D&D, Pathfinder (also D&D) and most of the OSR (still D&D) uses the d20. A nice flat distribution where every number has a 5% chance of occurring. This is good, the probability of rolling any particular range is easy to figure out, +1 matters, but not too much. There’s a reason it is still the default.
Runequest, Cthulhu, Mothership and plenty of others instead chose to go for added fidelity with 1% increments on a d100. This requires the use of weird ten-sided dice which are infamously NOT PLATONIC SOLIDS but that’s OK, the laws of physics were made to be broken right? Anyway the d100 makes everything super simple, a 67% Skill has a 67% chance of success, easy! I mean sometimes you have to add 25 to 67 to find your actual target but it’s good to practice maths, someday you might need it. Probably in another RPG.
Traveller decided 2d6 would be better, and then GURPS decided 3d6 would be even better. This was interesting because it brought triangular and bell curve distributions into task resolution systems. Game designers, crazy mavericks that they are, kept on going. Star Wars with its handfuls of d6 to add up and Shadowrun doing whatever the hell it was doing. I guess normal distributions is what it was doing.
The thing is none of this really matters, not in a “system doesn’t matter” way, but because in all of these games it’s not the actual probability of rolling individual numbers which is important. You pass a test when you hit one of a particular range of numbers, and when you start adding up the probabilities of rolling all those individual numbers, the differences start fading away.
Maybe some graphs would help?
Here’s your chance of hitting your target number on 3d6 and on 1d10 overlaid. They are 98% the same. Unless the GM spends a lot of time asking for players to make rolls of less than 4 or over 17 on 3d6, no one will notice the difference. And in fact, with those extreme rolls it would still take 20 tests on average to notice the difference.
What’s that? Powered by the Apocalypse games require the 2d6 triangle distribution and they account for 120% of all RPGs produced in the last ten years? Yeah, they work with linear dice too, you just need a slightly different look-up table.
| 2d6 | d12 | d6 | |
| Full Success | 10+ | 11+ | 6+ |
| Mixed Success | 7-9 | 6-10 | 4-5 |
| Fail | 6- | 5- | 3- |
So what’s my point?
What’s actually happening here is that dice systems are just providing you with a selection of pre-chosen steps along the number line from 0% to 100%. Mostly RPGs focus their dice rolls between about 30% and 90% because that’s where rolls are the most interesting; that’s where players are willing to take chances. These systems are further limited by the fact that each step along the path has to be significant enough that it feels like a meaningful jump. Usually, that’s about 5 to 10%. These restrictions don’t leave dice mechanisms much room to be functionally distinct.
D&D etc use 5% increments which can feel slow but that works because they want a slower progression for the epic zero-to-hero character arc.
Even dice pool systems tend to follow this pattern. Dice pools do produce diminishing returns so initial steps tend to be higher than 10%, and later steps lower than 5%, but it’s the same ballpark. Here’s the chances of success with between 1 and 10 dice in Free League’s Year Zero Engine:
Or if you look below you will see the odds of getting a pass, mixed success or failure in Blades in the Dark. Notice that each dice added increases the chances of success by about 12%, and after the Mixed Success chance settles down at 2d, the chances of a Fail diminishes by about 10% per dice.
If you still aren’t convinced the next chart shows what the Blades system looks like if you do the laziest hack ever and swap the dice pools for d10+stat and use the PbtA ranges for success. It’s not identical, but it’s pretty damn close!
So what was my point again? Well it’s not that system doesn’t matter, it does, but there are only so many ways to use dice well and most of them produce pretty similar results. The important part of a system is the bit where it tells you *when* to roll.
Traveller tells you to roll the dice to find out if your character has the right training for the task at hand, because it’s an RPG about middle aged fuck-ups in a capitalist hellscape, erm, in space!
Mothership tells you to roll to see if you freak out when you watch your crewmate get torn apart by an alien ball of claws and hate, because it’s a game of survival horror. In space!
Powered by the Apocalypse games are interesting because they each have unique triggers for which situations need dice rolls and rules. I could have probably actually just talked about them first and saved a lot of words. Oh well.
D&D 5e has you roll the dice to hit things and to try to seduce half-orc tiefligings and to see if you know that the person who is clearly lying to you is lying to you… Because 5e hasn’t got a clue what it’s on about!
So what is the best dice based random number generator for you tabletop RPGs? Well it is d20 roll-under blackjack with Advantage and Disadvantage where appropriate, but *that doesn’t matter* because, what you roll isn’t nearly as important as *why* you are rolling.
Although if your new RPG purchase is one of those where you are supposed to take your stat and your skill and roll those with a bunch of dice with some different things and your equipment adds other stuff and your friends can help with meta-tokens and all of that is then rendered into so much irrelevant busy work when it is finally compared to a target number that the GM pulled out of his ass? Just don’t.
Ominosity 