Dice pool mechanic

[Robert Scott Clark]

I'm toying around with a dice pool mechanic and I'm looking to see if there
is anything obvious I'm missing about the results.


Basically, I wanted a simple WWesque stat+skill pool that differentiated
between stat and skill in some way. A roll is made with a number of d6s
equal to the stat and a number of d10s equal to the skill. The result is
the highest number rolled on any die increased by 1 for every extra 6
rolled on a d6 or 10 rolled on a d10 - compare that to the target number
and voila.

Now, the results that jump out at me are...

High stats make you reliable but will not result in the highest values
without skill.

Minimal skill training can produce a high value.

Higher skill produces reliably higher results, but will only minimally
allow you to surpass what a lower skill can get (makes getting a 10 easier,
but low chance of going to 11 or higher)

Skill alone will not produce the highest results. Even without limiting
the max skill, high stats are more useful in producing exceptional results
(it's easier to roll the 6 on the d6 than the 10 on the d10)


I was also trying to think of some different benefit to rolling the 6/10,
that could be taken in addition to/instead of the +1. My gut reaction was
some form of minor critical success for a 6 and protection from
fumble/mishap for a 10 - to represent fluke success for the prodogy and the
experience of the high skill individual, but I coundn't figure out a good
system for fumbles/mishaps that didn't require a lot more rolling or
produce odd results (ie. having each 1 roll represent a minor mishap has
the silly effect of making more dice produce more mishaps.)

[James O'Rance]

Post by Robert Scott Clark
Basically, I wanted a simple WWesque stat+skill pool that differentiated
between stat and skill in some way. A roll is made with a number of d6s
equal to the stat and a number of d10s equal to the skill. The result is
the highest number rolled on any die increased by 1 for every extra 6
rolled on a d6 or 10 rolled on a d10 - compare that to the target number
and voila.

That's quite interesting. What range of difficulty numbers are you
looking at? What's the average difficulty?

A maximum difficulty of 10 should work. You could possibly go to 11
with the 6/10 +1 option, but these tasks will be so difficult that I
don't see the point. If you have some kind of botching mechanic (eg,
1s cancel out successes) then succeeding at a difficulty 11 task would
be very rare indeed.

An "average difficulty" of 5 feels about right to me. This gives the
unskilled character a reasonable chance to succeed at a task. An
average difficulty of 6 could also be okay, depending on how many dice
the character is likely to have in her dice pool.

However, if the average difficulty is set at 7 or higher, then it
becomes impossible for the unskilled character to accomplish (without
the 6/10 +1 option).

Post by Robert Scott Clark
I was also trying to think of some different benefit to rolling the 6/10,
that could be taken in addition to/instead of the +1. My gut reaction was
some form of minor critical success for a 6 and protection from
fumble/mishap for a 10 - to represent fluke success for the prodogy and the
experience of the high skill individual, but I coundn't figure out a good
system for fumbles/mishaps that didn't require a lot more rolling or
produce odd results (ie. having each 1 roll represent a minor mishap has
the silly effect of making more dice produce more mishaps.)

Hmmm. Maybe a variant on botch rules from the Revised Storyteller
games?

If you roll a 1 on 1d6, it cancels a success.
If all of your successes are cancelled, the action fails (having more
1s than successes has no additional effect).
If you roll a 1 on 1d6 and have no successes to cancel, you get a
mishap.
Rolling 1 on 1d10 has no special effect.

This means that high skill protects you somewhat from mishaps; high
dice pools don't produce a greater chance of mishaps than low dice
pools; and no additional dice rolling is required.

--
james o'rance
http://www.geocities.com/dragon-dreamer

[Robert Scott Clark]

Post by James O'Rance

Post by Robert Scott Clark
Basically, I wanted a simple WWesque stat+skill pool that
differentiated between stat and skill in some way. A roll is made
with a number of d6s equal to the stat and a number of d10s equal to
the skill. The result is the highest number rolled on any die
increased by 1 for every extra 6 rolled on a d6 or 10 rolled on a d10
- compare that to the target number and voila.

That's quite interesting. What range of difficulty numbers are you
looking at? What's the average difficulty?

