As I have written in the past, I struggled with the right dice to use in the game. I was tempted to use a d10 for quite some time, and I flirted with the normalization afforded by a bell curve of some kind.
The Twenty Sider
In the end, I decided to go with my old friend, the d20. To utilize it properly, however, I think one needs to attach special effects to the 1 and the 20. I mean, the point of the die is the thrill of rolling a 20 and the horror of rolling a 1. Since each has a 5% chance of coming up, it makes it clear that, for example, a d10 wouldn’t work. But I took it a step further, and actually put different special effects on 17, 18, and 19, as well as the 1 and the 20. That means that there’s 25% chance of something extra special happening on any given action–5% bad and 20% good. Here’s what I did:
On a 17 or 18, you get a Minor Effect. Basically, if this is a successful roll, you get to state something special that happens as a part of that action. Maybe in your jump you land with a flourish, or also knocked over the burning brazier onto the curtain to start a fire. Attack rolls with a Minor Effect either do additional damage or affect foes in some way–knocking them back, for example. On an 18, any points you spent on the action (such as through effort) are also reduced.
On a 19 or 20, you get a Major Effect. This is similar to the Minor Effect, but, you know, bigger in some way. You not only kick in the door, but momentarily intimidate all the foes on the other side. In battle, Major Effects do more damage or affect the targetsdramatically–knocking them down, disarming them, etc. On a 19, any points you spend on the action are reduced. On a 20, whatever points you spent return to your pool.
If the GM decides that the effect that you describe has a chance of failure, there may be a second roll involved. (Something simple, like adding damage in combat, never requires a different roll.) It’s worth noting that damage in Numenera is based on your weapon and other factors but isn’t random. So the variability in damage comes from Minor and Major Effects, not in a separate damage roll. Not only does this keep things moving faster, but it always means that if you roll well in combat, you’ll do better damage (or get a special effect).
On a 1, a GM can use “GM Intrusion” without awarding any experience points for it. Basically, things just get more complicated. This doesn’t always mean a “fumble” of course. It might mean that an opposing NPC does something really remarkable or lucky. It might mean that a seemingly random event makes things more interesting. It doesn’t really make narrative sense for someone extremely skilled at something should actually fumble at it 5% of the time. That’s also why if you are really good at a task, you can reduce the difficulty down to a point where you don’t even need to roll at all, and thus there’s no chance for a 1. (I’m experimenting in the playtests with a couple of ways that give you a chance at a good effect even in the “auto success” range.)
In the end, this means that with a typically sized group, there will be a special roll of some kind once or twice each round, which makes for very dynamic, exciting play.
This is why in the Numenera Kickstarter I put a unique dice set in as a stretch goal. I want to create a d20 with an indicator on the 5 results that have special effects. I’m still in talks with dice manufacturers to discuss the ins and outs of making this happen. (It’s not a matter of if it will happen–it will. It’s a matter of how it’s best done.)
The other dice that you need to play Numenera are 1d6 and 2d10 (percentile). The d6 is used only in rare instances. It’s almost inconsequential, but I figure everyone’s got a six sider laying around, so it’s not a strain to call for it occasionally. The percentile dice are used by the GM when rolling for various random things that only the GM rolls for. The Numenera book is going to be filled with lots of tables for the GM, mostly to help in adventure creation. Tables about the weird aspect of an isolated community, about mutations, about bioengineered effects, about incredible devices, and much, much more. As a game about ideas, I want to ensure that no Numenera GM ever runs out. (And the best thing about tables is, if you don’t like random generation, you can just pick what’s appropriate.) These tables will always be percentile based. So since GMs don’t roll for NPCs, the 2d10 become the GM’s dice.