I'm working on a homebrew system, and I'd like some feedback on the main resolution mechanic. I'm weighing a few different options.

Brief overview: This is a dice pool system. To resolve a task, a player rolls a number of dice equal to his character's relevant attribute plus any situational bonus dice. For example, Bob is playing Bobbar the Barbarian, and wants to roll to hit a goblin. Bobbar has a Strength of 3, so Bob will start with a pool of 3 dice and might add a few bonus dice to that pool, possibly bringing it up to 6 or even 7 dice. Other factors go into whether the goblin gets smacked or not, but it's this dice pool that I'm having trouble with.

So, my problem is in what dice to use, and I've narrowed it to four options:

Option 1: d6s are okay because everyone has them. They also work well in my system for low-to-mid-level characters, in terms of the probability for success. However, they ultimately result in overpowered characters at higher levels. This option might work, but I'd have to come up with some creative ways to limit high-level characters outside the core resolution mechanic.

Option 2: d8s are great because the probability distribution works very well for characters at all levels. However, most people don't have tons of d8s just lying around.

Option 3: d10s are okay because, again, most people have enough lying around. They also work well for high-level characters in my system. However, they lead to severely underpowered characters at low- and mid-levels. I think this is reason enough to scrap option 3.

Option 4: A combination of d6s and d10s works very well. As above, most people have plenty of d6s and d10s. Also, if d6s are used for attribute dice (the initial pool) and d10s are used for any bonus dice, the probability distribution for success is excellent; maybe even better than with pools of d8s. However, this option creates a little extra complexity in the mechanic, since two kinds of dice are used.

No one option is perfect. I lean toward option 4 because it solves the mechanical problems and doesn't alienate anybody by using unusual dice, but it does cause added complexity to the system. Do you think one of those options is better than the others? And do you think that option 4 adds too much complexity to the system?

I like Option 1, as it is close to a favorite dice pool based game of mine (Atomic Highway) and my own approach these days.

In my experience Option 1 is not overpowered so I'd like to hear your examples.  

Assuming there are no exploding dice, which I find highly problematic both statistically and practically.  

A question, do you have any modifiers to the dice roll?  For example, in the Atomic Highway d6 Engine system you might have a barbarian of Strength 3 (on a scale of 1-5) and a Axe skill 2.  You'd roll 3 (three) d6 then have 2 points (from the 2 skill) to add.  1 point to each of two dice, or two points to one die, to help you get a 6, which counts as a success.  

Giving extra d6 is really not a problem it just allows you to engage more targets or one much more successfully.  So a loner low level target is doomed against a high level adversary with many dice (so what that is the genre) but that same high level adversary has a challenge against say a dozen low level targets whom may have combined just as many dice as the high level adversary.

In all giving a d6 per level is basically Chainmail.  Not going to grab out the table but a hero being equivalent to x men, meant he got x dice, while the normal man at arms got 1 die.

My inclination is to say d6s, because you can find boxes of 30+ d6s for sale in pretty much any game shop.  No other die type is so easily accessible in bulk.  d10s come second on that criterion, thanks to WOD games spawning a market for five-packs of d10s.

But the main concern you're expressing here is that you seem to think that d6s end up being overpowered at high levels and d10s underpowered... but you don't tell us anything about why that might be, so we can't really evaluate the options in those terms.  What makes d6s more powerful than d10s?  Can that factor be adjusted so that any given die type will produce results matching your desired power level?

The system uses a roll-under target number, where the target number is variable by skill level. To continue the example from above, Bobbar rolls 3d6 (because he has a Strength of 3) and a few bonus dice. Let's say he has a total of 5d6. Bobbar is fairly new to barbarianing, so he only has a skill of 2. This means that any dice that come up 2 or less will count as successes. He'll have an 86% chance of at least one success, and that works well in the system. At a skill level of 5, any dice that comes up 5 or less will count as a success. So if he rolls 5d6, he is virtually guaranteed at least three successes. This alone isn't broken. The problem arises with powergamers, but they must be accounted for. If a character has a skill level of 5 and used some other mechanics to roll pools of 7 or 8 dice regularly, they will be able to trivialize virtually any encounter the GM could come up with (similar to creating a character that has a 95% chance to beat Difficulty Class 40 in D&D5e).

On the other hand, if I use d10s, it reigns in the powergamers because it makes each die much less likely to succeed. Hence, consistently maximizing the dice pool won't break the game. However, it also makes it unreasonably difficult for any character to get a success without having at least a 4 in the relevant skill.

Combining d6s and d10s largely fixes this problem. It makes the base pool much more important, in terms of generating successes. Bonus dice have less impact. But this also adds a layer of complexity to the basic resolution mechanic in a system that is intended to be relatively streamlined.