Still toying around, but I was thinking of always using opposed rolls -
easiest task 1d6, near impossible task 6d6+8d10. Average? 3d6 maybe?

I like the idea because it could differentiate between a failure because
of a faulty attempt (low player roll) or because of a complication (high
opposed roll)

(I just got a flash that maybe high skill is better at countering
complications than high stat is, so maybe a 10 could remove an opponent
die instead of adding the +1 to the result. I'll have to think a little
more about the ramifications of that - especially what that does to the
odds.)


I'm not set on that yet, as I can't really decide what the difference in
dice size would mean for an opposed roll vs anything other than a
character. I mean, if there's going to be different size dice to
represent the difficulty of defusing a bomb, there should be some
rationale for saying the difficulty is 2d6+2d10 instead of 6d6.


Just setting difficulties migth prove easier.

Post by James O'Rance
A maximum difficulty of 10 should work. You could possibly go to 11
with the 6/10 +1 option, but these tasks will be so difficult that I
don't see the point. If you have some kind of botching mechanic (eg,
1s cancel out successes) then succeeding at a difficulty 11 task would
be very rare indeed.

I'm thinking that instead of just having binary success, how much you
beat the target/opposed roll by would represent degree of success. So
while I wouldn't want many difficulties over 10, a roll of 11 or higher
would still probably represent greater success.

Post by James O'Rance
However, if the average difficulty is set at 7 or higher, then it
becomes impossible for the unskilled character to accomplish (without
the 6/10 +1 option).

I'm divided as to whether this is a bug or a feature. I like the idea
that a 7 or 8 difficulty is something that isn't _too_ difficult if you
have decent skill, but is still really hard even with very high stats.
Those difficulties could represent tasks that normally require some
training. The only thing I don't like is that there would still be no
way to have something that was almost impossible when untrained, but easy
with even a little training - but then, not many systems do that.

Post by James O'Rance

Post by Robert Scott Clark
I was also trying to think of some different benefit to rolling the
6/10, that could be taken in addition to/instead of the +1. My gut
reaction was some form of minor critical success for a 6 and
protection from fumble/mishap for a 10 - to represent fluke success
for the prodogy and the experience of the high skill individual, but
I coundn't figure out a good system for fumbles/mishaps that didn't
require a lot more rolling or produce odd results (ie. having each 1
roll represent a minor mishap has the silly effect of making more
dice produce more mishaps.)

Hmmm. Maybe a variant on botch rules from the Revised Storyteller
games?
If you roll a 1 on 1d6, it cancels a success.
If all of your successes are cancelled, the action fails (having more
1s than successes has no additional effect).
If you roll a 1 on 1d6 and have no successes to cancel, you get a
mishap.
Rolling 1 on 1d10 has no special effect.
This means that high skill protects you somewhat from mishaps; high
dice pools don't produce a greater chance of mishaps than low dice
pools; and no additional dice rolling is required.

There seems to be a little miscommunication here, as there would be no
way to have more than 1 success with that system. All the dice
collectively produce 1 number.

Maybe a 1 rolled on a d6 could be a -1, but then having a high stat isn't
really useful for someone with a high skill, as the chances of +1 on 6
and -1 on 1 would basically cancel each other out.

[Nis Haller Baggesen]

Post by Robert Scott Clark

Post by James O'Rance

Post by Robert Scott Clark
Basically, I wanted a simple WWesque stat+skill pool that
differentiated between stat and skill in some way. A roll is made
with a number of d6s equal to the stat and a number of d10s equal to
the skill. The result is the highest number rolled on any die
increased by 1 for every extra 6 rolled on a d6 or 10 rolled on a d10
- compare that to the target number and voila.

That's quite interesting. What range of difficulty numbers are you
looking at? What's the average difficulty?

Still toying around, but I was thinking of always using opposed rolls -
easiest task 1d6, near impossible task 6d6+8d10. Average? 3d6 maybe?
I like the idea because it could differentiate between a failure because
of a faulty attempt (low player roll) or because of a complication (high
opposed roll)
(I just got a flash that maybe high skill is better at countering
complications than high stat is, so maybe a 10 could remove an opponent
die instead of adding the +1 to the result. I'll have to think a little
more about the ramifications of that - especially what that does to the
odds.)

