Following up on my post about The Riddle of Steel RPG, here’s a broad question for readers at large, touching on many games: What, exactly, is the appeal of the “dice pool” method of generating outcomes?

I know it has a two-dimensional aspect to it, in that you can modify checks in two ways: you can both modify the “target number” that determines whether a die counts as a success, and you can modify the number of dice rolled. Thatย *sounds*ย like it offers something richer than the classic one-dimensional, dice-roll-plus-summed-modifiers method, and I don’t yet see anythingย *wrong*ย with the dice pool method. But I’m curious: Do dice-pool systems establish a clear, easily-followed master rule for what factors modify the target number and what factors modify the number of dice? Or are GMs often left wondering why the system specifies a modified target number forย *this*ย factor, but specifies additional dice forย *that*ย factor? (And similarly, whether to modify target number or change the number of dice when someย *new* factor comes up that hasn’t been specified by rules?)

More importantly, does the two-dimensional dice pool system generate notably more *meaningful* results than the classic method? If so, in what way?ย Or, if I may ask, do dice pools add a bit of complexity to rolling without returning some significant added benefit?

If it sounds like I’m a tad suspicious of dice pool mechanics, well, I’ll only admit that the method somehow feels more fuzzily abstract to me. But that’s doubtless the biased voice of my long experience with classic RPG dicing. I’d be interested in the wise words of gamers more versed in both styles. What say ye?

*Recommended side dishes*

## 18 Comments

## Blizzard

there is something to be said for having a handful of dice to express an attack or defence. It makes for a more tactile experience (our group used nice wooden salad bowls to hold our dice and a larger one to throw ours into).

Of course, Riddle of Steel has a deadly combat system so the number of combat encounters is usually less than a hit point based rpg. End result is that combat takes the same amount of time (in our group about 20 minutes, plus or minus), but fewer ’rounds’ happen than would in DnD. The exact results of each group of roles then becomes more important (thus players are more interested in them).

From a game design point of view, multiple dice create a more distrubted bell curve of possibilities. A single roll of a 20 sided dice is very unpredictable. Others play DnD with 2d10 or 3d6 to give a different flavour. As for determining whether to modify the target number, or the dice pool, that’s a matter of taste and experience. A lot of systems will just simply the process and alway use a specific target number, then only apply pluses/minuses to the rolls (if there is only one die).

## Perrin Rynning

All RPGs have to balance several different factors. One is letting the players feel like their characters have at least a calculable chance of accomplishing tasks at least partially outside of their (the characters’ AND the players’) control. Another is not making every task too easy to bother with. And a third is keeping things from being totally predictable; it’s just as as undesirable to know with certainty that your character will succeed as it is to know that he/she/it will fail.

Dice mods make it easier to calculate the odds of success, because you have a firm grasp of the base probabilities. The mods themselves can be known or hidden (‘cos the tricksy GM didn’t mention that the dungeon’s oxygen levels are rising, or something equally weird) but you know what your chances are to hit or exceed the success threshold on a single roll, and it’s easier to adjust your calculations from there. (Bonus: does anyone reading this column remember “column shifts”? They used to be all the rage, and now no one uses them. Care to guess why?)

On the other hand, dice pools make probability calculations more complex, because no one die roll result equals automatic success or failure. I might suggest visualizing it as playing at six craps tables simultaneously, with the wagers on each table modifying all of the others. Thus, it becomes much more difficult to tell whether or not a character will succeed at any task calling for dice rolls.

I believe that dice pools do a better job of making it more difficult for the player to predict the outcome of any contest, but that dice plus modifiers make the game flow more smoothly. Ultimately, it remains a matter of player choice which tools to use in the service of having fun.

## kenclary

As Riddle of Steel uses the term “dice pools”, it’s not just using a variable amount of dice in a roll. Like Shadowrun before it (iirc), “pool” means you have a number of dice each round that you get to allocate as bonuses to the various things you want to do. In RoS, you spend out of this pool to activate actions, and spend further to increase your chances with them. This is similar to some “action points” methods in some systems.

I think one appeal of this is that adding N dice to a roll is less predictable than adding N to a result. The “extra” chaos makes it feel less precise, which probably helps some players suspend disbelief.

