyond that, I'm not very clear. So it's hard to judge suitability.

Presumably you are happy with the probability distribution, and are asking for feedback mainly about what it will be like to play.

Overall it sounds a little bit fiddly. I think having to sort the Power dice into three(?) different piles to count them might be painful (although I still can't see how to roll a '0'). Custom dice would fix that problem. Or maybe you could simplify the reading of the power dice to two categories?

Having said this, I find counting BODY on Champions dice pretty easy (1 = 0, 2-5 = 1, 6 = 2). But that's because I played it a fair bit back in the day. Maybe regular play would get people to that point with your Power Dice.

I notice that your Skill die method Roll + Skill >=7 is functionally the same as Roll <= Skill. The latter would be quicker and simpler to my mind.

Okay, I've an alternative that I think I could make work.

You roll twice your power rating of d6's. 1's and 2's fail, everything else counts as a single success. Roll one extra d6 and roll equal or less then your skill to get a single extra success.

https://anydice.com/program/12740


Why not take a look at the Unisystem way of handling things?

In Unisystem you roll Stat + Skill + d10, and compare the result on a table. It's something like
  • 0-8 is a fail
  • 9-10 is 1 success
  • 11-12 is 2 successes
  • 13-14 is 3
  • 15-16 is 4
  • 17-20 is 5 (it goes a bit wobbly here)
  • 21-23 is 6
  • 24-26 is 7
(or something along those lines). It's a mostly non-linear curve but it basically works like a flattened dicepool.

Personally I'd adjust that a bit and say roll Stat + Skill + 2d6 (average 7), then
  • 0-8 is a fail
  • 9-11 is 1 success
  • 12-13 is 2 successes
  • 14-16 is 3 successes
  • +3 is +1 success.
Then your expected number of successes is 1/3rd of Stat + Skill, which is kind of like throwing Nd6 and counting 5-6 as successes, only with a flat distribution instead of a binomial distribution.
I messed around with this, but I don't like the amount of math. I did try doing stuff using df, but it never came out in a way I liked. Thanks for the idea though.
 

spaceLem

Green haired rodent
Validated User
#8
I messed around with this, but I don't like the amount of math.
Don't try to subtract 7 and divide by 3, that way lies madness! The trick to removing the maths is to write a table down and have it in front of you, which is how the Buffy the Vampire Slayer RPG did it (a game not known for being difficult to play).
 

Knaight

Registered User
Validated User
#9
The First Proposed System
On skill dice: Roll and add vs. 7 and roll at or under are effectively identical (it changes which numbers mean what, but the probabilities are the same), so you might want to swap them here to remove a constant.

As for how this would play, I'm seeing some points of concern. The big one is that there are pretty huge jumps when moving up in power - which might well be intentional, but is particularly massive at low levels, and really stand out if opposed rolls are used. Take the 0 to 1 gap, where even a 0.5 (Power.Skill) against 1.1 has only an 11.6% chance of victory. Comparing 1.5 to 2.1 though? That's a 22.15% chance now. 2.5 to 3.1 is a 26.99% chance. Then 29.95%. Then 31.99%. The increase in power matters most at the low end of the scale, and while this might be intentional it just feels off to me.

Power being 0-5 and skill 1-5 also just feels clunky.

The Second Proposed System
I'm liking the change to skill dice. Extracting the same marginal data, this time we see: 7.72%, 17.64%, 22.93%, 26.27%, 28.61%. Overall these are pretty similar curves, with the impact of the skill die reduced further. They actually look a bit better without the doubling, which also neatly removes a mathematical operation.

Working From Goals
The core of what you want, as described, as I understand it, is two things, with some suggested extras of lesser importance:
1) Variable number of successes, easily found (preferably counted).
2) A fairly flat curve.
3) Low granularity (I don't know if exactly 30 levels in 6 categories is aimed for, but the point is it's much more granular than most dice pool systems).
4) Usually beating the power level.
5) Minimal math