Even disregarding the problem of the odds I think this would be a bad idea.
First of all - Do you remove dice before or after the roll. If you do it
before, then who gets to roll first in a 'real' opposed test. If you do it
after, what happens if I remove that extra 10 which allowed yo to remove one
of my dice ?

Post by Robert Scott Clark
I'm not set on that yet, as I can't really decide what the difference in
dice size would mean for an opposed roll vs anything other than a
character. I mean, if there's going to be different size dice to
represent the difficulty of defusing a bomb, there should be some
rationale for saying the difficulty is 2d6+2d10 instead of 6d6.
Just setting difficulties migth prove easier.

I think it will.

[Jeff Heikkinen]

Robert Scott Clark, worshipped by llamas the world over, wrote...

Post by Nis Haller Baggesen
Even disregarding the problem of the odds I think this would be a bad
idea. First of all - Do you remove dice before or after the roll.

Given that it would be representing a skilled person's ability to react
to and/or avoid a complication, I would say after.

I would tend to think that that ability would tend to produce results
that were reliably a little better than those of someone without it;
your system unreliably produces ones that are spectacularly better.
Different intuitions, I guess, but are you sure that's what you want?

[Robert Scott Clark]

Post by Jeff Heikkinen
Robert Scott Clark, worshipped by llamas the world over, wrote...

Post by Nis Haller Baggesen
Even disregarding the problem of the odds I think this would be a bad
idea. First of all - Do you remove dice before or after the roll.

Given that it would be representing a skilled person's ability to react
to and/or avoid a complication, I would say after.

I would tend to think that that ability would tend to produce results
that were reliably a little better than those of someone without it;
your system unreliably produces ones that are spectacularly better.
Different intuitions, I guess, but are you sure that's what you want?

That's why I need to look at the odds.

[James O'Rance]

Post by Robert Scott Clark
Still toying around, but I was thinking of always using opposed rolls -
easiest task 1d6, near impossible task 6d6+8d10. Average? 3d6 maybe?
I like the idea because it could differentiate between a failure because
of a faulty attempt (low player roll) or because of a complication (high
opposed roll)

You've already described one of the main problems that I see with this
- how to make the "difficulty dice" concrete in situations -other-
than opposed characters.

Another drawback to this is that you'll introduce a new roll into
every action. This will slow the pace of the game somewhat.

An existing example is D&D magic versus psionics. If your wizard casts
a spell, the difficulty to resist the spell is 10 + spell level +
attribute modifier - a set number. If your psion uses a power, the
difficulty to resist is based on 1d20 + power level + attribute
modifier.

Using psionic powers makes the game a bit slower.

Post by Robert Scott Clark
(I just got a flash that maybe high skill is better at countering
complications than high stat is, so maybe a 10 could remove an opponent
die instead of adding the +1 to the result. I'll have to think a little
more about the ramifications of that - especially what that does to the
odds.)

Or remove "mishap dice" (whatever mechanic those might take).

Post by Robert Scott Clark
I'm thinking that instead of just having binary success, how much you
beat the target/opposed roll by would represent degree of success. So
while I wouldn't want many difficulties over 10, a roll of 11 or higher
would still probably represent greater success.

I wasn't assuming binary success/failure.

You stated that you wanted "a simple WWesque stat+skill pool", so I
assumed that levels of success would be determined in a similar way to
the Storyteller system (counting the number of dice that roll equial
to or higher than the target difficulty number).

Godlike has an interesting variation on this system.

Post by Robert Scott Clark
I'm divided as to whether this is a bug or a feature. I like the idea
that a 7 or 8 difficulty is something that isn't _too_ difficult if you
have decent skill, but is still really hard even with very high stats.
Those difficulties could represent tasks that normally require some
training.

Fair enough.

Post by Robert Scott Clark
The only thing I don't like is that there would still be no
way to have something that was almost impossible when untrained, but easy
with even a little training - but then, not many systems do that.