And rolling giant piles of dice is fun, sometimes ๐

## tbone

Looking at three comments on how dice pools differ from dice + mods, I see one point in common:

feel. It just feels different to hold a nice big handful of dice, and I can fully understand the tactile appeal of combat factors affecting the dice in your hand, not just an abstract number in your head. A couple of the comments also mention that shifting pool size and target numbers create probabilities more difficult to compute than an unvarying number of dice + mods. I can see how that might help keep number-crunching players more focused on play than on probabilities.kenclary also notes that in TRoS, the dice pool effectively doubles as an action point scheme. Good point.

Blizzard notes that more dice means a change in the probability distribution, something that’s true and needs to be considered by a designer. Although in the case of dice pool mechanics, I’m thinking that it’s an unintended effect, or possibly even a very minor negative side-effect. That is, I assume designers fully expect and want a pool of 10 dice (lots of bonuses!) to on average yield 5 times the successes of a pool of 2 dice (lots of penalties!), but I don’t assume that the designers

necessarilywant the 10-die pool to yield a very tightly-clustered number of successes that rarely stray far from average, and the 2-die pool to deviate more freely from its average. (In the same way, I don’t think RPG designers are looking for the effect that typical 1d sword damage is just as likely to deal max dam, min dam, or any dam in between, while a powerful 6d weapon will strongly cluster around average dam. That’s just the way things work out.) Then again, maybe some designersdohave reasons for liking the way a changing dice pool affects both average successes and deviation from average. I don’t know; it’s pretty harmless either way.Anyway, getting to my key reaction to the comments, it’s this: I was actually expecting that there was some more purely mechanical benefit to the dice pool system. That is, I had some expectation that compared to the conventional dice + mods method, which yields both a) success/failure result; and b) degree of success/failure with a single roll, the dice pool method with its added dimension might yield a) success/failure result, b) degree of success/failure, and c)…

somethingelse, which I’ve been overlooking. Maybe that’s not the case?Well, even if there isn’t an added benefit to dice pools in terms of crunchy useful results returned, or other ability to do tricks that dice + mods can’t, the points about fun and feel are well-taken. Lots of dice do make a nice sound. (I was thinking that I’ve never played a dice pool game, but some HERO mechanics are at least partially dice pool-ish, and where superheroes are concerned, those big handfuls of plastic do feel “right”. : )

Thanks to all for the comments and enlightenment!

(@Perrin: “Column shifts”? Actually, I do vaguely recall the mechanism, now that you mention it, though I can’t come up with specific games I might have played using it, or even exactly what’s involved. Any detail you could add to refresh my memory?)

## DBV

Hi there, I just read your post, and as others have mentioned, holding more dice feels more fun than a single die. That’s a big part of the appeal. In Shadowrun, when you hear 10 dice hit the table, you know that something serious is happening, and I feel that little bit helps enrich the presentation for players. That aside, dice pools derive their dual-dimensionality from the fact that you can modify target numbers (TN) and the size of the dice pools. This divide is also present in dice + adds, however. In Dice pools, the number of dice rolled generally reflect the entity’s skill or resources, while the TN references the difficulty of the action. Climbing a brick wall is TN 5, while climbing a sheet of glass is TN 8, for example. However, an invalid might roll one die, while a spider-being might roll 8, to generate his successes. This is an example of having higher skills or resources. In dice + adds style games, such as D20 system, both of these aspectrs are present. I have my clmb modifier, showing that I am an extremely proficient climber (+8) and on the other side, representing difficulty, I have the DC (the target number). If I become better at climbing, I get a +2 to my roll. If it is raining, there is a +2 to the DC. Both factors are modified. The only difference is that in a dice + adds system, the sum of the modifiers is generally applied as a single operation. Numerically, there isn’t a difference between a +8 vs. DC 18 and a +6 vs. DC 16, but those two represent different situations.

## tbone

I’ll take that as another vote for the fun factor of big dice rolls. Most of my gaming is in systems where those big handfuls come into play only for damage rolls. Fun to dish out, not fun to receive. : ) I can understand the appeal of systems using big, loud rolls in more situations.

