Are Dice Pools Good, Actually?

[Shiroiken]

I think that dice pools with successes work best with custom dice that can have multiple successes, penalties, weird effect, etc (look at the new L5R RPG for an example). The advantage of counting rolled dice is in a system that uses exploding dice, where you always have a chance, no matter how small.

 

[pemerton]

Most of what I'm playing these days involves dice pools of some sort or other.

Classic Traveller is mostly rolls of 2d6. Sometimes 3d6. The addition is easy. The smaller spread of results on 2d6 makes ties more common, and I like ties (when used with Let it Ride, to force the conflict onto a different field of endeavour).

Prince Valiant officially (ie as per the book) uses coins, with heads as success and tails as failure. This is the same as rolling dice labelled 0 and 1 and adding, I guess, but in our group we mostly use standard d6s and count successes - most of use are "evens" but our most diehard Burning Wheel fan is 4+. There is nothing "weird" about it that I've noticed, nor has anyone in our group complained. Pool sizes range from 1 or 2 to 15+ when a strong knight is jousting in heavy armour on a mighty steed in defence of lover or honour..

Burning Wheel is pools of d6s, generally similar in size to Prince Valiant. Normally 4+ is required but under peculiar circumstances it can be 3+ or even 2+. I have a set of charts that lists the probabilities for a given size of pool against a given obstacle.

What I like about the Prince Valiant and Burning Wheel pools is that (1) ties are more common, and (2) there is always a chance of failure when the dice are rolled. (BW is explicitly "say 'yes' or roll the dice", and Prince Valian works on a similar ethos.) This isn't something that you get with target number systems.

The most baroque dice pool system our group plays is Marvel Heroic/Cortex+ Heroic - the dice can be of all different sizes from d4 to d12, and the default is to add 2 to get the result to determine success against a target number, and then a third die determines the effect but based purely on its size, not its result. I haven't tried to create any probability charts for this as the maths would be above my pay grade!

The only system I play where the pool is rolled and you just keep the best die is Cthulhu Dark. By default there are no target numbers, only degrees of effectiveness of outcome. But sometimes an opposed die is rolled that sets the target number for success.

I think there are challenges in a roll and add vs target number system similar to those in a single die system like D&D or RQ, namely, that if the system allows lots of modifiers then if the maths is not carefully worked out success or failure can become automatic. Classic Traveller avoids this problem (in my experience) by keeping modifiers small. Rolemaster avoids it by having a mixture of auto-fails and open-ended rolls - the latter making it a dice pool system where the number of dice is randomly determined. I don't think the issue was all that well throught through in AD&D design, and I gather it can be a problem in 3E D&D as well. In 4e D&D I only encountered it at one point in one build - the Sage of Ages epic destiny.

 

[pemerton]

The advantage of counting rolled dice is in a system that uses exploding dice, where you always have a chance, no matter how small.

You can get this in Burning Wheel if you spend a player-side resource (fate point). Prince Valiant, on the other hand, has a hard cap (full successes is +1 success, but not exploding).

But you can do "always have a chance" in an addition-based system too, like Rolemaster. In two decades of play we saw some wild results (double- and triple-open-ended) that saved the day.

 

[Bedrockgames]

That being said, there are also disadvantages:

• Most people don't know the probabilities for success for dice pools by heart; this gets tremendously worse, if you combine dice pools with variable success thresholds

I think this is both an advantage and a disadvantage. I kind of like the muddiness on the player side. Most people don't know the actual probability of their success of jumping over a ravine (they just have a general sense of their chances). I think that translates well in play. On the GM side it is more complicated because it can result in the GM not really grasping how difficult he or she just made a challenge. Again though, I think as long as the general feel of the probability is there, it is still good.

 

[schneeland]

@Bedrockgames:
That's a fair point and a good observation. My feeling is, though, that most players (including me) will not develop a solid feeling for their chances of success, if there are too many variables (pool size, target number, no. of successes), so you're better off picking only one of them.
And of course, on the GM side, actual numbers are more important (as you already noted).