Difficulty 6 (using a ST-like dice pool) would cover this. The chance
that a moderately skilled person would get at least one 6 (on d10s) is
reasonable, whereas an unskilled person would have a much smaller
chance.

Post by Robert Scott Clark
There seems to be a little miscommunication here, as there would be no
way to have more than 1 success with that system. All the dice
collectively produce 1 number.

Definitely a miscommunication. I didn't realise that you wanted to get
a single, cumulative result from the dice pool. I assumed that each
die would be counted as a success or failure (or mishap) individually,
with the total number of successes giving you a qualitative success.

The Whispering Vault has a dice pool mechanic that produces a single
result number. Perhaps taking some inspiration from that system would
help.

--
james o'rance
http://www.geocities.com/dragon-dreamer

[James O'Rance]

Post by James O'Rance
The Whispering Vault has a dice pool mechanic that produces a single
result number. Perhaps taking some inspiration from that system would
help.

Okay, I've thought about this some more. It seems to me that you could
create a dice pool mechanic by borrowing bits of TWV and Godlike.

Opinions?

***

All actions are based on an Attribute + Skill combination. Attributes
and Skills are both rated from 1 to 5 (in this example; the range
could be greater).

Whenever you attempt an action, the GM calls from an appropriate
Attribute + Skill and (secretly?) determines the target difficulty
number. You roll 1d6 for each point of an Attribute, and 1d10 for each
point of a Skill. For example, if you had Intelligence 3 + Research 4,
you would roll 3d6+4d10.

You choose any number of matching dice to determine the success of
your roll. The total of the numbers on these dice is your result. If
the result is equal to or higher than the difficulty number, then the
action is successful.

The quality of success or failure is determined by the number of
matched dice in a roll. For example, you might choose a matched set of
five 3s. If the target difficulty is 15 or less, you get an astounding
success. However, if the target difficulty is greater than 15, then
you fail horribly.

For example, against a difficulty of 20 or less, a roll of four 5s is
a better success than a roll of five 4s.

If you roll a "10" on your action attempt, you can choose to change
that die to a number of your choice. For example, if you had rolled
three 5s and one 10, you could change the 10 to a 5. This makes your
action more likely to meet the target difficulty number, and gives you
a higher quality of success if you do.

***

This gives greater weight to skills, which can both roll higher and
get consistently higher levels of success (by forming more matched
sets). However, the untrained character still has a chance to succeed;
it will just be much less likely.

I like the number of dice in a matched set giving the quality of
success or failure. It makes for some interesting choices - if I'm not
sure of the difficulty, do I take a large set that might fail
horribly, or minimise the risk by taking a set of 1 die?

Nothing special happens when you roll a 6 on 1d6, or if you roll a 1
on any die.

Actual numbers for difficulties would probably have to be determined
by someone with more mathematical aptitude than me.

I'm keen to see what people think of this. If it's workable, I may
even use it myself. I have an idea that it would suit.

--
james o'rance
http://www.geocities.com/dragon-dreamer

[James O'Rance]

But an untrained character's successes will
be much less likely to be marginal; if the DC is 20, an
untrained character can never succeed with a match
of width less than four, while a trained character could
succeed with a width of two.

True. Again, the question is whether this is a quality or a bug.

This astounding success can only be achieved by an unskilled character
if they have an Attribute of 4+, which on a 1-5 scale is very high.

A target number of 20 is something that I picked to demonstrate the
importance of wide rolls; I suspect that it's actually quite high.
Generally the target number would be lower. An "easy" task might have
a target number of 8.

So on such difficult tasks, an untrained character needs to be
exceptional to have any chance of success at all. He would still
likely fail, but on the rare occasions when he did succeed, it would
be quite dramatic.

--
james o'rance
http://www.geocities.com/dragon-dreamer

[Rick Pikul]

In article <***@65.82.44.187>,
***@mindspring.com says...

{Foomph...}

Post by Robert Scott Clark
I'm not set on that yet, as I can't really decide what the difference in
dice size would mean for an opposed roll vs anything other than a
character. I mean, if there's going to be different size dice to
represent the difficulty of defusing a bomb, there should be some
rationale for saying the difficulty is 2d6+2d10 instead of 6d6.