Your next comment matches my understanding: in dice pools, the number of dice generally represents skill/resources, and the target number the task difficulty. You’re correct, too, that dice + adds mechanics incorporate those same two factors, though along a single axis, not two axes.

That’s where I see a potential problem with dice pools. With pools, opposing factors can cancel each other out, IF they’re on the same axis; that is, a factor that adds two dice gets cancelled out by a factor that subtracts two dice. No problem. But now imagine this: Some major difficulty factor adds (say) 6 to the target number, raising it from the game’s default of 5 to a new target number of 11. No problem, you think, because your PC is a Special Forces type whose awesome skills add way more than 6 added dice to his pool… but, wait, that’s still no good. You could have superhuman abilities adding

dozensof dice to your pool, yet using d10 dice you can’t get a single success against TN 11. Oops…This causes zero problem in a dice + adds game, where a -6 difficulty penalty is fully offset, or better, by bonuses totaling 6 or more. But how about with dice pools? The above problem must have cropped up during the first session of the first dice pool game ever, so I assume it’s long-fixed; can anyone fill me in on the solution?

## kenclary

The solution I recall Shadowrun using: when a d6 comes up 6, reroll it and add the new result. This can, in theory, keep looping as long as the same die keeps coming up as its highest number. So when your TN is above 6, and you’re rolling 20 dice, you immediately discard anything that wasn’t a 6, and reroll what’s left…

This, of course, makes the probability curve complicated (not to mention the way that TN 6 and TN 7 are essentially the same). But some people like it that way.

## tbone

So, applying that rule to my example of d10s, the Special Forces character with his big handful of dice will likely toss a few 10s, which will each receive the addition of another d10. So he may have a few dice resulting in 12 or 18 or whatever, letting him score successes against that TN of 11.

Fair enough; it gets around the problem. I have to say, though, in this area dice pools seem less elegant than dice + adds, which have no problem smoothly balancing any number of bonuses and penalties.

Hmm, here’s a random thought that just occurred. For gamers who like that nice “big handful of dice” aspect of dice pools but are playing dice + adds games, there’s an obvious way to crank up the dice: Convert big net bonuses to dice. So if the game has you making its universal 3d roll but you have a +8 bonus, roll 5d+1 instead.

Works for “high roll is good” systems, anyway.

## Esteemed Visitor

The main propblem of dice + adds in your example is that that special forces character that has +10 to a skill check never, ever, ever fails. Now if he were rolling a group of dice instead, he’s still going to succeed most of the time, but once in a while, he won’t.

Solid static bonuses have their own share of issues…

## tbone

True, the +10 bonus โ absent any mitigating mechanisms โ means the special ops character never fails. I don’t think that’s necessarily a bad thing! There are tasks at which the +10 character

shouldn’tever fail, barring factors that would whittle away at the bonus (e.g., eventual fatigure, if the guy is just making that roll again and again and again), or new circumstances that would dramatically change the bonus or the very nature of the roll.I yammer a bit on a related topic on this page, using the classic “unbeatable defenses” problem as an example. In short, I argue that, in some proposed “unbeatable defenses” scenarios, the defender

shouldnail every defense, parry after parry after parry, until something changes.Still, your point’s well taken. Normally, we assume that some sort of random factors

arealways happening around the edges, and that these should always make failure a possibility for Mr. Perfect +10. It would arguably be unsatisfying if the mechanics didn’t allow for this.But that’s where dice + adds systems typically do allow for the otherwise statistically impossible, through critical hits/misses. Whether it’s a crit system like GURPS’ (which essentially throws out all stats and mods on certain rolls), or some sort of exploding dice system (which still leaves the base stats and mods relevant, but squeezes in

somechance of the unexpected), there’s always some way to avoid static outcomes even with dice + adds.(Even if tossing big handfuls of dice still sounds more fun sometimes… : )

## Esteemed Visitor

Exploding dice solve that problem easily. And believe me, rolling a 10 on a d10 or a 6 on a d6 is fairly common (1 in 10/1 in 6).

## tussock

Never have seen a game that produced anything like a sensible range of target numbers, dice pool size, or required successes.

