Ability scores revisited

For a variety of reasons, I've always been a fan of 3d6 in order when generating ability scores for D&D characters. However, after spending an embarrassing amount of time generating characters at http://character.totalpartykill.ca/basic/ (Why is this SO MUCH FUN?!) some potential flaws have become apparent. Firstly, "hopeless" characters with no above-average scores and multiple way-below-average ones are a lot more common than you might think. Moldvay's suggestion of allowing players to discard such a character and roll a new is an acceptable patch, but is, to my mind, an inelegant solution at best.

Many people wouldn't consider the inverse of this, a character with no low scores and multiple very high ones, to be much of a problem. (I, on the other hand, have been sorely tempted to fudge scores downward if I roll an uber-character, because I find such characters profoundly uninteresting ...) Depending on the players and the group as a whole, it may not be, but in my groups, having one character with an 18 and a 16 and a couple 13s can easily make those who have a 14 in their prime requisite, a 7 or 8 in another score, and a bunch of 9s and 10s feel inadequate, especially if they're playing the same class as the super character. Imagine being the fighter with 14 Strength and 8 Dexterity on the same team as the guy with 18 Strength, 13 Dexterity, and 16 Constitution, and you might feel a twinge of resentment at being constantly upstaged.

One ubiquitous response to this has been to use methods that alter the distribution of ability scores -- typically toward the high end of the scale -- 4d6 drop lowest and other, even more extreme, dice-rolling schemes. What I'm looking at is something similar that will weight randomly-rolled scores toward the middle of the scale. The characters I find most appealing are the ones with abilities mostly in the average range, with one or two particular talents and weaknesses. That fighter above -- the St 14, Dx 8 guy -- would be a blast to play ... provided he doesn't have to work alongside a lot of people like that other guy who overshadow him in every way.

Anyway, to get to the actual point, I've been running some dice-rolling simulations on anydice.com, and the method that produces the most pleasing results to me is 2d6+3. Yes, I'm aware that doesn't allow the full official range of scores from 3 to 18, and I'm entirely OK with that. Let me show you why.

With 2d6+3, you'll get ability scores from 5 to 15, with an average of exactly 10 (compared to 10.5 for 3d6) with a smaller standard deviation. Extremes are a little less common, and the really extreme scores aren't possible to roll straight. A 13 score is a legitimate talent that only 16% of characters will have naturally. But wait, there's still the point-trading aspect of B/X to consider: You can increase an ability by 1 point, at the cost of lowering another by 2. By the book, you can only raise a prime requisite, and can't lower any score below 9, but we can expand that to allow raising or lowering of any score, but no score may be lowered below 6 (or raised above 18.) You want that +2 or +3 bonus? That kind of exceptional talent is reserved to those who work and sacrifice for it! You might be born with the potential to be amazingly strong, smart, charismatic or whatever (a 14 or 15 natural score) but developing it to that degree is a choice, one that requires trade-offs. This way, characters who are truly amazing in an ability are going to be relatively uncommon, and those with 16s or higher in more than one ability are exceedingly rare ... plus, they'll almost certainly be balanced with mediocrity in other areas.

As for the lower limit of 5 created by this method, I'm not going to lose any sleep over it. A -1 ability score penalty is a relative common thing, but I'm fine with -2 being exceedingly rare and -3 being beneath the range of a typical adventurer. This also leaves a little room for playing around with ability score loss due to curses, permanent injuries, etc. if that sort of thing tickles your fancy.

Obviously, this scheme isn't for everyone, but if you like the idea of adventurers being superior to the common man in courage and determination more than raw ability, it might be worth a try.