dice mechanics - what types are there?

[Raleel]

So, I want to take a look at dice mechanics a bit and think about different types of core mechanics in a game. I'll lay out a few examples so you can get an idea of granularity

• linear - this is d20 and d100. Roll a single die (even if it's multiple dice interpreted as a single die) and go against the target number, above or below. I don't care about crits or fumbles, as those are particular interpretations of results. Savage worlds ends up here as well, as they tend to have singular dice of a particular size. Exploding just adds to the single die value, and is not really a bell curve.
• bell curve - 2d6, 3d6, 2d10, etc. you take the total of the dice against a target number. The important feature is that it is a bell curve fudge dice end up in here as well, as they are functionally totaling the dice. 5e's advantage/disadvantage ends up here as well
• dice pool - WoD and SR are notable examples. Roll the pool of dice, count the number of results that hit a target number. the shape of this changes a bit based on the TN and the number of dice, but it's not linear exactly, and not really a bell curve. Truth, it is probably closer to linear in most ways, and it could be argued to be.

so, what other ones are out there?

 

[Dumarest]

There's that one where you roll two dice and subtract one from the other and add the result to an attribute, which I find needlessly clumsy and refuse to play.

 

[Thanks For The Memories]

Is roll two-pick one a bell curve? It's two linear dice for the same roll, you're just picking one or the other. Neither one changes the median range it just increases or decreases the chance for success, same as if you re-rolled one die.

 

[Luca]

Stolze's One Roll Engine used for Godlike.

It's a variation of the dice pool; you roll a bunch of D10s and count matching results. The number of matching dice is the "width" and the number rolled is the "height". So if you roll 5D10 and three dice come up with eight, the set would be a 3x8 = width 3 and height 8. Width is the speed with which you execute the action and height the degree of success. If you get multiple sets you can choose which one to use.

You also have "Hard dice" and "Wiggle dice" associated to powers. Hard dice always count as 10. Wiggle dice can be assigned after the roll.

 

[soltakss]

Some systems use Exploding Dice, which is nowhere near as fun as it sounds. basically , you roll and if you get a certain score, normally the maximum possible, then you roll again and keep adding until you don't get the maximum, allowing you to get a big score.

 

[Raleel]

There's that one where you roll two dice and subtract one from the other and add the result to an attribute, which I find needlessly clumsy and refuse to play.

that's a bell curve, just one die is negative

 

[Raleel]

Is roll two-pick one a bell curve? It's two linear dice for the same roll, you're just picking one or the other. Neither one changes the median range it just increases or decreases the chance for success, same as if you re-rolled one die.

jsut a very small dice pool, now that i think about it.

 

[Raleel]

Some systems use Exploding Dice, which is nowhere near as fun as it sounds. basically , you roll and if you get a certain score, normally the maximum possible, then you roll again and keep adding until you don't get the maximum, allowing you to get a big score.

ya, that's just a linear one, actually. the effective size only increases if you get a particular result. the target number doesn't really change.

if the Tn is over the die size, that might be something, but I think not.

 

[Raleel]

Stolze's One Roll Engine used for Godlike.

It's a variation of the dice pool; you roll a bunch of D10s and count matching results. The number of matching dice is the "width" and the number rolled is the "height". So if you roll 5D10 and three dice come up with eight, the set would be a 3x8 = width 3 and height 8. Width is the speed with which you execute the action and height the degree of success. If you get multiple sets you can choose which one to use.

You also have "Hard dice" and "Wiggle dice" associated to powers. Hard dice always count as 10. Wiggle dice can be assigned after the roll.

ya, that seems like a dice pool + linear. the width is the dice pool, and the target number is the die with the most results, I suppose.

 

[Dumarest]

that's a bell curve, just one die is negative

Still sucks, though.

 

[opaopajr]

There is only a few mechanics, whereas the dice formulae just provide the % gradient and limits:

Single Rolls
Pass/Fail (threshold) -- most rolls
Degree of Success -- distance from threshold
Specific Value -- the values itself provide new outcomes, e.g. Random Encounter Table

Multiple Rolls
Extended Contest -- more success needed over time
Contested Roll -- more success vs. opponent
Assisted Contest -- more successes provided by an ally

They may be combined for greater complexity.

