I heard it mentioned on the wizards D&D boards that indie rpgs look favorably upon dice pools instead of straight up percentage modifiers. Can someone tell me why this is? The only thing I can guess is that dice pool rolling is more tactice (storyteller vs d20), but I'm sure there's much better reasons than the amatuerish one I just gave.

I started down this problem because I (perhaps mistakenly) convinced myself that things would be easier if one can just throw one, or handfuls, of d20s to resolve anything.

1-20 translates to percentages very well. Each pip is a very manageable +/- 5%. Need a tighter bell curve? 2d10.

It began with frustration playing Deadlands non-d20. As written and unmodified, the play mechanics (poker hands, drawing cards for stats, dice pools for resolution) is great tactile fun but broken beyond belief. Using poker hands to resolve casting magic is like having the huckster players juggle live grenades blindfolded...awesome results at random times but nothing works when needed. I was retardedly trying to stay true to the official probabilities, and had to resort to writing programs to do so. Even storyteller and the old Heroquest boardgame lineage was easier to convert to pure percentages. Deadlands is, of course, one of the few extreme anomalies in mainstream RPG lines.

I can see the right-brain benefits of adding and taking away dice from a pool for Storyteller, but if it's actually easier to balance probabilities, or (for instance) appeals to 90% of narrative players or something, I've love to know all those reasons as well.

In your actual play account, what was happening in play when nothing worked when it was needed (in regards to the deadlands hand of cards)? It sounds important :)

On the main question, its a bit hard to answer - I'm not aware of any favour around here with dice pools. I am aware that the technical uses of dice pools are known around here. I'm not too sure myself, but I think they help create a very hard average - they eliminate freaky dice results in a way, as the results are quite predictable. In fact one poster here thinks the D20 system is actually a dice pool system, in a way, since you'll be rolling attack after attack, it's basically a pool of D20, all creating a hard average.

Hello,

Thanks for re-posting here. I'd like to focus on your examples from real play.

First, we ought to define our terms so I can make sure I'm really addressing your concerns. You've probably realized that people use the term uncritically all the time.

As far as I can tell, you are talking about a variable vs. fixed number of dice as your definition of dice pool. Check me on this: in the old game The Fantasy Trip, one rolled 3d6 and totaled them, hoping to reach a value equal to or under one's attribute of interest at the moment. Sometimes, as when an opponent parried or dodged, one rolled 4d6 instead. By your definition, this isn't a dice pool, right? Because the number of dice is fixed, or almost fixed, for all rolls. Or is that a dice pool to you? Some people use the term to indicate any resolution method which uses more than one die at a time. Is your definition that narrow?

Another example: The Shadow of Yesterday uses Fudge dice, reading two at a time (range from two pluses to two minuses). Although you may roll more than two, you only ever read two. Is this a dice pool to you?

Second, I've played a fair amount of Deadlands myself. I don't mind saying that its dice system is not elegant. However, its problems aren't specific to its using pools (a varying number of dice), but rather to its basic difficulty of having too many variables in the mix: number of dice, size of dice (number of sides), and target number. I am also not surprised to see you citing the so-called Storyteller system, because it, too, is mathematically irritating - one must assess a roll in terms of target number, and then assess it in terms of someone else's roll. The net effect (beyond the time wasted in arriving at the results) is to minimize degrees of success, such that most rolls end up not doing much of anything.

My thought on what you've written, then, is that you are mainly critiquing bad design rather than dice pool design. I think you may be confounding one of the features of a number of systems with problems with the problems.

Third, are you really asking whether independent RPGs favor dice pools? Is this actually of interest? It threatens to become a meaningless topic - one man's "favors" is another's "tends" is another's "historical detail" is another's "doesn't matter." The basic observation is yes, many independent games feature varying numbers of dice being rolled, although many do not. Of my four games, Sorcerer and It Was a Mutual Decision both involve varying numbers of dice per roll, Elfs does not (one always rolls 3d10), and Trollbabe either does or doesn't depending on how you want to interpret it.

I think the independent games, at least to some extent more than in the non-independent ones, display more elegant design. So if you were to look at some of the designs with varying numbers of dice, you'll also find that they do not include varying target numbers as well. In other words, the number of dice being rolled is easily processes by the mind as more or less difficulty and the whole "can't assess the outcome" issue goes away.

Best, Ron

Tactility is one explanation, but in my designs I find that the more dice a player rolls, the more information can be teased out of each roll.  This went back as far as UA where I didn't want to have initiative, to hit, and damage rolls separate.  In UA 1ed, initiative is mangled, but the same roll that resolves whether you hit or missed also gives damage.  (In the second edition, initiative got splinted at least.)  That was with a basic two die % roll.

ORE (used to drive GODLIKE, Wild Talents and REIGN) is a dice pool system (with optional target numbers AND penalties, thank you very much) uses a dice pool, but like the wily Inuit using every part of the walrus, it (potentially) uses every die in the pool, even those that don't contribute to success.

F'rinstance, suppose in REIGN I'm going to cast a spell and some jamoke is going to try and counterspell it.  We declare our actions and we roll our pools.  Our dice pools tell us the following.

-Did my spell work at all?
-Did his attempt to ruin it function?
-Which happened first?
-If my spell worked, is his attempt to foil it sufficient to shut it down?
-How much damage did the spell do?
-Where did the damage hit?  (Some spells have damage from Waste Dice, which means that 'failures' do damage in a different place.)

I can't see a way to pull that out of a d20 roll.

so-called Storyteller system, because it, too, is mathematically irritating - one must assess a roll in terms of target number, and then assess it in terms of someone else's roll.

So if you were to look at some of the designs with varying numbers of dice, you'll also find that they do not include varying target numbers as well.

Thanks for the feedback. Your insights should make it fairly simple to incorporate, or throw out, dice pools when needed. The challenge I started out with is how to make Deadlands fun, and the answer I teased out from you guys' experience is to make the resolution more forgiving.

On the storyteller system--the new storytelling system is definitely easier to manage for my players. It's extremely easy to explain to beginning players, and is essentially no harder than Heroquest (boardgame). I can explain that 2 dice gives you a ~50-50 change at succeeding, so even the most art-inclined math-averse player in my group is very comfortable with that. Storytelling->% conversions are much less troublesome...it's even possible to do storytelling->tohit/todam if I care to print out probabilies with some program.