Modern D&D has a big focus on character building – pre-choosing all the options in character advancement to maximize power or effectiveness for some sort of imagined play experience. This isn’t something the book comes out and advocates, but the way the rules have been structured since 3rd edition seem to have really inculcated the practice in players, and new players seem to pick it up up.

I hate it. I realize D&D isn’t about to go back to handing out character power very heavily based on actions taken in play (looting magical items, recruiting henchmen, etc.), but I don’t want a character’s life basically pre-scripted. I bet minor rules changes can sap some of the energy behind the obsession with builds.

Let’s start with something simple: randomizing ability score increases and feat acquisition.

1. Whenever you gain a level, roll 1d6. The indicated attribute increases by 1.
   • 1: Strength
   • 2: Dexterity
   • 3: Constitution
   • 4: Intelligence
   • 5: Wisdom
   • 6: Charisma
2. If this increases the attribute to an odd number, choose a half-feat (a feat that grants +1 to an attribute in addition to other benefits, e.g. Resilient or Observant).
   1. The half-feat you choose must be one that gives an ability score boost which could match the one you got – for example, if you got a +1 to Strength you could choose Resilient (which can give a +1 to any attribute), but not Observant (which can only give a +1 to Intelligence).
   2. You don’t gain the +1 attribute from the feat (in effect, you already got it from the roll), but you do get the other “half” of the feat.
3. At levels where you would normally get an Ability Score Increase, roll for two attribute increases. If they both increase attributes to odd values, you can take two half-feats, or one regular feat. This is the only way to get regular feats under this system.
4. If you aren’t using feats, then in place of a half-feat you can learn two normal languages, one exotic language, or any one proficiency (weapon, armor, tool, skill, or saving throw) of your choice.

Under this system, a character will gain 24 attribute points on the way to 20th level. However, those will be randomly distributed, increasing each ability score by 4 points on average. Since players don’t control how these increases are allocated, you won’t see characters of a given class looking identical even if their initial approaches to building their characters are very similar. As a side benefit, you can try dropping the cap on PC attributes – go ahead and let them exceed 20. Since the increases aren’t under player control, they aren’t likely to actually pass that limit, and for those who do their characters will be all the more distinctive, since their accomplishments won’t be readily copied by others.

24 attribute increases will average out to 12 times when an attribute increases to an odd value. This should be expected to result in around 10 half-feats and 1 full feat on the way to 20th level. That may sound like a lot, but how many groups actually go all the way to 20th level, or even get particularly close? I’m pretty sure most groups stop somewhere around 10th level, and in that case we’re looking at 11 attribute increases, which means about 5-6 half feats, and about a 1/3 chance of a full feat. If that still feels like too much – and despite all pretense of analysis in game design, how things feel is a much better guide to group satisfaction – then there’s an easy fix: just give out 2 randomly-rolled increases at the levels ASIs (Ability Score Increases) are noted in each class description, and 1 randomly-rolled increase at all other even-numbered levels. This will result in 16 attribute increases on the way to 20th level, with about 7 half-feats and 1 full feat over that period.