Something that comes to mind is that the 2d6+2d10 has one or more
'gotchas' that only experienced/skilled people tend to look for, while
the 6d6 is straightforward, if touchy.

[Robert Scott Clark]

First of all I would like to say that I think it looks quite neat.
However there might (Haven't done any math, just using my intuition)
be some strange effects in the model.
First of all the primary effect of having more dice is not so much to
increase you chance of performing well, but to reduce your chance of
performing poorly - So to speak. This works ok for my intuition about
how skills and practice work, but it does conflict with my intuition
about attributes, which I equate with inborn talent.

The intent was that at higher skill levels, the main function of
attributes would be the +1 bonus - meaning, as you imply, that higher
attributes means higher performance, not just less chance of poor
performance. But, yea, the 1/6 chance of that happening does throw a
damper on that somewhat. I'll come back to that in the next paragraph
though.

Also with just 2d6 (Which I would assume to be the lowest possible
number of dice for average - Leaving only 1d6 for the 'poor') you have
around 30% chance og getting a 6, while it take a lot of dice to
reliably get 2 sixes. I think this might result in attributes being
faily insignificant. This might also cause some problems with
representing things with very high attributes - Such as the strength
of an elephant. To get just an average +2 compared to the average guy,
that elephant would need 12 more dice. And +2 might not be enough, so
one might quickly end up with unhandy dicepools.

This is problematic. One possible solution is to say that the +1 can
come from 5-6 and 9-10 respectively, making the +1 happen more, and
somewhat more reliably. 4-6,8-10 would even be possible. To me this
seems more complicated for some reason, even though it isn't really, and
part of me just feels it's wrong. "Highest = something special" just
seems more natural, and even though I know that there's no reason for
that to be true, there's still a nagging feeling.

The second possible solution requires looking a little more closely at
what I was thinking of using the system for. One of the design ideas was
based somewhat on the way FengShui keeps all characters viable fighters
by making sure all their skills in fighting start somewhere around the
same score even if it's kind of silly that the archetype would fight
well. This is going for the same feel, except that it basically forces
the rolled results tighter together no matter how high the skill/stat
goes. In effect, the helpless love interest might fight off the big bad
for a few minutes until the hero gets there.

Now, conceptionally, this is pretty incompatable with elephants. Even if
I want to clump all human results together, an elephant should fall
outside that range. Some sort of scaling rules seem appropriate - maybe
just a big direct +X to whatever you roll, so an elephant would be rated
on a similar scale to humans, but on the "Bohemoth" scale along with the
general class of "big strong things". Then you might get the spunky
horse holding out just long enough in that tug-o-war against the
elephant, but the human stands no chance.

The big jump from unskilled to skilled, might also create some
difficulties when assigning difficulty numbers, espeiclaly to things
that can arguably be done quite well even with little skill, such as
climbing, jumping and other sorts of athletics.

I'm not sure I see this. I can't think of anything that an untrained
person has a decent chance of doing that it would bother me to have a
trained person succeed on reliably - excepting things that are mostly
luck.

As for botch and complication mechanics, you might use a simple linear
succes scale where rolling X lower than the target number was a
fumble, and Y over the target number is a superior success. In
addition or instead one might name one of the dice a 'wild die' and if
it came up 1 it indicated a complication, and if it came up maximum it
indicated a superior result (or vice versa) regardless of whether the
main roll was a succes or failure.

I really don't like that in most systems complications can only happen
when you fail the task, so the "wild die" is what I tend to go with.
This system is in some ways a simplified version of an older system I was
working on that relied heavily on such a "wild die" mechanic - so much so
that you were often rolling more wild dice than skill/stat dice.

[Nis Haller Baggesen]

Post by Robert Scott Clark

Also with just 2d6 (Which I would assume to be the lowest possible
number of dice for average - Leaving only 1d6 for the 'poor') you have
around 30% chance og getting a 6, while it take a lot of dice to
reliably get 2 sixes. I think this might result in attributes being
faily insignificant. This might also cause some problems with
representing things with very high attributes - Such as the strength
of an elephant. To get just an average +2 compared to the average guy,
that elephant would need 12 more dice. And +2 might not be enough, so
one might quickly end up with unhandy dicepools.