Quote from: O.Owl;938010The system uses a roll-under target number, where the target number is variable by skill level. To continue the example from above, Bobbar rolls 3d6 (because he has a Strength of 3) and a few bonus dice. Let's say he has a total of 5d6. Bobbar is fairly new to barbarianing, so he only has a skill of 2. This means that any dice that come up 2 or less will count as successes. He'll have an 86% chance of at least one success, and that works well in the system. At a skill level of 5, any dice that comes up 5 or less will count as a success. So if he rolls 5d6, he is virtually guaranteed at least three successes. This alone isn't broken. The problem arises with powergamers, but they must be accounted for. If a character has a skill level of 5 and used some other mechanics to roll pools of 7 or 8 dice regularly, they will be able to trivialize virtually any encounter the GM could come up with (similar to creating a character that has a 95% chance to beat Difficulty Class 40 in D&D5e).

Nifty mechanic.  The problem you have here is character power increases geometrically.   Not only do you get more dice you also get a better chance per die of succeeding.   That being said, a d8 may be the better way to go.   A lot of DMs have d8's galore.  I wouldn't worry too much about dice choice making the game hard to sell, any new RPG is hard to sell.

If you stick with d6 you can cap the number of dice you can apply, maybe by the nature/difficulty of the task.  Really difficult tasks allow only so much power (dice) to bear on the problem.

QuoteOn the other hand, if I use d10s, it reigns in the powergamers because it makes each die much less likely to succeed. Hence, consistently maximizing the dice pool won't break the game. However, it also makes it unreasonably difficult for any character to get a success without having at least a 4 in the relevant skill.

Combining d6s and d10s largely fixes this problem. It makes the base pool much more important, in terms of generating successes. Bonus dice have less impact. But this also adds a layer of complexity to the basic resolution mechanic in a system that is intended to be relatively streamlined.

The d6 / d10 distinction can also be used to advantage.  You can have some skills, bonuses, etc. only apply to one type of die.  There could be magic items that give you a d6 versus a d10, and vice versa.

Ya the main issue I see is so little grain between values, so there's not much subtlety between levels, especially for lower-powered people. It's hard to judge without knowing more about the system. It seems like the intent though is to have chunky power levels yet still have some uncertainty.

Quote from: O.Owl;938010The system uses a roll-under target number, where the target number is variable by skill level. To continue the example from above, Bobbar rolls 3d6 (because he has a Strength of 3) and a few bonus dice. Let's say he has a total of 5d6. Bobbar is fairly new to barbarianing, so he only has a skill of 2. This means that any dice that come up 2 or less will count as successes. He'll have an 86% chance of at least one success, and that works well in the system. At a skill level of 5, any dice that comes up 5 or less will count as a success.

While I see what you're trying to get at (skills give both more dice and an easier TN; bonuses give only more dice), I'm very skeptical of that kind of double-dipping, where the same thing gives both more dice and an improved chance of success per die.  I'm sure it would also give the "transparent probabilities" crowd fits, too.

If you're not familiar with it, you might want to take a look at Tenra Bansho Zero, which uses a fairly similar system:

Skills are rated on a scale from 1 to 5 and your skill rank acts as a roll-under target number, using d6s.

Attributes are your base number of dice to roll on a skill test, which can be modified by other factors, both situational/environmental and by spending aiki/kiai points (basically fate points, but not really).

The main difference from what you've described is that each thing (skill level, attribute level, bonuses, etc.) only affects one variable in the inputs to the resolution system (i.e., only the number of dice or the target number) instead of multiple variables.  This makes it a lot easier to tune the system, since the effects of any change are more constrained.

Quote from: nDervish;938104While I see what you're trying to get at (skills give both more dice and an easier TN; bonuses give only more dice), I'm very skeptical of that kind of double-dipping, where the same thing gives both more dice and an improved chance of success per die.  I'm sure it would also give the "transparent probabilities" crowd fits, too.

If you're not familiar with it, you might want to take a look at Tenra Bansho Zero, which uses a fairly similar system:

Skills are rated on a scale from 1 to 5 and your skill rank acts as a roll-under target number, using d6s.

Attributes are your base number of dice to roll on a skill test, which can be modified by other factors, both situational/environmental and by spending aiki/kiai points (basically fate points, but not really).

The main difference from what you've described is that each thing (skill level, attribute level, bonuses, etc.) only affects one variable in the inputs to the resolution system (i.e., only the number of dice or the target number) instead of multiple variables.  This makes it a lot easier to tune the system, since the effects of any change are more constrained.