 

[lordabdul]

It's funny to me that people talk about not having a clear idea of one's chances of success when, in my experience, the vast majority of D&D DMs I've seen in action would not tell the players what DC they're shooting for.

If you worry about making chances of success clear to the players, there's one and only one resolution system that's totally unbeatable: percentile systems. Roll D100 under your stat. Your percentage of success is right there on the character sheet.

Anyway, I like dice pools. They're fun, and you can do funky things with them.

1. You can easily get a multi-level success system, because it's easier to count the dice than to do a subtraction to figure out a margin of success/failure.
   1. This might give you "bumps" in success levels, say from moderate success to complete success to critical success (although you need a GM that can translate that to a satisfying narrative consequence)
   2. Most commonly it gives you extra stuff like extra actions, which is very easy to grok and GM... for instance 7th Sea does that, each success is an extra action you can perform in the action scene.
2. You can have weird resolution systems like trying to match dice for instance. Godlike/Wild Talents does that, and that gives you 2 measures out of the dice roll (the "width" and "height" of the match, like, say, 2 tens vs. 3 fours). I love to see designers explore other resolution systems like these, it shows that innovation isn't totally dead when it comes to rolling dice.

Dice pools have limitations though. I find that they lack granularity, and don't scale very well up and down. As such, they tend to be effective (IMHO) mostly in narrative-inclined games where it's less about wargaming and more about storytelling -- like, famously, the Storyteller system or the 7th Sea system. I don't think these games would have the appeal and "style" that they have without a dice pools.

 

[Wulffolk]

I agree with many of @lordabdul points.

I really love the nuance possible with dice pools. I love that you can modify so many aspects of a dice pool, such as: type of dice, number of dice, difficulty target, number of successes needed, immediate action vs extended actions, etc.

Some of this can be done with a d20, but I prefer the consistency of a dice pool's bell curve over the wild variability of a d20.

 

[Bedrockgames]

@Bedrockgames:
That's a fair point and a good observation. My feeling is, though, that most players (including me) will not develop a solid feeling for their chances of success, if there are too many variables (pool size, target number, no. of successes), so you're better off picking only one of them.
And of course, on the GM side, actual numbers are more important (as you already noted).

They don't need a solid feeling in my opinion, just a general sense. There is a difference. I like the muddiness because the players are not running numbers in their head, they are going off the GM description and just know something like (I have three 3d10 in this skill so I am pretty good). Real people don't run numbers in their heads when deciding to dart in front of a car, or leap over a ravine. Having a shifting TN makes it easier for the GM to set the difficulty appropriately.

 

[schneeland]

They don't need a solid feeling in my opinion, just a general sense.
...

I guess it comes down to preference here. At least half of my players tend to dislike systems where they don't have one (it's less of a problem for the people who lean towards the narrative side).

 

[Ovinomancer]

They don't need a solid feeling in my opinion, just a general sense. There is a difference. I like the muddiness because the players are not running numbers in their head, they are going off the GM description and just know something like (I have three 3d10 in this skill so I am pretty good). Real people don't run numbers in their heads when deciding to dart in front of a car, or leap over a ravine. Having a shifting TN makes it easier for the GM to set the difficulty appropriately.

Sure they do. I mean, there aren't TNs in real life, but there's a constant evaluation of risk versus capability. It's fundamental to existence. Not having a way to gauge risk in games is bad. Your example of 3d10 being pretty good is exactly what's being discussed -- this knowledge isn't helpful at all if you don't know the level of the challenge. If it's a really hard challenge, but you don't know it, then your faith in 3d10 will lead you to make a choice that's not desirable. TNs aren't magic, they're how you, as GM, can communicate difficulty to the player based on what their PC's perceive.

Now, I prefer less number centric descriptions, so when I run 5e, for instance, I'll say things like, "It looks like an easy climb," or, "that sounds like a really hard thing to do, are you sure?" My players know that easy is 10 (or nearabouts) and really hard is a 25 because I uses the easy/moderate/hard/very hard/near impossible scale for 10/15/20/25/30.