The Forge Archives

I was thinking about bringing up the issue of die mechanic transparency. I'm talking about "how clear is it what my chances are when I roll the dice"

I'm very much in favour of the precalculated roll x or higher to succeed. (The roll under x should be equally clear). Roll die + target number to beat a number can be reduced back to the previous stuff but it's not as clear, whereas dice pool techniques have the drawback of becoming rather unclear even though the general chance of success (do I have better chance of succeeding than him) might be clear in some cases.

So to summarize the types of rolls I can think of (let's ignore the issue of layering for now) are:

a) Roll x or higher to succeed.
- Example: roll 4 or higher on a D6 to succeed. (chance of success 2 in 6)
- Used by: many boardgames

a-2) Roll above x to succeed.
- Example: roll above 14 on a D20 to succeed (chance of success 6 in 20)

b) Roll under or equal to x to succeed.
- Example roll 45 or lower on a D100 to succeed. (chance of success 45 in 100)
- Used by: BRP style percentile systems

b-2) Roll under x to succeed.
- Example: roll under 4 on a D10 to succeed (chance of success 3 in 10)

c) Add x dice of type y, success if equal or higher than target number z.
- Example: roll 3D6 with result 13 or higher to succeed

c-2) Add x dice of type y, success if equal or higher than target number z.
- Example: roll 3D6 with result higher than 13 to succeed

(from now on I'm gonna skip the alternate versions, you're getting the point anyway)

d) Add x dice of type y, success if equal or lower than target number z.
- Example: roll 2D6 with result 5 or lower to succeed

e) Add x dice of different type together success if equal or more than target number y.
- Example: roll 1D6+1D20 with result 11 or higher to succeed
- Used by: Earthdawn

f) Roll one die and add rating to beat target number
- Example: roll 1D20 + skill (4) to with result 20 or higher to succeed.

g) Roll one die and subtract rating to roll under target number
- Example: roll 1D10 - 4 to roll under 5

h) Roll rating # dice of same type each to beat a target number, a minimum number of dice to beat target number is required for success.
- Example: roll 5D6 each die to beat 4. At least 3 dice must beat the target number for full success
- Used by: Vampire, Shadowrun

Can you help me fill this list?

Now I think the best transparency is to be had by the flat curve roll under or equal to target number systems because they readily give the chance (roll below or equal to 4 with a D6 gives the odds 4 to 6 straight away), but the roll over or equal to a target number with a single die  (a) is more straightforward and with low numbers even more intuitive.

Dice pools have a lot of advantages, but transparency doesn't seem to be one of them.

Mmm.. I really wanted to add a lot more to this thing before I sent it off, but I have to rush. What I want to do is to start with an overview of examples (please copy my list and fill in details like games using the different dice and such) and then we can analyze stuff matematically AS WELL as subjectively (I mean with the example systems we can easily rate stuff against each other right? - even though we might not always agree)

Quote
b) Roll under or equal to x to succeed.
- Example roll 45 or lower on a D100 to succeed. (chance of success 45 in 100)
- Used by: BRP style percentile systems

And a interesting variant:

Attributes and skills expressed as a potentially infinite precision percentile. Like 99.9999743%. The player rolls a single D10 aiming to roll under the first digit of the percentile. If equal, roll again against the next digit in the percentile. If lower, the success level (or concessions) is equal to the number on the D10 (0 -- 8). If above, the amount the D10 is above the digit is the extra complications (1 -- 9), the player needs to specify in order to succeed. Alternatively, the player can choose to let the opponent succeed.

- Used by: Ratio and Star Odyssey.

Ratio's advantages are:

[*]the player knows the exact percentage chance of succeeding (it's written on their character sheet);
[*]very low handling time;
[*]there's no math for successful rolls; and
[*]only one subtraction (D10 roll - skill digit = complications) or one addition (skill Digit + complications = D10)  for a "not successful" roll.
[/list:u]
For a generic SF and modern game, I haven't yet found anything better.

First something about Andrew's posting:

Interesting, especially that you too say it's "suitable for an SF or modern game".

Because here you touch on another aspect of die mechanics... the "Colour" it provides. Like someone pointed out elsewhere, heavy mathematics can destroy the feeling of the game, but sometimes it can be the other way around too. I could quite easily imagine an SF game with mechanics that "looked like" complex maths (it doesn't need to be real) to add that feeling to the whole game. Do you understand how I mean?

I guess this isn't anything new, I mean there are already games out there using cards, coins and even runestones (Wyrd).

However it also is something one might think about when one is designing mechanics. If it's SF it doesn't hurt to make it look a little complex even though it isn't (actually I felt the D6 in the original Star Wars destroyed the feeling a little because it seemed too simple, there wasn't enough technobabble in the die mechanics! ;) ), and on the other hand with Fantasy - don't make it like RM so that people have to bring their calculators (a group I played with all had calculators ready to whip out when combat started. YES, REALLY!)

Quote from: Pale FireInteresting, especially that you too say it's "suitable for an SF or modern game".

Because here you touch on another aspect of die mechanics... the "Colour" it provides.

Like someone pointed out elsewhere, heavy mathematics can destroy the feeling of the game, but sometimes it can be the other way around too. I could quite easily imagine an SF game with mechanics that "looked like" complex maths (it doesn't need to be real) to add that feeling to the whole game. Do you understand how I mean?

That's because for a SF or Modern setting, characters are far more likely to know their skill or attribute chances. For example, a factory made sentient machine will know it's success chance to the last decimal place, and most professional game players will know their statistical odds. For myself, I know from my diabete's nurse's clinic's testing machine that my muscle mass is +30% more than average. This measuring of human and machine ability will get better and more precise as time goes on, which naturally leads to specifying character ability in infinitely precise percentages.