I'm not familiar with Tenra Bansho Zero, but that's actually what I'm describing, in terms of the basic resolution mechanic. I guess I'm not explaining it clearly, but I wasn't intending any double dipping. Skills don't give both more dice and an easier TN - they only modify the TN. Attributes set the initial size of the dice pool. Additional dice can be added, but not by skills. Bonus dice would come from assistance from other characters, etc. The only way a skill would add more dice would be if the character specialized in a skill (so she would get a bonus dice when using that skill under very particular circumstances).

Quote from: O.Owl;937965This is a dice pool system. To resolve a task, a player rolls a number of dice equal to his character's relevant attribute plus any situational bonus dice. For example, Bob is playing Bobbar the Barbarian, and wants to roll to hit a goblin. Bobbar has a Strength of 3, so Bob will start with a pool of 3 dice and might add a few bonus dice to that pool, possibly bringing it up to 6 or even 7 dice. Other factors go into whether the goblin gets smacked or not, but it's this dice pool that I'm having trouble with.

This isn't enough information for us to give you a meaningful answer.

Quote from: O.Owl;938010The system uses a roll-under target number, where the target number is variable by skill level. To continue the example from above, Bobbar rolls 3d6 (because he has a Strength of 3) and a few bonus dice. Let's say he has a total of 5d6. Bobbar is fairly new to barbarianing, so he only has a skill of 2. This means that any dice that come up 2 or less will count as successes. He'll have an 86% chance of at least one success, and that works well in the system. At a skill level of 5, any dice that comes up 5 or less will count as a success. So if he rolls 5d6, he is virtually guaranteed at least three successes.

That's better.

Quote from: O.Owl;938010The problem arises with powergamers, but they must be accounted for.

Actually I'd be less concerned about #Powergamers or any specific die type and more concerned about how your game works as a whole.

Quote from: O.Owl;938010If a character has a skill level of 5 and used some other mechanics to roll pools of 7 or 8 dice regularly, they will be able to trivialize virtually any encounter the GM could come up with (similar to creating a character that has a 95% chance to beat Difficulty Class 40 in D&D5e).

Then don't let dice pools get to 7+ on a regular basis? Because we still don't know enough about your system to give you a meaningful answer.

If it helps, I used to run WoD (where Skills and Attributes cap at 5) with a system very similar to this, where the die pools started at 2 + Attribute, and could only be expanded using Blood Dice, which were lost if they rolled a Success. Worked fairly well at encouraging the kind of play I was after.

I use a 1d8 system in Salem World that is role under, but not a dice pool.  We cap the maximum scores at 7 so there is always a chance of failure.

I agree with others that the die type and power level, even the pool itself becomes less important than if it works.  Not knowing enough detail of the system in question makes it hard to speculate, but I'll try.

My Character has a skill of 5 as stated.  They have "other Mechanics" (I assume skills and such for bonuses) that add dice to that total.  So for a feat of strength, and my Strength (if that is how it works) is a 5, plus my Body Building of a 2 and a skill in Throwing of 2, and a natural Race/class bonus of 1 I get a Dice Pool of 10d10 (or whatever dice you end up with).  If I assume that you still roll for the 5, then there is a high probability of at least one 5 or less.  

Most characters will not know failure in that example.

As stated before, you can cap your pool at a given number.  This will act as a power regulator in the mechanics, allowing for a need for character growth.

Quote from: O.Owl;938116I'm not familiar with Tenra Bansho Zero, but that's actually what I'm describing, in terms of the basic resolution mechanic. I guess I'm not explaining it clearly, but I wasn't intending any double dipping. Skills don't give both more dice and an easier TN - they only modify the TN. Attributes set the initial size of the dice pool. Additional dice can be added, but not by skills. Bonus dice would come from assistance from other characters, etc. The only way a skill would add more dice would be if the character specialized in a skill (so she would get a bonus dice when using that skill under very particular circumstances).

Ah, ok, then I guess I misunderstood your earlier description.  TBZ is a translation of a Japanese game, designed to emulate high-powered anime insanity, with samurai who power up and transform into grotesque monsters, ninja jumping from rooftop to rooftop with an army of mirror image clones, young children piloting physics-defying powered armor suits, etc.  So the power level is almost certainly well above what you have in mind for your game, but scaling down attribute levels should be able to deal with that easily enough.

I appreciate the feedback. I think I've found a solution I'm satisfied with, by using small pools of d10s and adjusting the way the mechanic handles difficulty levels to fix the issues that concerned me. If, for some reason that doesn't work, I'll use the combination of d6s and d10s, since you don't think that mechanic would be too complicated.