When I was tapped to write for Dragon Age last year, one of the first things that struck me was the fact that character ability scores were randomly generated in order (with the caveats that the results were weighted to give more bonuses than penalties, the players could switch any two scores, and there are opportunities later to increase the bonuses).

This struck me at first as rather quaint and old-school, but when we generated characters for a playtest I was also reminded at how much fun random generation can be. In a way, it’s reflective of real life. You are born with certain traits and you make career choices based on your interests and what you were born with. I remember playing “iron man” (A)D&D (3d6 or 4d6-1 die in order) where I was thrilled that I could finally play a paladin as well as being disappointed by generating numerous “red shirt” characters that were only qualified to be poorly-skilled fighters.

Random generation also sparks players to create concepts that they’d never think of on their own. Depending on the RPG system, a PC could have an intensely rich background and career by the time the player is done rolling on a few charts.

Another benefit to random generation is that it’s easy to draw in new players. Essentially, six rolls in (A)D&D and its derivatives pretty much determine a character’s class and race as well as give the player an idea of her character’s physical, mental, and social traits. Other games go a step further and use random rolls to determine class/careers (Warhammer Fantasy Roleplay), physical appearance and personality (the pot-bellied lone wolves of various Palladium games) and even the size of your private parts (F.A.T.A.L.).

However, I also remember all of the arguments against character generation: it makes an unbalanced PC party, it encourages cheating, you may not get to play the concept that you want, and that the randomness of real life shouldn’t be imposed upon you when playing make-believe.

I’ve also seen many attempts at a middle ground, such as arranging scores as you see fit, rolling an extra die and dropping the lowest, automatically granting high scores in ability scores essential to your class (Rolemaster), making more rolls than necessary but keeping the best, or granting players that rolled less than the best-rolling players extra points to make up the difference (ironically often making the poorest-rolling player get a better character than the best-rolling player).

In one extreme example, I know of one D&D group that had each player roll 4d6 nine times. She’d keep the best six and drop the lowest die from each. Needless to say, most characters in that group had very high scores.

What this all boils down to is that I’ve seen players and GMs that embrace random generation as well as players and GMs that absolutely loathe it. I’ve even known some gamers to refuse to play in or run games that included it.

How about you? Do you embrace random generation in your games or do you tend to take steps to mitigate it? Have you ever added randomness to a point-buy system? How did it work out?