The math is genuinely unintuitive, and the probabilities are all extremely jumpy when the target numbers move off 50% per die, especially with large dice pools.

Then you get into designers losing themselves in the possible modifiers, without calculating how severely one can effect the others. Number of dice, target number, number of successes, ratio of success to failure, degrees of success, and so on. You end up with games were you’re more likely to accidentally kill yourself by being more skilled.

For anyone playing one: get yourself 4-6 dice to lower the chance of total failure below where you’ll notice it, then do everything you can to ease the target numbers. -1 TN is almost always worth more than another half dozen dice, and you’ll never have more dice than -2 TN is worth on d10’s.

If there’s degrees of success, and you want rare special success results, take a couple more dice. If there’s a ratio of success to failure, make sure you have an odd number of dice, and focus on minimising that target number at all costs. Never roll anything in play against high target numbers, take action to lower them first, range, lighting, magic, or whatever.

Really, it’s all just a clunky way of making very small modifiers have huge effects on your success at tasks, in ways that most people won’t understand.

## tbone

Thanks for the additional insights. Interesting stuff. I’ll take your comments as additional support for my personal disinterest in using dice pools for my own homemade efforts. I do agree with the appeal others have noted, mainly the joy of rolling big handfuls of dice, but in the end I like the clarity of a set number of dice with a single net modifier. It’s easy to understand the probabilities, and there’s no dithering over whether a factor should modify TN or number of dice or both.

## tussock

Something else I’m reminded of, many rolls of a single die also start to look like dice pools when there’s little opportunity for change in the target numbers throughout.

Any game that makes you roll 10+ checks to get a result becomes very tightly bound to some specific target number that will succeed almost all the time, while 1 step harder will fail almost all the time. DnD4 suffers from this in combats even with all the possible modifiers (because the mods are effectively constant at +2 to you or -2 to opponent and are cancelling out), and their “skill challenge” rules are a disaster by being more obvious about the results and having more variation between character’s target numbers.

Games really need to let the players see they’re losing a contest, and give them a way to change that relative to the unfolding situation (so you don’t just throw your obvious winning option first round).

## the Vicar

Not meaning to nitpick, but it’s a Triple-Dimensional system, in that there are 3 ways to influence the odds of success of any given roll of the dice: 1) Modifying the amount of dice rolled; 2) Modifying the Target Number that each die needs to achieve in order to be considered a “hit”; and 3) Modifying how many “hits” are required in order to make a dice roll a success.

I’ve played Shadowrun, which uses a Dice Pool system, and I have nothing but respect for their system. It’s their campaign world that I don’t care for.

And as others have said, tossing a bunch of dice is just better than tossing one die. I’ll stake my odds of success on 2d6 as opposed to 1d12 any day of the week, even if it is a lot harder with 2d6 to roll a 12 (1 out of 36 instead of 1 out of 12).

## tbone

No worry; I never equate clarifications and added info as nit-picking. All facts are good โ

I get what you’re saying about three “dimensions” to a dice pool; thanks for the input. What I’m wondering, if you or anyone can chip in, is this: How would all three dimensions be used in an example of play?

Say a PC is attempting a diffcult task. Her skills or other abilities aiding the attempt would typically, as I understand it, add to 1) number of dice rolled. The difficulty of the task would modify 2) the Target Number needed for each die to become a “hit”. What, then, would determine 3) the number of “hits” required for success? Or, perhaps, would you typically not determine 3) in advance, but rather use number of “hits” afterward to determine degree of success (as in, 1 “hit” a weak success, 3 “hits” a strong success, etc.)?

## Esteemed Visitor

Not to nitpick – and I know this is 6 years late – but it’s actually Quad-dimensional. You can also control the dice that are “kept”, as in Legends of the Five Rings.

## tbone

You’re right, that does add another dimension to what’s possible. And maybe a really fun one. I don’t know; it didn’t come to mind for me, as I’ve never played

Legendsor any game using the mechanic!(Well, there is

Zombie Dice, which sort of has a “keep these dice” mechanic for piling up your store of brains, but I don’t think that’s quite the same thing. I’ll have to look into how kept dice work inLegends.)