For a generic fantasy game, I wouldn't use a percentile system, I'd use something else, like Fudge levels as the characters don't know their own abilities as well. So levels like Good, Legendary and so on, make a lot more sense from the character's point of view.

- Roll x dice, find matching dice
- Roll x dice, do whatever and consult Special Die (which is basically Roll 1 die, consult chart)
- Roll x dice, count dice which are higher than target number (this doesn't mean "success/fail" at certain number of dice. See Chthonian Redux)
- Roll percentile dice, compare to target number, flip 10's and 1's digit if wanted
- Roll d20, read 1's digit and consult one chart, read 10's digit and consult another chart.
- Roll x dice, roll y dice, compare highest, compare second-highest, compare third-highest, etc.
- Roll 2 dice, compare them.
- Roll 3 dice, pick middle value.
- Roll x dice, consult chart (this can be added to a lot of different types of rolls)

Those are the ones I can think of. I use most of 'em, too. I've had the quickest and most fun experiences with:

- roll percentile (1 die, essentially), compare to difficulty (Call of Cthulhu)
- roll 2 dice, compare them. (Dying Earth, Shadows)
- roll x dice, pick highest, consult chart (InSpectres)
- roll 1 die, apply modifier, compare to difficulty (d20)
- roll x dice, count all that rolled above difficulty, consult chart (Chthonian Redux)

And an equal (or more) amount of fun than some of the above games, though it's slower (it's worth it! Each success = 1 player-provided fact):

- roll x dice vs. y dice, compare highest, count # of dice above loser's highest. (Donjon Krawl)

Hi PF,

One thing to consider in choosing a die mechanism is that linear die roll resolution systems pay a price for their transparency. Basically, modifiers on linear die rolls are not well behaved -- meaning that modifier-based mechanics don't do what they're supposed to do, which is produce a reasonable outcome distribution when many different factors are influencing the chance of success. Many many systems have suffered from this without, in most cases, even realizing it.

I'm going to explain this in detail on a new thread, so as not to hijack this one.

- Walt

Aside from Walt's point which is discussed elsewhere, maybe we can have a look at what is the easiest to handle.

At a glance, roll under x seems way easiest as it provides the probability straight off. The only drawback is that it's slightly counter-intuitive to roll low. Oh, and another advantage is that if you use a fixed system ranging from 1-max die result you can have the success being equal to the number rolled, like in Pendragon for example. However this is again a bit counter intuitive and I personally find it a bit unsettling.

If you are using modifiers, my opinion is that the smaller the die, the better. It's easier to add +2 to a D6 roll than +37 to a D100 roll. Subtractions are in general more difficult than additions. Also something to keep in mind.

Roll vs another roll is something I try to avoid these days, although it has its advantages. Two rolls allows you to add modifiers to both rolls rather than work out a target number and that might simplify things, especially if the dice is high, like D20 or D100. The bad thing is that two rolls usually means twice the amount of time. And of course the chance of success isn't as clear.

However, if you play with a certain system you soon get the feeling of odds, especially if the rolls are similar. For example if you roll xD6 against the opponent's yD6 for everything, then pretty soon you're gonna get a good feeling of the odds. On the other hand if you're dealing with ED's Dx+Dy+Dz where you have so many combinations that pretty much every situation is new, you don't get that benefit. That's why I'm saying I think SR works better than ED. In SR you get a feeling for the odds eventually. For ED that takes much much longer.

Do people agree or disagree so far?

Hey,

"... it's slightly counter-intuitive to roll low ..."

I must say, I have never understood this claim. I hear it repeated every so often, and it's like aliens talking. Not in all my years of role-playing among hundreds of different people, has a single person ever had a moment's confusion about "rolling over" or "rolling under."

Nor do I agree that roll vs. roll is necessarily more difficult. It is more difficult if you have to calculate success/failure for each roll (e.g. Feng Shui, Vampire), but that is not necessarily the case (e.g. Prince Valiant, Sorcerer). And certain systems (e.g. ROS, Hero Wars, Dust Devils) have even made within-roll successful/failure more quick and functional for opposed rolls (I use "roll" here loosely, meaning any Fortune system). You're being mighty quick to discount this approach.

Best,
Ron

Quote from: Ron Edwards

Quote from: Pale Fire... it's slightly counter-intuitive to roll low ...

I must say, I have never understood this claim. I hear it repeated every so often, and it's like aliens talking. Not in all my years of role-playing among hundreds of different people, has a single person ever had a moment's confusion about "rolling over" or "rolling under."

Agreement here.  I compared "rolling over" a succeedingly lower target number to "rolling under" a progressively higher target number.  I concluded that having a higher target number representing a higher skill level was more intuitive (not better).  It just didn't seem to make as much sense to have a lower skill rating for a 'better' skill.  It's a subjective choice at best, meaningless at worst.

Fang Langford

Ron: Notice that I only say it's slightly counter-intuitive. The way I see it this is a matter of one is used to, and what one would be used to before playing RPGs would be boardgames where it's almost always beneficial to roll high. (Usually the 6 gives you that extra roll and stuff).

Of course after getting used to rolling low it's no problem. I'm just saying that initially it will feel awkward, but I don't think anyone will have any problems UNDERSTANDING it.

As for the roll vs. roll, I don't like it because of the extra handling time. Of course there are ways to minimize that extra time. My greatest objection would be with games which roll D20 (or bigger) + large modifier against D20 + large modifier, because the values won't be immediately apparent. This takes a lot longer than say roll 1D10 above a target number. Or roll D100 under target number or whatever.

However, if you use simpler methods, like in Sorcerer I think the calculation of the target number in the D10 above TN example and a target number free opposed roll should be at least on equal terms.

So I'm not ruling out that as a good option.