This is problematic. One possible solution is to say that the +1 can
come from 5-6 and 9-10 respectively, making the +1 happen more, and
somewhat more reliably. 4-6,8-10 would even be possible. To me this
seems more complicated for some reason, even though it isn't really, and
part of me just feels it's wrong. "Highest = something special" just
seems more natural, and even though I know that there's no reason for
that to be true, there's still a nagging feeling.

Well one possible option would be to use d4's for the stat, but that has
other problems of course.

Post by Robert Scott Clark
The second possible solution requires looking a little more closely at
what I was thinking of using the system for. One of the design ideas was
based somewhat on the way FengShui keeps all characters viable fighters
by making sure all their skills in fighting start somewhere around the
same score even if it's kind of silly that the archetype would fight
well. This is going for the same feel, except that it basically forces
the rolled results tighter together no matter how high the skill/stat
goes. In effect, the helpless love interest might fight off the big bad
for a few minutes until the hero gets there.
Now, conceptionally, this is pretty incompatable with elephants. Even if
I want to clump all human results together, an elephant should fall
outside that range. Some sort of scaling rules seem appropriate - maybe
just a big direct +X to whatever you roll, so an elephant would be rated
on a similar scale to humans, but on the "Bohemoth" scale along with the
general class of "big strong things". Then you might get the spunky
horse holding out just long enough in that tug-o-war against the
elephant, but the human stands no chance.

Of course this doesn't have to be a problem if you games normally only deal
with humans. I just mentioned it because I'm obsessive about getting my own
systems to scale nicely. And also can be a probelm in games like Shadowrun
where the scale in the player group can vary widely.

Post by Robert Scott Clark

The big jump from unskilled to skilled, might also create some
difficulties when assigning difficulty numbers, espeiclaly to things
that can arguably be done quite well even with little skill, such as
climbing, jumping and other sorts of athletics.

I'm not sure I see this. I can't think of anything that an untrained
person has a decent chance of doing that it would bother me to have a
trained person succeed on reliably - excepting things that are mostly
luck.

I don't have any probelm with the trained person succeeding more reliably,
but I would have a problem with. My problem is that the skill greatly
increases the difficulty range a person can handle and or the success rate.
This just doesn't jive with my experience from when I wen't to athletics or
now that I'm climbing.

Even a year of training did not significantly increase my maximal long jump
distance compared to when I started (Or high jump, or my best 100m dash
time). What it did do was allow me to perform at my limit much more often.
Same thing with climbing (Although I must admit I've just strted that, so
training may mean more). Training will undoubtedly make me better, but it
wont make me grow those 5 cm that will allow me to reach some different
grips. But maybe my experience is just unusual.

Post by Robert Scott Clark

As for botch and complication mechanics, you might use a simple linear
succes scale where rolling X lower than the target number was a
fumble, and Y over the target number is a superior success. In
addition or instead one might name one of the dice a 'wild die' and if
it came up 1 it indicated a complication, and if it came up maximum it
indicated a superior result (or vice versa) regardless of whether the
main roll was a succes or failure.

I really don't like that in most systems complications can only happen
when you fail the task, so the "wild die" is what I tend to go with.

I would go with the 'wild die' mechanic as well - And for much the same
reason. However I think it is important to keep it integrated into the main
roll in some way.

mvh

Nis

[Robert Scott Clark]

Post by Nis Haller Baggesen
Even a year of training did not significantly increase my maximal long
jump distance compared to when I started (Or high jump, or my best
100m dash time).

I think this is a case of different skills behaving in different ways. The
linear progression you see in most games seems wrong to me, as higher skill
should probably produce a logarithmic increase in distance. Higher skill
levels are about squeezing in fractions of an inch. The main problem is
that this sort of thing can't be modeled very well with a single skill
system that is assumed to represent everything.


But then my idea of learnign something about jumping and climbing is
figuring out how not to hurt myself.