If you’re designing a roleplaying game system, which dice should you use?
I believe that dice in an RPG have only one purpose, and that is to introduce randomness to a given attempt. The faster and easier this is made for the GM, the better the game will flow.
 There may be some minor exceptions such as Randall Hansen’s epic dice mechanic, e.g. dice granted by the GM to reward epic, awesome and cool actions by a player.
I’m going to make the case for d100 – that is, two ten sided dice, also known as d10s.
 With d100 you roll two ten sided dice to get a result between 1 and 100. One die represents the tens digit and another the ones digit. There are even dice called decaders – that’s a d10 marked (10,20,30,40,50,60,70,80,90,00) which make reading the result even easier.
Arguments in favor of the d100
1- D100 are perhaps the fastest and easiest dice to read because you don’t need to do math in your head to get a result. The game moves faster, and everyone instantly knows success or failure just by reading the dice.
2 – with d100 you get a lot of granularity. As a GM I use the degree of success or failure to flavor the description of the result. Thus if you need 60 to succeed and the player rolls a 60, I’d describe just the barest possible success. A roll of 61 would still be just barely succeeding, but a little better. This makes coming up with interesting descriptions easier for me.
3 – D100 is quite easy for the GM to assess difficulty as we’re quite accustomed to dealing with probabilities expressed as percentages. Our lexicon is filled with them: fifty-fifty, eight out of ten, a ten percent chance, a third of the time, etc. The faster I as a GM can give a difficulty to a player, the more the game will flow.
4 – Using d100 is also quick for players to calculate risk. If I know I only need a 20 to succeed it’s easy for me to say I have a good chance of success. But d20 + 3d6 + 2d12? Not clear. As a player, if I clearly understand the risk, it’s easier for me to take action.
5 – D100 best lends itself to prosaic descriptions from the GM, a cornerstone of the game system I’ve been working on with a couple of friends of mine for the last five years. More on that later.
Arguments against the d100 and my counterarguments
I’ll add arguments against the d100 here as I become aware of them.
But there’s no skill progression
D100 is fixed. Your maximum roll will always be 100. No matter how good you get, your skill doesn’t effect the outcome. Skill is only taken into consideration by the GM when assessing difficulty. Another way of putting this: if the skill difference between, say, level 1 and level 2 isn’t part of the player’s die roll, what is the point of having skill progression?
Adding dice over time just creates an arms race. All this does is require the GM to ratchet up the number required for success over time. It complicates the process for no real gain, other than the physical satisfaction of chucking a lot of dice.
Skill checks should always be challenging for a player. We become bored quickly when things are easy.
Yes, characters can do things easier as they level up. But only a poor GM would keep putting the same low level challenges in front of a high level player. Boring!
Next, consider for a moment whether it is even possible to answer this question: “What is the difference, mathematically, between level 1 and level 2 in a all skills?”
The answer is, “it doesn’t matter.” The game does not improve if we try to answer this question. A game system improves if it is more fun, engaging, easier to play, or gives more options without adding undue complexity.
 It has to be for all skills, or we doom ourselves to endless edge cases and bickering over granularity.
Please let me know your comments or arguments against the d100. I’ll try to address them in edits of this page.