Still Searching for the Perfect Die-Roll System

[Tom]

I'm back... After experimenting with Roll-and-Keep, a chance find led me to something that I think is the perfect system.

I've written up a slightly long in-depth analysis over here: https://dl.dropboxusercontent.com/u/9638874/DiceProbabilities.pdf


The idea is very simple: Roll 2-10 dice, take only the highest two and sum them up.

The result is a wonderful probability curve, with everything I've wanted in a dice system. It's easy and fast (only comparison and one addition), it gives smooth, consistent curves (no odd edges like with exploding dice systems) and the holy grail: Higher skill leads to not only higher success chance, but also higher consistency (less spread) - most dice systems work the other way around!


Comments?

[Ron Edwards]

If it makes you happy!

You probably already know that when someone says "This is the perfett mechanic," my constant response is "for what," but I think that's not the conversation you're looking for. All I can say now is, enjoy.

Best, Ron

[Tom]

I know where you're coming from.

You will notice that I am not looking at any specific "for". I am looking at dice mechanics purely for their statistical values, under assumptions that I've laid out elsewhere, I think posted to this forum as well.

Basically, what I want in a dice system is a simulacrum of realism, and I've years ago set up basic tenets for that. I made this public first in 2013 at a gaming convention in a talk, here's the point where a slide sums it up fairly good: https://youtu.be/hYHnZjASkNo?t=21m13s -- sorry, it's all in german.


The idea is that no matter what your gaming system, setting or mechanic (skills, magic, whatever) you have:

a) a character
b) the world

influencing the probability of success. Many dice systems actually model exactly this without making it explicity - the most common mechanics have a number or type of dice that the player rolls vs. a threshold or difficulty that the circumstances define.

But there are also four more aspects that a good mechanic needs to satisfy:

* intuitive correctness
* diminishing returns
* the effect of skill (e.g. in a dice game your gambling skill is less important than in a card game where you can use strategy)
* spread

All of these requirements are not from any setting, but simulate to some extent how the real world works. Except for settings like dreamlands, we expect the mechanics to behave in what feels to be right to us. This is what I am trying to look at, irrespective of setting.


On that foundation, more can be built. For some systems, you want luck to be a bigger factor. For some systems you want various ways for the players to influence the results. For some systems, you want a lot of modifiers so as many factors as possible are correctly simulated. For other systems you want the GM simply eyeballs a difficulty number.

However, just as system matters, so do dice mechanics. If your dice mechanic is to throw 5 d100 and take the square root of the difference between the highest and the lowest, your mechanic is shit because it is easy to come up with a much simpler system to generate a skewed distribution between 1 and 10.

The exact dice mechanics need to fit the setting and the game, of course. But we can objectively say that some dice mechanics are better than others. Sorcerer has a brilliant mechanic with almost perfect marks in all requirements. But anyway I've tried for three years to invent something better. :-)

[Ron Edwards]

I didn't ask for an explanation, but you've definitely outlined your priorities clearly. Again: I'm glad you're happy, and enjoy.

Best, Ron