What exactly are you trying to discover?

 

[Raleel]

just options I think. part of this was born out of bell curves vs linear thoughts. part of it is figuring out dice pools for shadowrun and how to convert things. and part of it is what is not worth my time to work on

 

[KrakaJak]

I see Dice Pools as an exponential curve. Each die you add to a pool has an exponentially smaller impact on your probability of success. It helps discourage players (at least the mathematically inclined ones) from specializing too much on a single stat or skill.

Special result/complex dice pools - FFGs Star Wars is probably the best example of this: Dice with varying sides and special symbols form a pool where you can get mixed results, i.e. Success but with a critical misfortune (your blaster fires true and kills the stormtrooper, but it's circuits have fried, turning it into a paperweight). Vampire 5e has a limited version of this with Hunger Dice.

 

[Aos]

WEG Star Wars has the dice code system. Each test is based on a number of dice specific to the skill/vehicle or what have you, rolled against a base difficulty of 5,10,15 or 20. Difficulty can increase past 20 due to mods.
The number of dice rolled is called a dice code. So if you fire a blaster at a point blank target and have 2d6 code- well, I’m sure you get the rest. Damage is based on a simple level of success.
There is no dice code upper limit.
chases are done with opposing roles, however.

 

[Aos]

There's that one where you roll two dice and subtract one from the other and add the result to an attribute, which I find needlessly clumsy and refuse to play.

ICONS has this. It drove me back to MSHRPG. That ansd the absolute shit organization of the rule book.

 

[Bilharzia]

There's the mutant year zero / Forbidden Lands system which is a dice pool with pushes. You build your pool from attribute dice + skill dice + gear dice, make your initial roll. All "1"s are fails, all "6"s are a success. To succeed you usually just need one 6, regardless of whether you roll any 1s on this roll. To push, you re-roll any dice which did not come up 1 or 6. On this pushed roll you might score extra 6s which give extra benefits, if you roll 1s you might damage your gear or harm yourself (and now previous 1s you rolled with your attribute dice can now harm you).I don't know exactly how that shapes the probability graph, the rules do include a table showing the probabilities. The system is built around encouraging players to risk pushing as the odds are pretty low without it

.

 

[Ladybird]

that's a bell curve, just one die is negative

It's also zero-centred, which is an interesting property - it means your number also doubles as it's "passive" equivalent for people rolling against you, rather than having to add a number to it to get a target.

 

[Caesar Slaad]

• dice pool - WoD and SR are notable examples. Roll the pool of dice, count the number of results that hit a target number. the shape of this changes a bit based on the TN and the number of dice, but it's not linear exactly, and not really a bell curve. Truth, it is probably closer to linear in most ways, and it could be argued to be.

That's actually two mechanics. WoD is binomial ("count successes"), SR and WEG SW are additive dice pools.

Another method other than adding or binomial is "take the best", like Blades in the Dark.

A minor trend nowadays is to give some dice different significance. For example, in AGE, one die is a different color. If you get doubles, it generates stunt points. Otherwise it's a basic bell curve.

Then there is the custom/symbol dice trend, like Genesys and The One Ring.

 

[Mankcam]

Modiphius 2D20 is both linear and dice pool

 

[dbm]

‘Roll & Keep’ might qualify as different? EABA uses this, and I think Legends of the Five Rings might have in earlier iterations, but I’m not sure on that one.

In EABA, your skills and stats are expressed primarily in dice, getting you more and more of them. But a human-level being can only keep their best three dice (more powerful beings can keep more). So, as you become better your average will increase but the limits of your ability stay roughly the same.

And how could people forget the amazing new dice mechanic introduced in Cyborg Commando where you roll two d10 and multiply the rolls together? I think it might have been called dx, but there my memory starts to let me down...

 

[Nick J]

Surprised nobody mentioned the beloved classic "Cyborg Commando?" Roll 2d10 and compare the product of the result against a skill rated to 100 (I think? I confess it's been thirty years since I got duped into picking up that dumpster fire of a game out of a Value Village bargain bin).

Edit:
Whoops! Credit to dbm for singing the praises of this "classic" and boo on me for skimming on my phone.

