Forum changes: Editing of posts has been turned off until further notice.

Main Menu

[How Many Points] Atribute Distribution Question

Started by Comte, October 21, 2003, 10:37:03 PM

Previous topic - Next topic


This is a two part question.  First of all I have seen mention of things like probability curves for dice and things like that.  I would like to know if there are any resources on the net that has all this information listed somewhere.  I ask this because I seem to of turned into a googleing retard and the resource library didn't yeild anything helpful.  

I need this information because I am home brewing a game that uses a d12 dice pool, one is the lowest stat 12 is the highest.  For the char gen I was just going to drop a number of point that the players divide between 3 stats.  Is there a formula to determain how many points to give them?  Of course the number of points may be altered through play testing but I just need somewhere to start.
"I think where I am not, therefore I am where I do not think.
What one ought to say is: I am not whereever I am the plaything of my thought; I think of what I am where I do not think to think."

Mike Holmes

Just to restate, you have three stats that range from one to twelve, and you're wondering about creating a point system to allow players to select their own stats?

This is just way too little information, unfortunately. We'd need to know a lot more about the game to help out.
1. Is the stat curve more or less linear, or something else? That is, is a character with a 4 "twice as good" as a 2, and an 8 twice as good as a 4? Or is the curve exponential? For example, 3 is twice as good as 2, and 4 is twice as good as 3. Or some other curve?
2. Do 1 and 12 represent the minimum and maximum ratings for humans, or for everything (no human is as strong as an elephant; do stats need to cover elephants)?
3. What represents the "average" statistic for an individual? In either dramatic or in-game terms.
4. Are the PCs statistically "average"? Or are they extraordinary (or incompetant)?
5. How does the resolution system work?

Without answers to all of those, it's hard to reply in anything other than generalizations. But I can give you some stats advice. I'll start out with the remedial stuff, just in case. Don't be insulted if it's stuff you already know.

First, the average roll for any die is equal to the sum of the numbers on the sides, divided by the number of sides. 1+2+3+4= 10/4=2.5= the average roll for a 4-sided die. A quick simplification for calculating this for dice with an even distribution is to take the number at the middle of the range. That is, for a d10 with possible results from 1 to 10, the middle is 5.5 (leaving 5 numbers less, and 5 numbers more). Only on rare odd-sided dice, like the recent 7-sider, do you get an integer for the average - 4 in the case of the 7-sider.

Most gamers know the above. What they're not aware of is that for a single roll of one die, the average means nothing. The average can only become statistically relevant over multiple rolls. Still, that's enough predictive power for some designers to use just one die. (Bell curves have the advantage that the average rolls, or ones adjacent, are the most common).

The assumption in many designs is that a 50% target is considered an advantageous place to start. So, for a system in which, for example, a player had to roll a d12 less than or equal to their stat to succeed, they then assume that the "average" stat should be the one that gives them a 50% chance to succeed. So for a d12, that would be a 6, obviously, the number just below the average. Note that with an odd number range that you can't get a 50% chance distribution (with only one die), because you either have to choose the average number itself, or one above or below it.

This is all rather cliche design, but it would work if you had the same sort of goals.

Anyhow, once you've looked at your system, and decided what your "average" stat should be, then the question is how "average" do you want the characters to be? If, in the hypothetical, you wanted characters to "average on average" in terms of stats, and you decided that 6 was average, then for three stats, you'd want to give out 18 points. Because that would result in an average of exactly 6 distributed over three stats. If you want characters to be above average, give them some number on the scale multiplied by three or therabouts. 21 points will result in an average of 7, etc.

Again, this is all very cliche design. It assumes a ton of things that may not be true for the game in question (or even in most games that use such methods). But it does show how, quite often, such decisions are made, and the sorts of methods that are used to make them.

Now, I suspect that you might know some or all of this. What may be plaguing you is the probabilities of certain results. And that can be more difficult depending on how your system works (though if it's just a d12 roll-under, that's really simple). Anyhow, the thread I usually refer people to for this is:

Some of the math is very easy. Some of it is actually somewhat difficult, however. It'll all depend on how your system is constructed as to what formulae to use. In any case, if you give us more information, we'd be better armed to help out.

Member of Indie Netgaming
-Get your indie game fix online.


Well, we know it's a d12 "dice pool" system... which, I'm assuming, is going to mean "take a number of d12s equal to your score, and roll 'em."  But what counts as a success?  Do you add them all up and compare?  Do you count all D12s that roll higher than 8, or higher than 10?  Do you add up all the D12s that roll odd, and ignore the evens?  Are you looking for matches?  Is it simply Xd12, take the highest?

And as Mike asked, what is the average?  And how much better and/or worse are PCs supposed to be, when compared to that average?

I do have one suggestion, though:  whatever # you choose, don't make it divisible by three.  That's an easy numerical way to ensure that at least ONE statistic is different from the others, if a player goes for the "flat line" sort of character.  This suggestion is largely aesthetic, but it's something that I think is important.
Alexander Cherry, Twisted Confessions Game Design
Maker of many fine story-games!
Moderator of Indie Netgaming

Jack Aidley

QuoteI do have one suggestion, though: whatever # you choose, don't make it divisible by three. That's an easy numerical way to ensure that at least ONE statistic is different from the others, if a player goes for the "flat line" sort of character. This suggestion is largely aesthetic, but it's something that I think is important.

I would like to strongly disagree with this suggestion. By doing as Lxndr suggests you remove an option from the player. It annoys me that in many systems I can' t create a completely flat line character. Particularly in 'low-number' systems, where that extra point in strength (say) makes a big difference. Effectively you make it so your PCs must be either 'strong', 'fast' or 'smart' (taking an arbitary guess at what your three stats are).
- Jack Aidley, Great Ork Gods, Iron Game Chef (Fantasy): Chanter