One of my top 5 design goals for YARPS was to not impede characterization (or personification, if you prefer). To that end, I eliminated the intelligence, wisdom, and charisma attributes. In my experience, players rarely play down to these attributes when they're low, and it's rather hard to play above your own human limitations when they're high. So it will be up to the players, rather than the system, to decide how smart, stupid, wise, foolish, charming, or obnoxious their characterization is.
That leaves the three physical attributes: strength, dexterity, and constitution (I prefer the term toughness, but I worry that changing the attribute name would introduce too much confusion for too little benefit). To that I added two new attributes: speed and comeliness (a measure of how pleasing a character's physical appearance is). These, then, are YARPS five standard attributes.
To generate a character's attributes, a player rolls 4D6 five times and assigns the score, the sum of each roll, to the attributes in order. Why did I choose 4D6 instead of 3D6? In a word: probability.
There are 216 possible results for a roll of three six-sided dice. This means that the probability of rolling the lowest and highest score (3 and 18 respectively) is 1 in 216, or 0.46%. Applied, it means that 1 in 216 characters will have the highest (or lowest) possible score for a given attribute [2]. That's just too often for my liking.
A roll of four six-sided dice has 1296 possible results. A character rolling the highest (or lowest) score for an attribute 1 in 1296 times (or 0.08%) is still too frequent for my liking, but it's much more tolerable. And summing the roll is still mentally manageable (I've noticed once you add 7 or 8 dice into a summed pool players start slowing down while tallying; this matters to me because I'll be using 4D6 as my "core" competency mechanic for YARPS as well, instead of 1D20).
I'm still going to have a 1-3 attribute modifier range (-3 to -1 for low attributes, +1 to +3 for high attributes), but having a character with a -/+2 or -/+3 modifier will be more exceptional. This is how it will break down:
Score Modifier Probability
----- -------- --------------------
4 -3 1 in 1296 ( 0.08%)
5-7 -2 24 in 1296 ( 2.62%)
8-11 -1 275 in 1296 (21.21%)
12-16 none 676 in 1296 (52.17%)
17-20 +1 275 in 1296 (21.21%)
21-23 +2 24 in 1296 ( 2.62%)
24 +3 1 in 1296 ( 0.08%)What I like about this table:
- Less than .2% of characters will have the the lowest or highest score for a given attribute (-3 or +3)
- Roughly 5% of characters will have an exceptionally low or high score (-2 or +2), often enough that a given party should have a few such scores
Comparing this to Moldvay Basic:
- -/+ 3 goes from 0.92% in Moldvay Basic to 0.16% in YARPS
- -/+ 2 goes from 8.3% to 5.24%
- -/+ 1 goes from 42.56% to 42.42%, nearly the same
With this I've accomplished what I wanted, which is the same probability for -/+ 1, but a lower probability for -/+ 2 and -/+ 3. With this, I can design the system such that +1 is a real boon (or penalty) for an action test or effect roll, without worrying too much about balancing +1 against +2 and +3. More on how I intend to do that in my next post on YARPS action tests.
The most-mentioned concern I've heard when posing the 4D6 idea to others is that a maximum (or minimum) roll will now be so rare that you could game for a year (or years) and never see it during character generation. I considered this, and based on my experience and anecdotes, I think it will still happen, just much less frequently (which is my desire). I guess play-testing and time will tell. It'll be an interesting experiment, regardless.
[1] The design "base" for YARPS is Moldvay/Cook B/X D&D, but I'm also leaning heavily on White Box D&D.
[2] Treating an attribute roll as a single event when trying to determine overall probability of exceptional attribute scores is of course incorrect; you must consider all five attribute roll events. But in the interest of simplicity, I'm foregoing proper math.