 

[Morblot]

Surprised nobody mentioned the beloved classic "Cyborg Commando?"

...have we been reading the same thread?

 

[Tom B]

‘Roll & Keep’ might qualify as different? EABA uses this, and I think Legends of the Five Rings might have in earlier iterations, but I’m not sure on that one.

In EABA, your skills and stats are expressed primarily in dice, getting you more and more of them. But a human-level being can only keep their best three dice (more powerful beings can keep more). So, as you become better your average will increase but the limits of your ability stay roughly the same.

And how could people forget the amazing new dice mechanic introduced in Cyborg Commando where you roll two d10 and multiply the rolls together? I think it might have been called dx, but there my memory starts to let me down...

I'll toss in the EABA probability chart. It shows the percentage chance of success based on difficulty and the total net attribute+skill dice.

 

[AsenRG]

I classify them as follows:
Roll Over+/-modifiers
-Single die, with or without modifiers (Talislanta, Witchcraft)
-Multiple dice (Kuro, BoL, Zenobia, Classic Traveller)
-Varying number of dice (Fractured Kingdoms, for example, and T&T)
-Varying sides of the die (Fighters United). That's a mixed group with Stepdice.

Roll Under+/-modifiers
-Single die (Pendragon, Mythras, Flashing Blades, Dragon Warriors, Maelstrom)
-Multiple dice (GURPS)
-Varying number of dice (Traveller 4 and Traveller 5).

Dicepool
-Success-counting (Exalted, Atomic Highway)
-Additive (D6 system, T&T...yes, I know I included T&T above, too, and that's not an accident)
-Roll-and-keep (L5R, the ONE AND ONLY edition of 7th Sea).
-Roll X, keep N (EABA v2)
-Keep Highest (Silhouette, Sorcerer)
-Keep Lowest (Blue Planet 2nd edition - and 3rd edition, too, judging by the Quick start)

Stepdice:
These are only its own group because sometimes they use different kinds of resolution for different tasks. Beat TN and count successes, Take Highest and Compare, Take All and Add Modifiers: Usagi Yojumbo/Ironclaw/Myriad Song, Savage Worlds and Earthdawn are all examples of a mix of the approaches).


And finally, the weird:
Roll Over-or-Under:
Trollbabe!

 

[Picaroon Jack]

Stolze's One Roll Engine used for Godlike.

It's a variation of the dice pool; you roll a bunch of D10s and count matching results. The number of matching dice is the "width" and the number rolled is the "height". So if you roll 5D10 and three dice come up with eight, the set would be a 3x8 = width 3 and height 8. Width is the speed with which you execute the action and height the degree of success. If you get multiple sets you can choose which one to use.

You also have "Hard dice" and "Wiggle dice" associated to powers. Hard dice always count as 10. Wiggle dice can be assigned after the roll.

Have you played Godlike? How did it go?

 

[Edgewise]

It's also zero-centred, which is an interesting property - it means your number also doubles as it's "passive" equivalent for people rolling against you, rather than having to add a number to it to get a target.

You can get the same curve and values by 2d6-7. The advantage of this is two-fold: (a) no subtraction, and (b) no need to keep track of which die is negative and which is positive.

 

[dbm]

You can get the same curve and values by 2d6-7. The advantage of this is ... no subtraction,

I think you might want to read your process again...

 

[Ladybird]

I think you might want to read your process again...

The advantage of doing it using fate dice is that there's less subtraction; sure, you can get a similar zero-centred curve with 2d6-7, but then there's a subtraction on every roll (People tend to be worse at, or more put off by, subtraction than addition) or you roll that 7 into all your numbers and everything looks weird.

Obviously you can still get negative results on a fate roll, but the range of numbers is much smaller and people who have difficulties with maths are better able to cope with those.

 

[AsenRG]

You can get the same curve and values by 2d6-7. The advantage of this is two-fold: (a) no subtraction, and (b) no need to keep track of which die is negative and which is positive.

But the disadvantage is that your dice can only explode in one direction .

 

[sharps54]

I like ORE in theory, in fact it should work very well in a PbP game, but I haven’t ever actually played it.