This strongly suggests that you want to stick to relatively few dice, and change the dice themselves as much as possible. It's not really a dice pool system at all anymore if you do this, but it can keep success counting and other dice-pool concepts around. The obvious ways of doing this also tend to veer into bad practice (i.e. changing both the number of dice and the number to succeed, which has largely been dropped from modern mechanics for good reason. That said there are ways to do this involving more modern distinct colored dice techniques that also allow for some fine graining, which can get you down to fairly few dice pretty easily, albeit often in ways that compromise the flatness or take a lot of dice colors (a 1-30 scale can be fit on 1-5 dice pretty easily, though the methods for doing so satisfyingly often get pretty janky somewhere).
 

jakinbandw

Registered User
Validated User
#10
The First Proposed System
On skill dice: Roll and add vs. 7 and roll at or under are effectively identical (it changes which numbers mean what, but the probabilities are the same), so you might want to swap them here to remove a constant.

As for how this would play, I'm seeing some points of concern. The big one is that there are pretty huge jumps when moving up in power - which might well be intentional, but is particularly massive at low levels, and really stand out if opposed rolls are used. Take the 0 to 1 gap, where even a 0.5 (Power.Skill) against 1.1 has only an 11.6% chance of victory. Comparing 1.5 to 2.1 though? That's a 22.15% chance now. 2.5 to 3.1 is a 26.99% chance. Then 29.95%. Then 31.99%. The increase in power matters most at the low end of the scale, and while this might be intentional it just feels off to me.

Power being 0-5 and skill 1-5 also just feels clunky.

The Second Proposed System
I'm liking the change to skill dice. Extracting the same marginal data, this time we see: 7.72%, 17.64%, 22.93%, 26.27%, 28.61%. Overall these are pretty similar curves, with the impact of the skill die reduced further. They actually look a bit better without the doubling, which also neatly removes a mathematical operation.

Working From Goals
The core of what you want, as described, as I understand it, is two things, with some suggested extras of lesser importance:
1) Variable number of successes, easily found (preferably counted).
2) A fairly flat curve.
3) Low granularity (I don't know if exactly 30 levels in 6 categories is aimed for, but the point is it's much more granular than most dice pool systems).
4) Usually beating the power level.
5) Minimal math

This strongly suggests that you want to stick to relatively few dice, and change the dice themselves as much as possible. It's not really a dice pool system at all anymore if you do this, but it can keep success counting and other dice-pool concepts around. The obvious ways of doing this also tend to veer into bad practice (i.e. changing both the number of dice and the number to succeed, which has largely been dropped from modern mechanics for good reason. That said there are ways to do this involving more modern distinct colored dice techniques that also allow for some fine graining, which can get you down to fairly few dice pretty easily, albeit often in ways that compromise the flatness or take a lot of dice colors (a 1-30 scale can be fit on 1-5 dice pretty easily, though the methods for doing so satisfyingly often get pretty janky somewhere).
Thanks for your feedback! This was incredibly helpful. You were right on all 5 goals I had, and stated them clearer than I managed in my design notes. Taking your points into account as best as I could (and trying to keep Jank to a minimum) this is what I came up with:


Third Proposed System

Roll a number of power dice equal to your power ranking. Each roll less than 6 is a success. Roll a single skill dice. If the result is equal or less than your Skill or Power ranking you get a success. If it's equal to or less than both your skill and power ranking you get two successes. On an opposed roll on a tie, the character with the highest power wins. If they have the same power, the character with the higher skill wins. If the skills are the same as well, then the winner is whoever rolled the lowest on the skill dice.
https://anydice.com/program/128a2

This keeps the odds of succeeding against someone stronger than you about the same throughout as far as I can tell. It also means that now Skill can be ranked 0-6 along with power being ranked 0-6. That said, someone with 0 Skill and 0 Power will never succeed, which I think is fair!

It also removes the jump between power levels, or at least reduces them by quite a bit. It keeps the number of dice down to a max of 6 and makes the power dice easy to count (remove all 6's). The only increase in difficulty comes from the skill dice, but I don't think it's too bad. It also means that if you have Power 5 and Skill 5 you will either get 2 successes on your skill dice, or none at all. This is kinda weird, but the number of power dice you are rolling tends to make up for it probability wise. It's just a bit odd.

[Edit] Just realized that this doesn't actually need d6's. You could have skills and power go from 0 to 3 and use a d4 and the probabilities don't shift too much.
 
Last edited:
Top Bottom