 

[ffilz]

I'll add one more, a true normal distribution (bell curve). A college friend came up with a way to use the actual normal distribution to generate modifiers to be added to attack rating to be compared against defensive rating. The idea is that instead of using d10 to roll the digits of 1d100 or 1d1000, you use d10s to generate the decimal digits of a number between 0 and 1 (so if you roll a 5 and a 6, you rolled a 0.56). He then took the normal distribution and scaled it so that one standard deviation above average is a +20/3 modifier (on this scale, that 0.56 is a +1). If you roll close to 0 or close to 1, you roll additional digits until you have 2 digits past a leading string of 0s or 9s. So if you roll 0.99, roll 2 more dice, say a 9 and 7, so now you have 0.9997, so one more die, a 3 gives you 0.99973. Now there is a downside, you now need to look up on a chart (0.99973 is a +23). Here is an Excel Spreadsheet that computes the table: https://ffilz.github.io/Gaming/cddnorm.xls (my friend didn't have the luxury of this back in the 1980s... But his computation only has 3 minor differences). The table lookup doesn't end up being all that bad for two reasons. First, in most cases you need better than a +0 to succeed (no worse than D&D where you often need a 15+ to hit), and second, especially given that you typically need better than a +0 to succeed, it's not that hard to remember the first 5-10 numbers. The result is typically if you need to do a table lookup, things are exciting, you're actually about to actually hit something real hard to hit, you're expert fighter is actually about to fumble (to fumble you need to roll at least a modified attack rating of -5 or lower, which given most folks have at least an attack rating of 3 or 4 is at worst about 10% and that's for a 1st level character who doesn't have great stats), or you're about to score a really nice critical hit (because the normal distribution is open ended, critical hits are scored if your modified attack is 7 to 10 (or more for really tough monsters) above defensive rating you score double damage, with an additional damage multiplier for each additional +2.

Now one could argue that this is part of a category still not listed - roll on a table (I'm pretty sure there are some systems like this - heck, Avalon Hill board games fit into that category if you don't want to count things like D&D magic item tables). I think in some way this normal distribution is still a different kind of animal than an Avalon Hill CRT.

Frank

 

[Edgewise]

I think you might want to read your process again...

OK, you totally got me. I guess what I was really going after was that it's easier to subtract a constant value (7) from a dynamic value (2d6) than one dynamic value (1d6) from another (1d6). But I hadn't even thought it through till now.

 

[Edgewise]

But the disadvantage is that your dice can only explode in one direction .

That's actually totally true. As you know, Lark compares two 1d6 roll and both dice explode. I came up with a weird mechanic to simplify this to a single 2d6 roll, which is to have custom dice. I call these dice "explosion dice," slightly distinguished from the concept of exploding dice.

Anyway, they come in pairs, and one of the dice has the faces {-2, 2, 3, 4, 5, 6} and the other has {1, 2, 3, 4, 5, 9}. This ends up producing a bell curve that's flattened at the edges and extends from -1 to 15. Subtract seven and you have a normalized curve from -8 to 8.

For a +/-2 "explosion," it just ranges from -7 to 7 with a similar shape. In both cases, you still have a 1/6 probability of a result of zero, making me think that would also be true for other +/-X explosion dice. Probably within certain bounds for X, but notice that you always subtract seven.

 

[Ladybird]

That's actually totally true. As you know, Lark compares two 1d6 roll and both dice explode. I came up with a weird mechanic to simplify this to a single 2d6 roll, which is to have custom dice. I call these dice "explosion dice," slightly distinguished from the concept of exploding dice.

Anyway, they come in pairs, and one of the dice has the faces {-2, 2, 3, 4, 5, 6} and the other has {1, 2, 3, 4, 5, 9}. This ends up producing a bell curve that's flattened at the edges and extends from -1 to 15. Subtract seven and you have a normalized curve from -8 to 8.

For a +/-2 "explosion," it just ranges from -7 to 7 with a similar shape. In both cases, you still have a 1/6 probability of a result of zero, making me think that would also be true for other +/-X explosion dice. Probably within certain bounds for X, but notice that you always subtract seven.

That's a clever idea. I do have a pair of sicherman dice which I dremeled myself, but I've never had the confidence to actually use them in a game.